energy performance measurement and indicators - unido · 2/18/2016 1 energy performance measurement...

125
2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and 25 February 2016 1 Authors Liam McLaughlin Vilnis Vesma Luis Marques Almanza Acknowledge support of the Austrian Energy Agency and Marco Matteini of UNIDO 2

Upload: phamthuy

Post on 30-Jun-2019

218 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

1

Energy Performance Measurement

and indicators

Liam McLaughlin

Luis Marques Almanza

Tehran, UNIDO office

24 and 25 February 2016

1

Authors

Liam McLaughlin

Vilnis Vesma

Luis Marques Almanza

Acknowledge support of

the Austrian Energy Agency

and Marco Matteini of UNIDO

2

Page 2: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

2

Scope of the training

Focus is individual organisations

• Industry

• Large Buildings

• Public sector

Not dealing the policy level

• National EE

• Sectoral EE

• Sectoral benchmarking

3

WHY ARE WE HERE?

4

1

Page 3: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

3

Why are we here?

Purpose Importance

Stop climate change

Reduce energy cost

Use less energy

Improve energy performance

“My boss sent me”

Other

5

Requires performance improvement

• Only ISO management system standard that requires this

Requires monitoring of performance

Is that what organisations require?

Role of ISO 50006 (Energy Performance Indicators - EnPIs

and Energy Baselines - EnBs)

6

Energy Management System - ISO 50001

Page 4: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

4

How do you measure energy

performance now? Actual cost compared with budget?

kWh last month compared with the same month last year?

kWh/m2 compared with another facility

kWh/unit of production

Moving total of 12 months kWh

More complex method

7

Is this good performance?

8

kW

h o

f N

atu

ral G

as p

er

year

Page 5: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

5

Objective support for decision making

• Often subjective reasons

How much energy we are consuming?

Is consumption increasing or decreasing?

Is performance improving or not?

• Energy Performance indicators (EnPIs) & Baselines (EnB)

Are we meeting targets?

Can we verify savings from improvements?

Are we meeting budgets?

How to allocate costs

Purpose of energy metrics

9

Performance measurement options 1. M&V of a project or operational improvement

2. Critical operating parameter showing effect of

an operational change

• For example, combustion analysis results

3. Observation after awareness training

• For example, number of PCs switched off

4. Normalised whole facility indicator

5. Other normalised indicators

The last 2 are the main focus of this training10

Page 6: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

6

11

EnMS - ISO 50001 simplified

kWh(€ + CO2)

Commit to

change

Plan the

changes

Make the

changes

Check the

results

12

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 7: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

7

13

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Terminology

Energy use v energy consumption

Energy driver, factor, relevant variable, independent

variable

Expected energy consumption

Energy performance, saving, efficiency, conservation

Energy Budget

14

Page 8: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

8

Related topics

EnMS (ISO 50001)

EnMS as a metering system

Monitoring and Targeting (M&T)

Measurement and Verification (M&V)

Building Management System (BMS)

Building Energy Management System (BEMS)

15

Relevant standards

ISO 50001 (EnMS) and ISO 50004 (EnMS Guidance)

ISO 50006 (Baselines and EnPIs)

ISO 50015 (M&V of an organisation’s savings)

ISO 17747 (Calculating energy savings)

UNIDO Energy Management Capacity Building Program

16

Page 9: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

9

Is it easy to improve?

17

Discussion

What is your previous experience or views on

energy performance measurement?

18

Page 10: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

10

BUILD COMMITMENT

19

DELUSIONS AND BARRIERS

(TO IMPROVEMENT)

Build Commitment

20

2

Page 11: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

11

21

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

"How many managers have been told by their

staff that bad coal consumption was due to low

output? How is it possible for them to judge

whether this is an excuse or a reason?”

These are the opening words from a fuel efficiency bulletin, published in

1943 by the Ministry of Fuel and Power, which criticises the "ton of coal per

ton of output" metric as a misleading indicator of fuel efficiency.

The author was Oliver Lyle, managing director of the eponymous sugar

refinery, a very knowledgeable and eminent engineer who had no time

whatever for the Specific Energy Ratio. Any works engineer today will know

that SERs vary continuously for reasons nothing to do with energy efficiency.

22

Page 12: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

12

Typical example

23

Which was the worse energy

performance?

Foundry industry

WorseBetter

Typical example

24

0.805

0.870

Foundry industry

Page 13: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

13

Energy per unit of production

100

126

99 95

0

20

40

60

80

100

120

140

2008 2009 2010 2011

25

Car assembly industry

Which is right?

26

-16.74 %

+2.19 %

-8.94 %

Brewing industry

Page 14: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

14

Uses for SEC

Cost allocation

Legal or corporate compliance

Benchmarking (?)

Not useful for energy performance measurement

• Except if negligible baseload and only one relevant

variable

27

Uses for absolute energy trends

Annualised view is good for setting future budgets

Annualised view is good for monitoring spending

against budget

Good overview

Not useful for energy performance measurement

• Except if no relevant variables

28

Page 15: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

15

Uses for normalised models

Typically the only effective way to know if

performance is improving or not

Are targets being met

Whole facility

Individual SEUs, buildings, departments

29

30

Basic terminology

Energy performance indicator

Energy Baseline

Energy Target

Energy Improvement

Refe

rence E

nP

Ivalu

e

(baselin

e p

erio

d)

Curr

ent

EnP

Ivalu

e

(report

ing p

erio

d)

Energy Baseline

Energy Target

Target

Achieved !

Actual value

Source: Adapted from ISO 50006

Page 16: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

16

Energy performance indicators

31

EnPI type Example Problems

Measured energy value Annualised

consumption

Misleading results

Do not use variables

Do not measure energy efficiency

Ratio kWh per unit Misleading results

Does not account for baseload

and non-linear effects

Regression Y=mX+C

X: variable value

C: baseload

Complex if it is not linear

Uncertainty

Must be maintained and adjusted

Engineering model Energy simulation Complex

Must be maintained and adjusted

Source: ISO 50006

Energy performance indicators: Criteria

Attributes we need:

• Only responds to changes in energy performance

• Unaffected by weather, production outputs or other

relevant variables

• Direction and magnitude of change consistent with

change of performance

32

Page 17: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

17

33

But …

“MWh/tonne” does not meet these criteria

Two or more relevant variables? We cannot even

calculate such a ratio

ISO 50004 advises against Specific energy consumption

unless there is no baseload and only one variable

• Has anyone an example of such an organisation?

34

Variable

kWh

Using SEC shows

not only non-precise results (YELLOW)

but usually contrary results (RED)

REGRESSION

TREND

SEC: kWh / UNIT

baseload

baseload

Regression vs SEC

Page 18: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

18

Is it easy to show improvement?

35

Where are we going?

36

But…Are we really improving?

How can we be sure??

YES!!!

I’ll show you!!

We want to:

• Develop a model for expected performance.

• Compare actual with expected

• Quantify performance, +ve or -ve

• React to deviations

• Communicate to build commitment

Page 19: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

19

Discussion

Is Specific energy consumption (SEC)

useful?

37

MANAGEMENT INFORMATION

Build Commitment

38

3

Page 20: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

20

Management Information

Communication

Support

Commitment

Decision making

Reporting

39

DEVELOP INFORMATION

AND PLANS

40

Page 21: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

21

PAST AND PRESENT

CONSUMPTION

Develop information and plans

41

4

42

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 22: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

22

Analyze Energy Use & Consumption

Collect past and current monthly consumption data at the

facility level (energy bills).

