NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.

Modeling and Optimization of Commercial Buildings and Stationary Fuel Cell Systems

2013 Fuel Cell Seminar and Energy Exposition

Keith Wipke (presenter, NREL) for Chris Ainscough (NREL) and Dustin McLarty, Ryan Sullivan, Jack Brouwer (University of California Irvine)

October 23, 2013 Columbus, Ohio

NREL/PR-5400-60904

STA44-5

2

Acknowledgements

This project is supported by the U.S. Department of Energy’s Fuel Cell Technologies Office Project Manager and Champion, Jason Marcinkoski

3

Introduction

Tool Name: Distributed Generation Building Energy Assessment Tool (DG-BEAT) DG-BEAT is designed to allow the stakeholders (OEMs, end users, building energy managers) to assess the economics of installing stationary fuel cell systems in a variety of building types in the United States. It enables sizing and controls optimization between commercial buildings, fuel cells, energy storage, and renewables. Compiled version is available (for free).

4

Questions the Model Answers

1. What size fuel cell or size of other DG technology is best for my building?

2. What are optimal system sizes that can serve the broadest market in the most economical fashion?

3. How should fuel cell systems be dispatched to achieve benefits (energy, emissions, economics)?

4. How much do thermal and electrical energy storage systems benefit stationary FC installations?

5

Software Layout

DG-BEAT

Components

Fuel Cell

Chiller

Thermal Storage

Batteries

Solar

Wind

Demand

16 Building Types

16 ASHRAE Climates

3 Vintages

CEC Building Data

Controls

Base Load

Diurnal Dispatch

Weekly Dispatch

Load Following

Economics

Component Costs

O&M Costs

Grid Costs

Nat. Gas Costs

Manufacturing Scale Up

Analysis

Single Scenario

Multi-Building

Sensitivity

Validation

National Survey

6

16 Building Types Modeled

Modeled buildings represent ~67% of national inventory Sizes range from 5 kW to 1,500 kW

02468

10121416

Nat

iona

l Mar

ket

(GW

)

7

Economics

Utility Costs EIA forecasted NG costs Time-of-use electricity rates

20+ pre-loaded rate structures State average energy costs

Equipment Costs Installation costs, scales with capacity (kW) Operation/maintenance costs, scales with capacity and operation (kW and kWh)

Energy costs vs. demand charges Zero-export constraint 20-year NPV analysis

Includes stack replacement Analyze financing methods, interest rates, offset electricity charges

0

20

40

60

80

100

Baseline With DG

Demand Charges

Energy Charges

8

Energy Storage Shift

0

5

10

15

20

25

0 1 2 3 4 5 6 7 8

Dem

and/

Gen

erat

ion

(kW

)

Day

Electric Demand

Shifted Demand

Motivation Move demand to lower-cost times by using cold or hot water storage

9

Preliminary Results—Without Storage

• Building: L.A. Clinic/Small Hospital (avg. demand ~100 kW) • Utility: SoCal Edison • Control Strategy: Base Load

• Results: o Base load = 34 kW ~23.7%

of demand o Grid cost is $0.126/kWh

delivered o 20 year savings (NPV) is 6%

3.24 million$ 3.02

million$

6% difference

10

Preliminary Results—With Cold Water Storage

• Building: L.A. Clinic/Small Hospital (avg. demand ~100 kW) • Utility: SoCal Edison • Control Strategy: Base Load • CW Storage: 3,600 kWh • Results:

o Base load = 34 kW ~23.7% of demand

o Grid cost is $0.126/kWh delivered

o 20 year savings (NPV) is 13%

3.24 million$ 2.83

million$

13% difference

Takeaway: Hybridizing fuel cell systems for buildings can result in significant savings (double, in this one case)

11

Dispatch Controller Flowchart: How it Works

Are renewable generators present?

Renewable Power Profile

Yes Subtract

from demand Profile

Dispatch Controller

Demand System

Are there chillers or cold thermal storage

No

Apply cold

storage shift

Chillers only

Both

Subtract absorption

chiller output from cooling

demand

Meet cooling demand with

electric chillers

Are there batteries?

