effective hci practices workshop - canadian biomass · wpac safety conference effective hci...

WPAC Safety ConferenceEffective HCI Practices Workshop

Human Centered Solutions®

Helping People PerformTM

www.applyHCS.com

Prince George, BCJune 13, 2017

This document and the information contained herein are proprietary of Human Centered Solutions, LLP, and are confidentially disclosed to representatives of the Wood Pellet Association. This document and the information contained herein are made available only to employees of the aforementioned organization. This document, and any copy thereof, and the information contained herein shall be maintained in strictest confidence and shall not be disclosed or distributed (a) to WPA member employees for whom such information is not necessary in connection with their assigned responsibilities, or (b) to persons who are not bound by a non-disclosure agreement with WPA. There shall be no exceptions to the terms and conditions set forth herein except as authorized in writing by HCS.

© 2017 Human Centered Solutions® LLC/LLP



Introduction Case Study HCI PracticesSA & Psych Wrap-Up

Workshop Protocol

Ø Questions?Ø Breaks?Ø Share Your ExperiencesØ Expectations?Ø Describe challenges you have had?Ø Parking Lot

¾ Additional Information?

Participation is encouraged

©2017 Human Centered Solutions, LLP; All rights reserved. 2




HCI Workshop Agenda13:3013:5014:30

15:00

15:30

16:40

Ø IntroductionØ Situation Awareness & Psychology Crash-courseØ HCI “Best in Class” Case Study

¾ Break

Ø HCI Effective Practices¾ Operator Interface¾ Alarm Management¾ Control Room & Console Design¾ Operations Procedures

Ø Wrap-up





What is Human Factors?

Ø Human Factors (Chapanis, 1985)¾ Discovery and application of information about human behavior

� abilities, limitations, and other characteristics¾ To the design of:

� tools, machines, systems, tasks, jobs and environments¾ For productive, safe and effective human use

Ø Objectives (Sanders & McCormick, 1993)¾ Enhance the effectiveness and efficiency of human activities,

often with focus on work.¾ Enhance desirable human values including:

� improved safety, reduced fatigue and stress, increased comfort, greater user acceptance, increased job satisfaction and improved quality of life





Human Centered SolutionsØ We apply human factors principles and design

methodologies to:¾ Enhance effectiveness and safety of human activities in and

around control rooms

Roles & Responsibilities

Training

Control Room& Console Design

Operator Interface Design

Alarms

Operating ProceduresManagement of

Change

Operator Staffing& Workload

FatigueManagement

CorporateCulture Process

Control





Human Centered SolutionsHelping People Perform

Founded in 1994

Creating a new paradigm for the operation of

complex industrial plants, with solution concepts

that improve Operations’ ability to prevent and respond to abnormal

situations.

www.asmconsortium.org

Abnormal Situation ManagementA Joint Research & Development Consortium





Sources of Abnormal Events

Established in literature ; confirmed by 18 plant studies - US, Canada, & Europe

People:• Fail to detect problems in

reams of data• Are required to make

hasty interventions• May be unable to make

consistent responses• May be unable to

communicate well

Process 20%

Mostly Preventable

Equipment 40%

Often Preventable

People 40%

Almost AlwaysPreventable





ASM & HCS

Ø Principal and Contributing Authors of ASMC Guidelines ¾ Effective Operations Practices¾ Effective Console Operator

Interface Design Effective Procedural Practices

¾ Effective Alarm Management Practices

¾ Effective Alarm Rationalization Practices

¾ Effective Change Management Practices in HMI Design

Ø Technical leadership and contributions as HCS on ASM Projects¾ Procedural Operations ¾ Future Role of the Operator ¾ First-line Leadership ¾ ASM HMI Business Case Study ¾ Advanced Operator Interface

Framework ¾ HMI Visualization Techniques¾ Alarm Trend Display¾ Alarm Sounds for Multi-Console

Control Rooms™ ¾ Root Cause Analysis & Process Safety

Metrics¾ Value of Integrated Trending in HMI

8©2017 Human Centered Solutions, LLP; All rights

reserved.





15:00

15:30

16:40


¾ Break

Ø HCI Effective Practices¾ Operator Interface¾ Alarm Management¾ Control Room & Console Design

Ø Wrap-up





What is Situation Awareness (for an individual)?

Ø Put simply, Situation Awareness is “knowing what is going on round you so you can figure out what to do” (Adam, 1993)

Ø Aviation research identified poor situation awareness as a key factor:¾ Military aviation mishaps (Hartel, Smith & Prince, 1991)¾ Accidents among major airlines (Endsley, 1995)

Ø These findings have motivated studies in decision-making and training methods to improve pilot situation awareness

Situation Awareness

ExternalCues

Mental Model

Perception Comprehension Projection Actions





Abnormal Situation Management

Ø A major responsibility of process operators is to prevent and respond to abnormal situations

Ø Abnormal situations extend, develop, and change over time increasing the complexity of interventions