Determine what other data may be available for analysis.

• Sub-meter data

• Interval data

• Equipment information

• Other data

Determine PAST and CURRENT consumption by use.

Note: The time period for data collected will depend on

your organization and what data is available.

43

What are my energy sources,

uses and consumption levels? Electrical, natural gas, propane, hydro, wind?

What facilities, systems or equipment are using energy?

What data do we have? Where?

What data do we need? Where?

How much energy are we consuming?

How much did we consume in the past?

What are energy consumption trends for the future?

44

Page 23: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

23

Moving total of previous 12 months (or 52 weeks, etc)

• Removes seasonal effects

• Gives a real view of comparison v budget

• Effects of a change stay for next 12 periods

• Absolute numbers

• No allowance for changing activity levels

Very useful for forecasting, you can quickly judge what

next 12 months use will be

• You need to correct for known changes in output or other activity

Annualised trends

45

What does this tell us?

46

Food industry

0

200000

400000

600000

800000

1000000

1200000

1400000

1600000

1800000

2000000

01/2

009

02/2

009

03/2

009

04/2

009

05/2

009

06/2

009

07/2

009

08/2

009

09/2

009

10/2

009

11/2

009

12/2

009

01/2

010

02/2

010

03/2

010

04/2

010

05/2

010

06/2

010

07/2

010

08/2

010

09/2

010

10/2

010

11/2

010

12/2

010

01/2

011

02/2

011

03/2

011

04/2

011

05/2

011

06/2

011

07/2

011

08/2

011

09/2

011

10/2

011

11/2

011

12/2

011

01/2

012

02/2

012

03/2

012

04/2

012

05/2

012

06/2

012

07/2

012

08/2

012

09/2

012

10/2

012

11/2

012

12/2

012

01/2

013

02/2

013

03/2

013

04/2

013

05/2

013

kW

h p

er

month

Page 24: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

24

Is this good information?

47

0

200000

400000

600000

800000

1000000

1200000

1400000

1600000

1800000

2000000

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

2010

2011

2012

2013

Electricity data in annualised view

48

12/2

009

01/2

010

02/2

010

03/2

010

04/2

010

05/2

010

06/2

010

07/2

010

08/2

010

09/2

010

10/2

010

11/2

010

12/2

010

01/2

011

02/2

011

03/2

011

04/2

011

05/2

011

06/2

011

07/2

011

08/2

011

09/2

011

10/2

011

11/2

011

12/2

011

01/2

012

02/2

012

03/2

012

04/2

012

05/2

012

06/2

012

07/2

012

08/2

012

09/2

012

10/2

012

11/2

012

12/2

012

01/2

013

02/2

013

03/2

013

04/2

013

05/2

013

15000000

15500000

16000000

16500000

17000000

17500000

18000000

18500000

19000000

19500000

20000000

kW

h p

er

year

(ELE

C)

Page 25: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

25

Actual annualised electricity

usage and costs

49

1000000

1050000

1100000

1150000

1200000

1250000

1300000

1350000

1400000

1450000

1500000

12/2

009

01/2

010

02/2

010

03/2

010

04/2

010

05/2

010

06/2

010

07/2

010

08/2

010

09/2

010

10/2

010

11/2

010

12/2

010

01/2

011

02/2

011

03/2

011

04/2

011

05/2

011

06/2

011

07/2

011

08/2

011

09/2

011

10/2

011

11/2

011

12/2

011

01/2

012

02/2

012

03/2

012

04/2

012

05/2

012

06/2

012

07/2

012

08/2

012

09/2

012

10/2

012

11/2

012

12/2

012

01/2

013

02/2

013

03/2

013

04/2

013

05/2

013

15000000

15500000

16000000

16500000

17000000

17500000

18000000

18500000

19000000

19500000

20000000

Euro

per

year

(ELE

C)

kW

h p

er

year

(ELE

C)

Consumption

Cost

Common mistakes

1. Year-to-date reporting

• Inaccurate near start of year

• Moving annual totals or averages better

• Calendar has no significance

• Why waste information from prior periods?

• Long-term history gives superior analysis

50

Page 26: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

26

Exercise 01

This data comes from the same plant

as the previous data.

• 1. Calculate and represent the annualised

trend for 2013

• What is the annual consumption in the year

ending July 2013

• 2. Which is the % change in consumption in

2013 compared to 2012?

51

Exercise 01 - Solution

52

Page 27: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

27

SIGNIFICANT ENERGY USES

(SEUS)

Develop plans and information

53

5

54

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 28: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

28

What are energy uses?

• “manner or kind of application of energy1”

• The service provided

• e.g. Light, heat, pump, cool, ventilate, convey, etc.

Significant energy uses

• Large energy uses

• Uses with good potential for savings

SEU is a central and key concept of an

EnMS

1Source: ISO 50001

What are SEUs?

55

Brainstorm:

• What do you think are the large uses?

• Where do you think there are good savings opportunities

List them

Tools:

• Motor list

• Thermal process list

• Lighting list

What to do in a multi-building organisation

• Is it the biggest buildings?

Exercise 02 Identify your SEUs

56

Page 29: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

29

Motor List

57

Source: UNIDO EnMS Tools

SEUs

58

Electricity Heat

8,651,145

1,584,839

821,876

1,525,654

809,185

1,063,759

3,838,640

47%

9%

4%

8%

4%

6%

21%

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 5,000,000 10,000,000

kWh per year

4,250,300

6,146,639

9,297,205

22%

31%

47%

Steam

Hot water

Dryers

0 5,000,000 10,000,000

kWh per year

Electricity + Heat

170,012

245,866

371,888

692,092

126,787

65,750

122,052

64,735

85,101

307,091

8%

11%

17%

31%

6%

3%

5%

3%

4%

14%

Steam

Hot water

Dryers

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 200,000 400,000 600,000 800,000

euros per year

Page 30: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

30

5,079,109

7,345,234

11,110,160

20,961,723

3,840,065

1,991,406

3,696,660

1,960,655

2,577,488

9,301,024

7%

11%

16%

31%

6%

3%

5%

3%

4%

14%

Steam

Hot water

Dryers

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 10,000,000 20,000,000 30,000,000

kWh per year

5,079,109

7,345,234

11,110,160

22%

31%

47%

Steam

Hot water

Dryers

0 10,000,000 20,000,000 30,000,000

kWh per year

20,961,723

3,840,065

1,991,406

3,696,660

1,960,655

2,577,488

9,301,024

47%

9%

4%

8%

4%

6%

21%

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 10,000,000 20,000,000 30,000,000kWh per year

SEUs

59

Electricity Heat Electricity + HeatPrimary Energy Primary Energy Primary Energy

799

1156

1748

2855

523

271

503

267

351

1267

8%

12%

18%

29%

5%

3%

5%

3%

4%

13%

Steam

Hot water

Dryers

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 1000 2000 3000

tCO2 per year

799

1,156

1,748

22%

31%

47%

Steam

Hot water

Dryers

0 1,000 2,000 3,000

tCO2 per year

2,855

523

271

503

267

351

1,267

47%

9%

4%

8%

4%

6%

21%

Refrigeration

Comp Air

Lighting

Dryers

Pumps

Ovens

Others

0 1,000 2,000 3,000

tCO2 per year

SEUs

60

Electricity Heat Electricity + HeatCO2 Emissions CO2 Emissions CO2 Emissions

Page 31: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

31

PERFORMANCE MODELS –

ONE RELEVANT VARIABLE

Develop information and plans

61

6

62

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 32: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

32

Expected consumption

We can compute expected consumption accurately if we

consider the variables that cause consumption to vary

• Production throughput?