Apply electric storage

shift

Dispatch Fuel Cells

Use efficiency curve to

calculate fuel use

Calculate un-met electric demand

Grid Electricity

Fuel Used

Self Generation

Yes

No No

𝑩𝑩𝑩𝑩𝑩𝑩𝑩𝑩 𝑫𝑫𝑫𝑫𝑩𝑩 ≠ 𝑮𝑫𝑩𝑫𝑮𝑫𝑮𝑩𝑮𝑩 Due to renewable power and energy storage

Energy Shifting

12

Base Load

800

900

1000

1100

1200

1300

0 6 12 18 24

Dem

and/

Gen

erat

ion

(kW

)

Hour

0

50

100

150

200

250

300

350

0 6 12 18 24

Grid

Sup

ply

(kW

)

Hour

Simplest type of dispatch Grid Load

13

Diurnal Peaking

800

900

1000

1100

1200

1300

0 6 12 18 24

Dem

and/

Gen

erat

ion

(kW

)

Hour

Series1diurnal peakingDemand Generation

0

50

100

150

200

250

300

350

0 6 12 18 24

Grid

Sup

ply

(kW

)

Hour 4 x DFC300 = 1,200 kW, 4:3 turndown ratio, 10%/hour slew

Grid Load

Motivation Increased stack life (~equal kWh spread over longer time) Higher-value generation during on-peak hours Larger FC installation

14

Weekend Dip

700750800850900950

10001050110011501200

0 24 48 72 96 120144168192216

Dem

and/

Gen

erat

ion

(kW

)

Hour

Series1weekend dipDemand Generation

0

50

100

150

200

250

300

350

0 24 48 72 96 120

144

168

192

216

Grid

Sup

ply

(kW

)

Hour

Grid Load

Motivation Same as daily dispatch, but with fewer fuel cell transients

15

Control Strategy

Comparison of dispatch for our LA Clinic example 20 Year NPV Percent reduction in NPV relative to baseline

Takeaway: With storage, proper dispatch of the FC can reduce NPV even further. Diurnal peaking increases %NPV saved from 13% to 15% (relative to no FC system).

2.83 MM$

3.03 MM$

2.77 MM$

2.97 MM$

16

Optimal Size $/

kW fo

r 20

Year

Life

span

Payb

ack

Perio

d (y

ears

)

» Optimal Size: 91 kW » NPV: Savings 15% over baseline

20 Year NPV Percent reduction in NPV relative to baseline

DG-BEAT can find the optimal (lowest lifecycle cost) system for an application (again, the LA Clinic example)

If the system is too large, it will actually cost more than the baseline. Negative % reduction

17

Viable Market

Viable: Any building where 20-year NPV w/ FC is less than 20-year NPV w/o FC

DG-BEAT can estimate the viable market for a particular size of fuel cell system, using commercial building inventory data

Takeaway: For this case, the most viable markets are the NE and California

18

Viable Regions for DG

Avg. Net Present Value: Weighted average of each building type (weighted by size and # buildings in state)

Includes buildings which lose money by converting to FC

Users can calculate state-by-state values for a system type under consideration (may be useful for OEMs)

19

Future Work Topics

• Cost optimization instead of area under demand curve (brute force for energy charges only)

• Scale installation to demand with fixed FC size multiples of baseline system (e.g., 4 x 300 kW)

• Advanced controls (e.g., predictive control) with physical DG system models and comparison to ideal dispatch

• Additional building energy demand profiles o PNNL building profiles o ASHRAE 90.1-2007 standard o Easier upload of different formats by users

• CBECS o Update building inventories based on output from CBECS 2012

• Refinement of economics for scenario analyses o State incentives database (DSIRE) o Projections for changes in utility prices o Regional NG prices o Regional variation in energy vs. demand charges o Variation in utility costs and energy vs. demand costs with scale of installation o Variation in building sizes of same class (e.g., not all hospitals = 1,200kW)

20

Summary of Functionality

1) Build a DG system out of any number of subcomponents (FC, chiller, batteries…)

2) Apply to 768 building types or build campus of multiple buildings

3) Choose from four dispatch strategies 4) Evaluate lifetime costs with 30+ real world utility rates,

or build your own 5) Perform sensitivity analysis 6) Describe national market with detailed application-

specific evaluation