• Weather?

• etc

We must be able to measure these variables

• (Also known as ‘driving factors’, ‘relevant variables (in ISO 50001 and ISO 50006)’,

‘energy factors’, ‘explanatory variables’, ‘independent variables’, or ‘drivers’).

63

Relevant variables

Measurable

Routinely variable

Cause consumption to vary

(or are plausibly correlated)

• Production activity…

• Weather…

• Hours of darkness…

• Distance driven…

• … etc …

64

Page 33: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

33

Static factors

Some things influence consumption but do not routinely

vary…

For example:

• Size of process equipment

• Number of luminaires in a lighting system

• Size of a building

65

Exercise 03

List some possible variables that affect energy

consumption in your organisations

• Do this for one or two SEU’s

Categorise them as variable and static

Do you know an SEU whose energy consumption is not

affected by a variable?

66

Page 34: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

34

Scatter diagram

67

Represent consumption VS

relevant variable

See the trend

Observe the dispersion

Obtain the formula

0

200

400

600

800

1,000

1,200

1,400

1,600

0 10 20 30 40 50 60

kW

h/w

eek

CDD 15

y = 18,572x + 167,84R² = 0,8926 Remember: Y= mX + c

• c and m are constants

• X is a measured “relevant variable” variable

0

200

400

600

800

1,000

1,200

1,400

1,600

0 10 20 30 40 50 60

kW

h/w

eek

CDD 15

y = 18,572x + 167,84R² = 0,8926

Scatter diagram

68

You can also use formulae in

excel c: =INTERCEPT (known_y's,known_x's)

m: =SLOPE (known_y's,known_x's)

R2 =RSQ(known_y's,known_x's)

Remember: Y= mX + c

• c and m are constants

• X is a measured “relevant variable” variable

Page 35: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

35

Understand and interpret results

Intercept:

• What does the intercept mean?: Consumption when all the

variables are 0 at the same time.

• It is the baseload in most of the cases, unless that case is outside

of the model range.

69

R2:

• What does the R2 mean?: % of variation explained by variables

• High R2:

a) If all predicted variables were included:

a) Strong correlation. Not necessarily good performance.

b) If not all predicted variables were included. Think why.

a) The other ones were not really variables.

b) Saving Opportunities in operational control.

• Low R2:

a) There are other variables.

b) Saving Opportunities in operational control.

70

Understand and interpret results

Page 36: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

36

A technical example: boiler

71

y = 1.3761x + 189.84R² = 0.9933

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

0 2000 4000 6000 8000

kW

gas

kW steam

0%

10%

20%

30%

40%

50%

60%

70%

80%

0 2000 4000 6000 8000

eff

icie

ncy

kW steam

• 1,3761 kW of gas to get each kW of steam.

• Standing losses of 189.84 kW of gas

• The efficiency is lower when the output (and input) is lower.

Example: glass furnace

72

Fixed 220,000

kWh per week

Variable

355 kWh

per tonne

Page 37: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

37

Exercise 04

73

Suppose weekly consumption characteristic is

• 220,000 kWh/week + 355 kWh/tonne

What would it be at daily intervals?

• 31,400 kWh/day + 50.7 kWh/tonne

• 31,400 kWh/day + 355 kWh/tonne

• 220,000 kWh/day + 355 kWh/tonne

• 220,000 kWh/day + 50.7 kWh/tonne

Exercise 04

74

Look at these data, from a SEU

(compressed air) in the demo

plant.

• 1. Use a scatter diagram to analyse

the relation between sliced

products volume (t) and

consumption.

• 2. What is the intercept telling us?

• 3. How many kWh will we need to

produce 10 t more each month?

Page 38: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

38

Exercise 04 - Solution

75

PERFORMANCE MODELS -

MORE THAN ONE RELEVANT

VARIABLE

Develop information and plans

76

7

Page 39: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

39

77

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Straight-line models are most common

More complex models may be appropriate

• Curved characteristics

• Multiple relevant variables

• Modelling from first principles

Expected-consumption formulae

78

x - + (

≠ ÷ √

Page 40: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

40

Expected consumption =

• c kWh per week (or per day, month etc)

• + m1 kWh per tonne of product A

• + m2 kWh per tonne of product B

• + m3 kWh per tonne of product C

Multiple relevant variables

79

Discussion

Consider a car

What are all the relevant variables for fuel

consumption?

Which are practical to measure?

Which are economical to measure?

80

Page 41: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

41

81

Have we chosen appropriate relevant

variables?

Use technical knowledge

and common sense

Test the significance of

each factor in the model

82

Testing significance of relevant variables

Use Excel’s regression

analysis tool

Page 42: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

42

83

Testing significance of relevant variables

84

Testing significance of relevant variables

Page 43: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

43

85

Testing significance of relevant variables

P-value < 0.1?

86

Testing significance of relevant variables

Page 44: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

44

87

Testing significance of relevant variables

P-value < 0.1?Coefficients

P-value:

• What does the P-value mean?: Probability of being significant.

• Low P-value:

a) The variable is significant.

• High P-value:

a) The variable is not significant.

b) Some variables are correlated. Colinearity. Check it.

c) The variable is significant but there are other variables.

d) Saving Opportunities in operational control.

88

Understand and interpret results

Page 45: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

45

89

Testing significance of the model

Significance F < 0.1?

Significance F:

• What does this mean?: Probability of being significant.

• Low Significance F:

a) The model is significant.

• High Significance F:

a) The model is not significant.

b) Some variables are non-linear.

• But:

• Low significance F + high P-value

a) Colinearity

90

Understand and interpret results

Page 46: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

46

91

Regression analysis – key points

Regression analysis is only a statistical estimate of the

effect of each relevant variable

Technical understanding of the process is critical

Operational control is an un-measurable relevant variable

• Important concept

• Often very significant

THE IMPORTANCE OF

WEATHER

Develop information and plans

92

8

Page 47: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

47

93

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Weather-related energy demand

94

Energy consumption varies because of the

weather in many industries

Space heating and cooling

Cold stores

Industries with refrigeration as an SEU

Clean rooms in pharmaceuticals,

microelectronics, etc.

Is it feasible to relate energy consumption

directly to outside-air temperature?

Page 48: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

48

What about weather-related demand?

95

We did this for an energy-intensive process…

Can we do something similar for weather-related energy

consumption?

Heating Degree Days (HDD)

and Cooling Degree Days (CDD)

“Base temperature”:

• HDD base: outside temp. above which no artificial heating is required.

CDD base: outside temp. below which no artificial cooling is required.

• Default in the UK & IRL 15.5ºC (Austria is 12C)

• Other countries differ: Lower HDD base in countries with high

standards of weatherisation

• Depends on the building construction and internal heat gains

• Can be calculated in a daily/monthly/yearly basis.

96

Page 49: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

49

How to get HDD and CDD?

www.degreedays.net

97

How to get HDD and CDD?

98

City name and press

Station Search

Choose station

Choose data options

Generate

Wait and download

Page 50: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

50

Temperature-related demand

99

Heating base

temperature

Days

De

gre

es

Temperature-related demand

100

Shaded area is proportional

to heat energy requirement

Days

De

gre

es

Page 51: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

51

Heating ‘degree-day’ figures

101

Weekly gas consumption

Notice similarities between the shapes

Plot energy against degree days

Temperature-related demand

102

Weekly degree-day values

Page 52: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

52

Temperature-related demand

103

Knowing the degree day figure, we can read off expected

gas consumption

Temperature-related demand

104

12.7

900

Expected consumption= 900 kWh/week + 12.7 kWh/degree day

Page 53: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

53

What does outside temperature affect?

Building heating and cooling energy

Refrigeration as an SEU

• Food and drink industries

Industries with critical indoor environmental conditions

• Microelectronics

• Car assembly (painting is a SEU)

Humidity can have a similar effect

105

Changing base temperature

Sometimes another base temperature is needed to get year

round data.

• E.g. many zeros at the traditional base.

• Typical in industries with refrigeration during all the year

106

Page 54: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

54

A typical cooling degree-day history

107

Weather in multivariate regression

108

In general HDD is used in Heating analysis, and CDD in

Cooling Analysis

But in some cases both need to be included

When the same system is used for heating or cooling:• heat pump

• boiler combined with absorption chiller

• Electrical heaters and cooling.

• ...

Page 55: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

55

Weather in multivariate regression

109

ELEC= (316.09*CDD15.5) + (326.30*HDD15.5) + (36415.95*OCC) – 104796.15

Total consumption

Buiding with electrical

heaters + cooling

Daylight in regression models

Significant variation in some latitudes

Prolonged overcast also possible

Affects photocell-controlled lighting

Possible sources of data

• Photocell controlling hours-run meter

• Standard tables

110

Page 56: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

56

Sources of degree-day data?

111

Published figures

Degreedays.net

Subscription services

Measure your own

• Weather station

• SCADA system

• Building energy management system

Degreedays VS Averages

112

Week 1 Week 2 Week 3

ºC kWh ºC kWh ºC kWh

Monday 9 1750 23 250 8 2000

Tuesday 10 1500 21 250 10 1500

Wednesday 14 500 18 250 12 1000

Thursday 19 250 15 250 14 500

Friday 14 500 14 500 19 250

Saturday 22 250 11 1250 23 250

Sunday 24 250 10 1500 26 250

Average temperature (ºC) 16 16 16

Total heat energy (kWh) 5000 4250 5750

Page 57: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

57

REVIEW DAY 1

113

DATA COLLECTION

Develop information and plans

114

9

Page 58: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

58

115

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Preparatory and routine phases

116

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

What data will I need?

Where will it come from?

How will I acquire it routinely?

Page 59: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

59

Data requirements

Consumption data

Relevant variable data

117

Data requirements

Equal intervals

Synchronized with assessment

interval (may be more frequent)

Continuous history

Correctly time-stamped

No extraneous values

Free of estimates

118

Page 60: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

60

Data requirements - poor layout

Discontinuous

Incorrect time-stamps

Extraneous values

Estimates

Unequal intervals

119

Relevant variables - principles

Gross throughput, not saleable

Monitor each process step separately?

Energy use should be same time period as production

• Beware of storage and WIP

120

+

Page 61: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

61

Beware “strange” data

Relevant variables - principles

121

General advice

Relevant variables and consumptions both important

• intervals to match analysis/reporting

• whether metered or not

• continuous histories

Start immediately with available data

• completeness is not essential

• do not wait for perfect coverage

122

Page 62: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

62

Data is objective information. Information is power to:

• Analyze your activities.

• Understand your organization.

• Help you taking correct decisions.

• Judge your past decisions and react.

• Measure your performance.

Comparisons are not difficult if you choose the right scale.

• Good data analysis will show the correct indicators.

• Correct indicators will help you succeed.

• Correct M&V will let you know (and show) your success.

Key Importance of data collection

123

STATISTICAL MODELS

INCLUDING ERROR AND

UNCERTAINTY

Develop information and plans

124

10

Page 63: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

63

125

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Importance of uncertainty

Data

Appropriateness

Uncertainty of the model

Avoid fruitless work

Resulting uncertainty of conclusions

126

Page 64: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

64

Scatter diagram

127

Represent consumption VS

relevant variable

See the trend

Observe the dispersion

Obtain the formula

0

200

400

600

800

1,000

1,200

1,400

1,600

0 10 20 30 40 50 60

kW

h/w

eek

CDD 15

y = 18,572x + 167,84R² = 0,8926 Remember: Y= mX + c

• c and m are constants

• X is a measured “relevant variable”

0

200

400

600

800

1,000

1,200

1,400

1,600

0 10 20 30 40 50 60

kW

h/w

eek

CDD 15

y = 18,572x + 167,84R² = 0,8926

Scatter diagram

128

You can also use formulae in

excel c: =INTERCEPT (known_y's,known_x's)

m: =SLOPE (known_y's,known_x's)

R2 =RSQ(known_y's,known_x's)

Remember: Y= mX + c

• c and m are constants

• X is a measured “relevant variable”

Page 65: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

65

R2:

• What does the R2 mean?: % of variation explained by variables

• High R2:

a) If all predicted variables were included:

a) Strong correlation. Not necessarily good performance.

b) If not all predicted variables were included. Think why.

a) They were not really variables.

b) Saving Opportunities in operational control.

• Low R2:

a) There are other variables.

b) Saving Opportunities in operational control.

129

Understand and interpret results

Low R2 vs High R2

130

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

16/0

1/2

013

23/0

1/2

013

30/0

1/2

013

06/0

2/2

013

13/0

2/2

013

20/0

2/2

013

27/0

2/2

013

06/0

3/2

013

13/0

3/2

013

20/0

3/2

013

27/0

3/2

013

03/0

4/2

013

10/0

4/2

013

17/0

4/2

013

24/0

4/2

013

01/0

5/2

013

08/0

5/2

013

15/0

5/2

013

22/0

5/2

013

29/0

5/2

013

05/0

6/2

013

12/0

6/2

013

19/0

6/2

013

26/0

6/2

013

03/0

7/2

013

10/0

7/2

013

17/0

7/2

013

24/0

7/2

013

31/0

7/2

013

07/0

8/2

013

14/0

8/2

013

21/0

8/2

013

28/0

8/2

013

04/0

9/2

013

11/0

9/2

013

18/0

9/2

013

25/0

9/2

013

02/1

0/2

013

09/1

0/2

013

16/1

0/2

013

23/1

0/2

013

30/1

0/2

013

06/1

1/2

013

13/1

1/2

013

20/1

1/2

013

27/1

1/2

013

04/1

2/2

013

11/1

2/2

013

18/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

23/0

1/2

013

06/0

2/2

013

20/0

2/2

013

06/0

3/2

013

20/0

3/2

013

03/0

4/2

013

17/0

4/2

013

01/0

5/2

013

15/0

5/2

013

29/0

5/2

013

12/0

6/2

013

26/0

6/2

013

10/0

7/2

013

24/0

7/2

013

07/0

8/2

013

21/0

8/2

013

04/0

9/2

013

18/0

9/2

013

02/1

0/2

013

16/1

0/2

013

30/1

0/2

013

13/1

1/2

013

27/1

1/2

013

11/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

23/0

1/2

013

06/0

2/2

013

20/0

2/2

013

06/0

3/2

013

20/0

3/2

013

03/0

4/2

013

17/0

4/2

013

01/0

5/2

013

15/0

5/2

013

29/0

5/2

013

12/0

6/2

013

26/0

6/2

013

10/0

7/2

013

24/0

7/2

013

07/0

8/2

013

21/0

8/2

013

04/0

9/2

013

18/0

9/2

013

02/1

0/2

013

16/1

0/2

013

30/1

0/2

013

13/1

1/2

013

27/1

1/2

013

11/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

12 variables (CDD0 and different production parameters)

R2: 0.92

1 variable=Production

R2: 0.64

Brewing industry

1 variable=CDD0

R2: 0.47

Brewing industry

Page 66: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

66

More about R2

131

y = 51.888x + 354660R² = 0.2144

0

100000

200000

300000

400000

500000

600000

0 100 200 300 400

kW

h

CDD15.5

• Main variable must be CDD.

• Regression shows low R2.

• We would have expected high R2

and a higher slope.

• Saving opportunities in

operational control. It consumes

the same in winter and in summer.

Electricity data taken from an office building in Spain.

P-value:

• What does the P-value mean?: Probability of being significant.

• Low P-value:

a) The variable is significant.

• High P-value:

a) The variable is not significant.

b) Some variables are correlated. Colinearity. Check it.

c) The variable is significant but there are other variables.

d) Saving Opportunities in operational control.

132

Understand and interpret results

Page 67: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

67

Colinearity

Two or more variables consistently change together.

Use the one that has a greater impact on consumption.

133

y = 0.3946x + 217.15R² = 0.7904

0

100

200

300

400

500

600

700

800

900

1000

0 500 1000 1500 2000

t S

liced

t Cooked

One example:

Sliced product volume

is related to cooked

product volume.

Data problems

Bad data or missing data:

• Prevention:

• Staff must understands the importance of data collection.

• Instrumentation must be always calibrated.

• Detection:

• Check data before doing a regression analysis and during monitoring

• Solution:• Check if that datum can be predicted from other variables (co-linearity).

• If it cannot be predicted:

– IN ANALYSIS: Do not use it.

– IN MONITORING: Assume that that period is neutral.

» Actual consumption = expected consumption

134

Page 68: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

68

Data problems

Non available data

• SEUs:

• Difficulty to measure some of them: Lighting.

• Variables:

• Essential variable: Start measuring it as soon as possible.

• Important (non-essential) variable: Consider reducing accuracy.

Replace it by other co-linear variable if possible.

135

Data problems

Data collection timing

• Data collection timing has to be the same for all the variables

• Beware of storage and Work in Progress (WIP)

• In regression analysis and performance reporting, it is not an

advantage to have daily consumption data if you cannot have daily

variable data.

• The less frequently collected variable sets the analysis limit.

136

Page 69: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

69

Data problems

137

Categorical variable data:

• Non-numerical variable, e.g.: Days of the week:

Variables to use

Data problems

138

Categorical variable data:

• Non-numerical variable. e.g.: Days of the week:

Sunday:(Intercept)…………….....7489.27 kWh

Saturday: (+370.50 kWh)………..7859.77 kWh

Weekday: (+6380.86 kWh)…….13870.13 kWh

Page 70: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

70

Limitations of conventional statistics

Model range. Limits of extrapolation.

139

But we cannot say this.

The relation is only linear inside the model range.

Limitations of conventional statistics

Margin of error.

• Errors in meter calibration.

• Standard error in regression analysis.

• Data frequency and different data timing:

• Which is better?

– Monthly data: R2= 0.98 Number of variables:2

– Daily data: R2= 0.75 Number of variables:8

• IT DEPENDS.

– Daily data is more specific and tend to be more reliable.

– Monthly data sometimes can help to reduce different data

timing problems.

140

Page 71: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

71

Limitations of conventional statistics

Static factor change.

• It means being out of the model range.

• A new analysis need to be done.

React to paradoxical outcomes:

• Detect and reject coincidences:

• Non-significant variables that are apparently significant.

• Negative intercept: It is rare, but possible in same cases in

multivariate regression when the intercept is outside the model

range.

Need of interpretation from a technical point of view

141

Need of technical understanding

Think. Predict variables.

Think. Predict relation between variables and consumption.

Do regression analysis. Confirm predictions.

Think. Understand and interpret results.

Think. React to paradoxical results.

142

REGRESSION IS ONLY A TOOL

Page 72: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

72

143

These are data from the demo plant

• 1. Do a multivariate regression analysis

with production volumes and CDD.

• 2. Which variables are significant?

• 3. Which variables are not significant?

Why?

• 4. Which model should we use?

• 5. What are the R2, P-value, Significance

F, intercept and the coefficients telling us?

Exercise 05

Exercise 05 - Solution

144

Page 73: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

73

ENERGY PERFORMANCE

INDICATORS AND BASELINES

Develop information and plans

145

11

146

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 74: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

74

147

Basic terminology

Refe

rence E

nP

Ivalu

e

(baselin

e p

erio

d)

Curr

ent

EnP

Ivalu

e

(report

ing p

erio

d)

Energy Baseline

Energy Target

Target

Achieved !

Actual value

Source: Adapted from ISO 50006

Energy performance indicator (EnPI)

Energy Baseline (EnB)

Energy Target

Energy Improvement

EnPI & EnB

EnB: Expected consumption = 1163449.22+(517.27*CDD5)+(1594.81*Cured)

EnPI: A comparison of baseline (expected consumption) and actual consumption

148

Page 75: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

75

EnPI & EnB

EnB: Expected consumption

EnPI: A comparison of baseline (expected consumption) and actual consumption

149

Different views, same story

150

Baseline

Brewery industry (12 variables)

Page 76: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

76

Different views, same story

151

Baseline

Brewery industry (12 variables)

Different views, same story

152

Baseline

Brewery industry (12 variables)

Page 77: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

77

Different views, same story

153

Baseline

Brewery industry (12 variables)

CUSUM

CUSUM: CUmulative SUM of deviation from expected

consumption

• Equals the sum of the residuals

Key technique for…

• Target-setting

• Diagnosing changes in performance

• Tracking savings achieved

154

Page 78: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

78

Understanding cusum

Which chart(s)…

indicate a single fault has recently occurred and not yet been cured?

indicate a short period of waste, which has now been corrected?

indicate a successful energy-saving measure has been implemented?

indicate potential to set a more aggressive target?

suggest the target may have been set too aggressively?

155

FORECASTING: TARGET

SETTING AND BUDGETING

Develop information and plans

156

12

Page 79: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

79

157

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Terminology

Forecasting

Target setting

Budgetting

Long term planning

158

Page 80: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

80

Forecasting

Predicting or estimating future energy consumption

Predicting future energy prices

Estimating savings from:

• Energy saving projects

• Operational control

• Monitoring and corrective actions

159

Three classes of “target”

1. Aspirational

• “Top down” possibly corporate

• Aggregate e.g. site-wide

• May be arbitrary

• Deliberately “stretching” ??

2. Bottom up based on action plans

• Based on what can actually be achieved

• Revised continuously

• Agreed with interested parties

3. Based on previous best performance

• CUSUM

In all cases should be “tough but fair”

160

Page 81: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

81

Target (kWh)

Action plans (kWh)

What we want to

achieve

How we are

going to

achieve it

EnPis

How we measure if

we are being

successful

161

1. Aspirational targets

Top down: to challenge and drive improvement

• Corporate target, e.g. 5%

• Based on national targets e.g. EU2020

Should be reflected in budgets

• Padded budgets are a licence to waste energy

Should be reflected in performance monitoring

• i.e. reduce expected consumption by x%

162

Page 82: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

82

2. Bottom up approach

163

Identify all opportunities

Decide which you will action

The total of these is the target savings

• Consider the effect of operational control

• And reaction to deviations

Subtract from expected consumption

3. Previous best performance

Previous best performance based on regression

model(s)

Not necessarily best possible performance

Has been achieved with existing equipment and

people

• No investment required

164

Page 83: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

83

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

165

Setting an “aggressive but achievable”

target

Raw data

y = 501,86x + 14343R² = 0,878

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

166

Setting an “aggressive but achievable”

target

Regression line

Page 84: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

84

167

Setting an “aggressive but achievable”

target

REMEMBER:

CUSUM: Cumulative difference

between the actual consumption

and the expected consumption

0

10000

20000

30000

40000

50000

60000

70000

80000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51

Initial CUSUM

168

Setting an “aggressive but achievable”

target

Period of best performance identified

0

10000

20000

30000

40000

50000

60000

70000

80000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51

Page 85: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

85

169

Setting an “aggressive but achievable”

target

y = 501,86x + 14343R² = 0,878

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

Raw data

170

Setting an “aggressive but achievable”

target

y = 501.86x + 14343R² = 0.878

y = 550.44x + 7969.9R² = 0.977

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

Regression of best performance period only

Page 86: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

86

171

Setting an “aggressive but achievable”

target

y = 550.44x + 7969.9R² = 0.977

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

Best performance adopted as target

172

Setting an “aggressive but achievable”

target

Directly read

potential savings

0

50000

100000

150000

200000

250000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51

CUSUM relative to best performance

Page 87: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

87

Low R2 vs High R2

173

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

16/0

1/2

013

23/0

1/2

013

30/0

1/2

013

06/0

2/2

013

13/0

2/2

013

20/0

2/2

013

27/0

2/2

013

06/0

3/2

013

13/0

3/2

013

20/0

3/2

013

27/0

3/2

013

03/0

4/2

013

10/0

4/2

013

17/0

4/2

013

24/0

4/2

013

01/0

5/2

013

08/0

5/2

013

15/0

5/2

013

22/0

5/2

013

29/0

5/2

013

05/0

6/2

013

12/0

6/2

013

19/0

6/2

013

26/0

6/2

013

03/0

7/2

013

10/0

7/2

013

17/0

7/2

013

24/0

7/2

013

31/0

7/2

013

07/0

8/2

013

14/0

8/2

013

21/0

8/2

013

28/0

8/2

013

04/0

9/2

013

11/0

9/2

013

18/0

9/2

013

25/0

9/2

013

02/1

0/2

013

09/1

0/2

013

16/1

0/2

013

23/1

0/2

013

30/1

0/2

013

06/1

1/2

013

13/1

1/2

013

20/1

1/2

013

27/1

1/2

013

04/1

2/2

013

11/1

2/2

013

18/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

23/0

1/2

013

06/0

2/2

013

20/0

2/2

013

06/0

3/2

013

20/0

3/2

013

03/0

4/2

013

17/0

4/2

013

01/0

5/2

013

15/0

5/2

013

29/0

5/2

013

12/0

6/2

013

26/0

6/2

013

10/0

7/2

013

24/0

7/2

013

07/0

8/2

013

21/0

8/2

013

04/0

9/2

013

18/0

9/2

013

02/1

0/2

013

16/1

0/2

013

30/1

0/2

013

13/1

1/2

013

27/1

1/2

013

11/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

0

50000

100000

150000

200000

250000

300000

09/0

1/2

013

23/0

1/2

013

06/0

2/2

013

20/0

2/2

013

06/0

3/2

013

20/0

3/2

013

03/0

4/2

013

17/0

4/2

013

01/0

5/2

013

15/0

5/2

013

29/0

5/2

013

12/0

6/2

013

26/0

6/2

013

10/0

7/2

013

24/0

7/2

013

07/0

8/2

013

21/0

8/2

013

04/0

9/2

013

18/0

9/2

013

02/1

0/2

013

16/1

0/2

013

30/1

0/2

013

13/1

1/2

013

27/1

1/2

013

11/1

2/2

013

25/1

2/2

013

kW

h

Actual

Regression model

12 variables (CDD0 and different production parameters)

R2: 0.92

1 variable=Production

R2: 0.64

1 variable=CDD0

R2: 0.47

Brewing industry

-1000000

-500000

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

09/0

1/2

013

16/0

1/2

013

23/0

1/2

013

30/0

1/2

013

06/0

2/2

013

13/0

2/2

013

20/0

2/2

013

27/0

2/2

013

06/0

3/2

013

13/0

3/2

013

20/0

3/2

013

27/0

3/2

013

03/0

4/2

013

10/0

4/2

013

17/0

4/2

013

24/0

4/2

013

01/0

5/2

013

08/0

5/2

013

15/0

5/2

013

22/0

5/2

013

29/0

5/2

013

05/0

6/2

013

12/0

6/2

013

19/0

6/2

013

26/0

6/2

013

03/0

7/2

013

10/0

7/2

013

17/0

7/2

013

24/0

7/2

013

31/0

7/2

013

07/0

8/2

013

14/0

8/2

013

21/0

8/2

013

28/0

8/2

013

04/0

9/2

013

11/0

9/2

013

18/0

9/2

013

25/0

9/2

013

02/1

0/2

013

09/1

0/2

013

16/1

0/2

013

23/1

0/2

013

30/1

0/2

013

06/1

1/2

013

13/1

1/2

013

20/1

1/2

013

27/1

1/2

013

04/1

2/2

013

11/1

2/2

013

18/1

2/2

013

25/1

2/2

013

kW

h

R2=0.92

R2=0.64

R2=0.47

CUSUM: Low R2 vs High R2

174

Actual-Expected CUSUM

Act-Exp CUSUM+1%

Act-Exp CUSUM+10%

Act-Exp CUSUM-1%Act-Exp CUSUM-10%

Page 88: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

88

Discussion

A tough but fair target yields a cusum chart with…

• sustained horizontal sections and

• no sustained downward sections

175

Budgets

Estimate next period’s consumption

• Current performance adjusted by known changes in:

• Production and other relevant variables

• Savings targets

Estimate energy prices

176

Page 89: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

89

Forecasting consumption

Take the formula from the regression model

Insert the forecast values foe each variable

The result is the forecast energy consumption

177

We did a regression exercise with data from the demo

plant, and we found that the expected consumption was:

kWh=830128.88+(298.52*CDD5)+(1193.06*Cured)+(282.88*Cooked)

• 1. Forecast the monthly

consumption in 2013 if the

expected production is 25%

higher than in 2012.

• 2. What is the budget for 2013 if

electricity is 12c/kWh.

Exercise 06

178

Page 90: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

90

As we expect an increase of 25% in production volumes, the

forecast formula is:

kWh=830128.88+(298.52*CDD5)+(1193.06*Cured*1.25)+(282.88*Cooked*1.25)

Exercise 06 - Solution

179

MONITORING, VERIFICATION

AND REPORTING

180

Page 91: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

91

MONITORING ENERGY

PERFORMANCE

Monitoring, Verification and Reporting

181

13

182

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 92: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

92

Monitoring performance

Budget v actual spending

Expected v actual consumption

• Actual minus expected

• Actual divided by expected

Target v actual consumption

• Actual minus target

• Actual divided by target

CUSUM

183

Routine monitoring

ISO 50001 §4.6.1 (e)

• Demands regular comparison of actual and expected consumption

Our interpretation:

• Consider once a week

• Daily is better but is unusual

• Monthly is too long – waste will accumulate

• We need rapid detection and prioritisation of unexpected excess consumption

184

Page 93: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

93

Data from 2011 used to develop the

expected consumption formula

Monitoring Performance

185

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Actual consumption in 2012

Monitoring Performance

186

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Page 94: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

94

Monitoring Performance

Expected consumption is the BASELINE.

It is the consumption that we should have if

the performance is the same as last year,

based on the relevant variables

187

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Monitoring Performance

The Energy Perfomance Coefficient is

the Actual Consumption divided by the

expected consumption

188

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Page 95: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

95

Monitoring Performance

The actual savings are

the difference between

actual consumption and

expected consumption

For example, in January

we saved 26682 kWh

189

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Monitoring Performance

The actual savings CUSUM are the

cumulative savings from the beginning

For example, from January to June we

saved 198320 kWh

190

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Page 96: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

96

Monitoring Performance

The target consumption is the

consumption we want to have.

For example, the target here is to save

2.5%

191

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Monitoring Performance

We can also compare our consumption with the target.

For example, from January to May the target savings

were 227890 kWh and we have saved 198320 kWh, so

it is less than the target.

192

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

Page 97: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

97

193

Expected consumption = 116344.22+(517.27*CDD5)+(1594.81*Cured)

-250000

-200000

-150000

-100000

-50000

0

50000

12/1

1

01/1

2

02/1

2

03/1

2

04/1

2

05/1

2

06/1

2

07/1

2

08/1

2

09/1

2

10/1

2

11/1

2

12/1

2

Actual Savings

Target Savings

Baseline

Monitoring Performance

194

• 1. Did the demo plant meet the target in 2013 (2.5%)?

• 2. How many kWh did they save before August? And in

the whole 2013? Which was the best month?

• 3. Compare results with annualised results (Exercise 01)

Exercise 07

Page 98: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

98

Exercise 07 - Solution

195

Exercise 07 - Solution

196

Page 99: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

99

Detect deviations

• Green: Actual consumption lower than the target consumption.

• Yellow: Actual consumption lower than expected consumption but

higher than the target.

• Red: Actual consumption higher than expected consumption.

197

-

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000

1,600,000

1,800,000

2,000,000

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000

Gas E

nerg

y (

GJ)

Urea (t)

y = 23.94x

Regression vs SEC

198

Petrochemical

industry

DateDriver

Urea (Ton)Gas Energy

(GJ)

92-1 60,975 1,459,756

92-2 60,439 1,433,852

92-3 60,714 1,419,236

92-4 55,317 1,387,274

92-5 50,877 1,308,811

92-6 60,266 1,453,399

92-7 56,554 1,353,021

92-8 57,929 1,379,231

92-9 55,308 1,431,928

92-10 26,606 796,450

92-11 24,672 835,078

92-12 57,553 1,398,561

Ratio

23.94

23.72

23.38

25.08

25.73

24.12

23.92

23.81

25.89

29.93

33.85

24.30

average

23.94

-

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000

1,600,000

1,800,000

2,000,000

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000

Gas E

nerg

y (

GJ)

Urea (t)

y = 23.94x

y = 17.989x + 364474R² = 0,978

-

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000

1,600,000

1,800,000

2,000,000

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000

Gas E

nerg

y (

GJ)

Urea (t)

y = 23.94x

Page 100: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

100

Regression vs SEC

199

+0.16 %

-1.82 %

Cement

industry

REGRESSION

RATIO

2 variables:

- Clinker

- Cement

R2=0.93

Regression vs SEC

200

Cement

industry

REGRESSION

RATIO

Page 101: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

101

MEASURE AND VERIFY THE

RESULTS

Monitoring, Verification and Reporting

201

14

202

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 102: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

102

Why is measurement of savings

important?

Measurement & Verification (M&V)

Requirement of ISO 50001

Requirement of good management practice

Improves trust in results

Gets backing for further similar projects

Underpins performance contracting

May reveal avoidable underperformance

203

Flavours of “expected” consumption

What we should have used

• For routine ongoing assessment of performance

• Based on tough-but-achievable formula

What we would have used

• Based on historical performance characteristic

204

Page 103: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

103

Evaluating savings

Use the expected-consumption formula that applied

before the energy-saving project

• called the “historical baseline” formula

Feed in the relevant-variable values measured after the

project

This tells you (approximately) how much would have been

used if the project had not been carried out

205

Example

Before improvement, a set of air compressors consumed:

50 000 kWh per week plus 0.12 kWh per m3 of air

After improvement, in a particular week the air throughput

was 1 900 000 m3 and electricity consumption was

273 000 kWh

At baseline performance consumption would have been

50 000 + ( 0.12 x 1 900 000 ) = 278 000 kWh

So savings were 278 000 - 273 000 = 5 000 kWh

206

Page 104: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

104

Exercise 08

207

We start with a baseline model of performance.

• Baseload 600 kWh per week

• Product X 300 kWh per tonne

• Product Z 7 kWh per litre

• Space heating 40 kWh per degree day

Exercise 08

208

Suppose we implement an energy project…

…then in the ten-week period following its installation we

record the following totals:

• Gas consumption 36,600 kWh

• Product X 35 tonne

• Product Z 3,000 litre

• Weather 80 degree days

How much gas did we save during those ten weeks?

Page 105: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

105

Exercise 08 - Solution

209

kWh

Standing loads 600 kWh per week x 10 weeks = 6,000

Product X 300 kWh per tonne x 35 tonne = 10,500

Product Z 7 kWh per litre x 3,000 litre = 21,000

Space heating 40 kWh per degree day x 80 degree day = 3,200

Total expected consumption 40,700

Actual consumption 36,600

Avoided energy consumption 4,100

ISO 50015

Relate M&V process to ISO 50001 philosophy,

methodology and specific vocabulary (2014)

Establish a common set of principles and guidelines to be

used for measurement and verification (M&V)

Can be used independently, or in conjunction with other

standards or protocols.

It is can be used to verify savings in single projects or as a

part of an EnMS (ISO 50001)

210

Page 106: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

106

ISO 50015

M&V Principles

M&V Plan

Implementation of M&V Plan

Uncertainty

Documentation

211

ISO 50015

M&V Principles

• appropriate accuracy and management of uncertainty

• transparency and reproducibility of M&V process

• data management and measurement planning

• competence of the M&V practitioner

• impartiality

• confidentiality

• use of appropriate methods

212

Source: ISO 50015

Page 107: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

107

ISO 50015

M&V Principles

M&V Plan

• Establish and document an M&V plan

• Data-gathering. Aligned with ISO50006.

• Verify the implementation of Energy Performance Improvement

Actions (EPIAs)

• Conduct M&V analysis

• Report M&V results and issue documentation

• Review the need to repeat the process

213

Source: ISO 50015

ISO 50015

M&V Principles

M&V Plan

Implementation of M&V Plan

• Data gathering

• Verification of the implementation of the EPIA(s)

• Observation anticipated or unforeseen changes

• M&V analysis

• M&V reporting

• Review the need to repeat the process

Uncertainty

Documentation

214

Source: ISO 50015

Page 108: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

108

The M&V Plan - contents

• Scope and purpose

• EnPIAs (=ECM)

• M&V boundaries

• Preliminary plan

assessment

• Relevant variables and

static factors

• EnPIs

• Calculation Method

• Data-gathering plan

• Baseline and

adjustments

• Resources

• Roles and

Responsibilities

• Documentation

215

Source: ISO 50015

IPMVP

216

International Performance Measurement

and Verification Protocol (1994-1995)

Vol. I - Concepts and Options for

Determining Energy and Water Savings

Presents a framework and defines terms

used in determining‘savings’after

implementation of a project

IPMVP is more strict with R2 values. Only

models with R2 higher than 0.75 are

recommended. It uses the T-stat instead of

the p-value.

Page 109: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

109

EFFECTIVE REPORTING

Monitoring, Verification and Reporting

217

15

218

EnMS – conceptual cycle

BUILD COMMITMENT

Decision making and

support

Reporting

DEVELOP INFORMATION

AND PLANS

Opportunity list and action

plan

Technical audits and

operation control review

Collect energy bills and

sub-meter data

Analyze past and present

energy consumption

Forecasting: Targets and

budgets

Develop (or review)

baselines and EnPis

Identify and quantify SEUs

Identify relevant variables

and collect past data

IMPLEMENTATION

Procurement and Design

Operational control

Training

Implement action plan

CHECKING

Verify results of action plan

Compare actual and target

(or expected) consumption

Investigate and correct

significant deviationskWh(€ + CO2)

Page 110: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

110

Purpose

To help you communicate the results of analysis

In an appropriate manner

Without annoying the recipients

Support decision making

Who needs what?

Levels of detail for different levels of management

219

Compare and contrast

220

Page 111: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

111

Less can be more

Bar lengths easier to compare

Ranking contains information

Monochrome• Readable as a photocopy

• Or on a defective projector

• No colour perception problems

Note also…• No 3-D effects

• No borders or background

• No grid lines or tick marks

• No values

• Horizontal orientation: easy to read legend

221

Beware Excel’s repertoire

222

Page 112: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

112

“Line” chart type

223

X-Y scatter diagram

224

Page 113: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

113

“Stacked bar” chart

225

Home energy-saving measures

0 20 40 60 80 100 120

Loft insulation

Wall insulation

Draughtproofing

Heating controls

Efficient boiler

Low-energy lighting

Double glazing

Already fitted Fitted by respondent

“Column” chart

226

0

5

10

15

20

25

30

35

1960- 1965- 1970- 1975- 1980- 1985- 1990- 1995- 2000-

Page 114: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

114

Column chart with overlapping time axis

227

CONSUMPTION

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

4 5 6 7 8 9 10 11 12 1 2 3

MONTH

KW

H

98/99

99/00

00/01

01/02

Why do managers ask for reports?

For the record

To help make choices

To trigger action

228

Page 115: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

115

Report progress

• Green: Actual consumption lower than the target consumption.

• Yellow: Actual consumption lower than expected consumption but

higher than the target.

• Red: Actual consumption higher than expected consumption.

229

Reporting: summary

Use charts to give quick impressions of trends and

relationships

Use tables where specific detail must be accessible

In action reports and ‘dashboards’, focus filter and flag

230

Page 116: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

116

ADDITIONAL TOPICS

ISO 50006 ENPIS AND

BASELINES

Additional Topics

16

Page 117: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

117

ISO 50006

Measuring Energy Performance using Energy Baselines

and EnPIs

1. Obtain relevant energy performance information from

the energy review

2. Identify energy performance indicators

3. Establish energy baselines

4. Use EnPIs and energy baselines

5. Maintain and adjust energy EnPIs and energy

baselines

233

Source: ISO 50006

ISO 50006

1. Obtain relevant energy performance information from

the energy review

• Define boundaries

• Define and quantify energy flows

• Define and quantify relevant variables

• Define and quantify static factors

• Gather data

234

Source: ISO 50006

Page 118: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

118

2.Identify energy performance indicators

• Identify uses

• Determine the specific energy performance

characteristics to be quantified

Beware of ratios

ISO 50006

235

Source: ISO 50006

ISO 50006

3. Establish energy baselines

• Determine a suitable baseline period

• Determine and test energy baselines

236

Source: ISO 50006

Page 119: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

119

ISO 50006

4.Use EnPIs and energy baselines

• Determine when normalization is needed

• Calculate energy performance improvements

• Communicate changes in energy performance

237

Source: ISO 50006

ISO 50006

4.Use EnPIs and energy baselines

238

Source: ISO 50006

Page 120: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

120

ISO 50006

5. Maintain and adjust EnPIs and EB. Why?

• Static factor changes

• Energy use change

• Data availability

• Data frequency

• Target change

• Following predetermined method.

• Management Review.

• …

239

Source: ISO 50006

HOW MUCH ENERGY CAN AN

ENMS SAVE IN YOUR

ORGANISATION?

Additional Topics

17

Page 121: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

121

How much could we save here?

241

y = 51.888x + 354660R² = 0.2144

0

100000

200000

300000

400000

500000

600000

0 100 200 300 400

kW

h

CDD15.5

• Main variable must be CDD.

• Regression shows low R2.

• We would have expected high R2.

• Saving opportunities in

operational control. It consumes

the same in winter and in summer.

Electricity data taken from an office building in Spain.

y = 501,86x + 14343R² = 0,878

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

242

How much could we save here?

Regression line

Page 122: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

122

243

How much could we save here?

y = 501.86x + 14343R² = 0.878

y = 550.44x + 7969.9R² = 0.977

0

10000

20000

30000

40000

50000

60000

0 10 20 30 40 50 60 70 80 90

kW

h

HDD15.5

Regression of best performance period only

Process

For multivariate show best recent performance v average

performance

Show best coefficients v average and worst coefficients

for each variable

Quantify potential savings

This is an indication of how much better management

could save with existing people and equipment.

244

Page 123: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

123

And remember…

Low R2 can show you opportunities to improve

BUT:

High R2 does NOT mean good performance

High R2 does NOT mean lack of low cost saving potential

High R2 JUST shows a strong correlation.

245

BENCHMARKING

Additional Topics

18

Page 124: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

124

Benchmarking

247

If MWh/tonne is not meaningful…

How can we compare one plant with another?

How can we compare one building with another?

Same building on opposite sides of a street?

248

Benchmarking: Multivariate regression

Page 125: Energy Performance Measurement and indicators - UNIDO · 2/18/2016 1 Energy Performance Measurement and indicators Liam McLaughlin Luis Marques Almanza Tehran, UNIDO office 24 and

2/18/2016

125

SUMMARY AND NEXT STEPS

19