Ø To prevent and respond to abnormal situations, operators must continuously maintain situation awareness¾ Where the process is, where it is going, and how quickly it is

going there¾ What tasks to perform, how to perform them, and when to

perform them





SA Failure Studies in Aviation

Ø Research in the fields of military and civil aviation has identified situation awareness failures as the leading contributor to accidents

Ø One study found that 88% of major airline accidents involving pilot error were problems with SA rather than decision making or flight skills (Endsley, 1995)





SA Failure Studies in AviationØ NASA Aviation Safety

Reporting System reports show the following Situation Awareness causes of accidents

Perception (76%)Ø Information not observedØ Information not availableØ Information difficult to detectØ Memory error Ø Misperception of informationComprehension (19%)Ø Misinterpretation of dataØ Over reliance on expectationsØ Incomplete/incorrect mental modelProjection (5%)Ø Attention overloadØ Over-reacting to current trendsØ Incomplete/incorrect mental model





Transmission Pipeline Rupture (CA)Ø In 2010, a 30-inch diameter segment of a gas transmission pipeline

ruptured in a residential area in San Bruno, CA resulting in gas release, explosion & fire

Ø Investigations found multiple root causes, however a few were related to control room operations:¾ Operators lacked assigned roles &

responsibilities ¾ Operators did not notice the dropping

pressures at one station after the rupture

¾ Alerted by staff at another facility so ittook longer to stop the flow of gas to the rupture site – Level 1 SA Perception Failure

Ø What would have helped:¾ Specific assignments or operators to dedicated areas¾ An overview display showing critical process values (e.g., dropping pressures) to

help operators perceive important/critical information©2017 Human Centered Solutions, LLP; All rights

reserved. 14




Improving Situation AwarenessØ Four basic principles for

improving situation awareness:1. Employ what we know of human

information processing to influence information perception

2. Understand the influence mental models have on operator performance and evaluating information

3. Understand how alertness can effect human performance and situation awareness, and how alertness levels can be influenced

4. Understand what comprises effective teamwork required for situation awareness and abnormal situation management recovery

MaintainingAlertness

Processing Information

Teamwork

Evaluating Information

EffectiveSituation Awareness





Human Information Processing

Ø Attention influences each of the information processing elements

Situation Awareness

ExternalCues

Mental Model

Perception Comprehension Projection Actions Attention Resources

Long-term Memory

Short-term Memory

CognitionSensory

Processing Perception Response Selection

Response Execution






15:00

15:30

16:40


¾ Break


Ø Wrap-up





Operator Interface Case StudyASM Consortium® HMI Evaluation

ObjectiveØ Develop a case study to illustrate the potential impact

of Human-Centered Operator Interface concepts¾ Hypothesis: the human-centered style of operator

interface improves operator performance for incident avoidance and in abnormal situations, in comparison to the traditional design

ApproachØ Compare operator performance in high fidelity

simulator environment with two HMI stylesErrington, J., Reising, D., Bullemer, P., DeMaere, T., Coppard, D., Doe, K., and Bloom, C. (2005). Establishing human performance improvements and economic benefit for a human-centered operator interface: An industrial evaluation. Proceedings of the Human Factors and Ergonomics Society 49th Annual Meeting, Orlando, FL, pp. 26-30.





GU U

U GG U

U

Traditional Console Simulator

GU U

U GG U

U

Traditional Console Simulator

The Side-by-Side Comparison

vs. ASM-style Console Simulator

GG

GG

G GPCPC

ASM-style Console SimulatorASM-style Console Simulator

GG

GG

G GPCPC

Human-Centered Console Simulator





Operator Interface Case Study

Biases against our Hypothesis…

Ø As a comparison, these “traditional” displays represent a “better practice” than the industry norm for operating schematics ¾ This fact makes for a conservative comparison in this

study’s performance testing

Ø This unit’s training simulator console more closely matched that of the actual plant than that of the Human-Centered interface¾ This fact made the study an even more conservative

evaluation of the hypothesis





Operator Interface ConceptKey Design Features

Ø Single, Integrated View of Multi-Level Hierarchy

Ø Simultaneous views of increasing detail¾ Level 1 – Console Overview ¾ Level 2 – Unit Summary ¾ Level 3 – Equipment detail ¾ Level 4 – Group & Point detail

Ø Effective Window Management and Layout

Ø Effective Navigation SchemeØ Visual Coding SchemeØ Integrated alarm managementØ Integrated Trending





Case Study Experimental Design

Ø The case study involved 2 groups of operators ¾ Pre-test – Establish if there were any

differences in operations and plant experience between the 2 groups

¾ Scenario testing – Establish if there were any performance difference in incident detection, incident prevention between the 2 interfaces� Tested the operators on 4 matching scenarios

Ø A total of 21 operators: ¾ 10 for the Traditional interface design ¾ 11 for the Human-Centered interface design





Case Study Scenarios

Ø Used 4 scenarios in the operator performance evaluation¾ Scenarios were matched so that there was a similar

development time and matching instrumentation� This allowed for better isolation of the effect between the

operator interfaces on operator performance for each scenario

Ø The 4 scenarios were¾ A main process steam turbine vacuum problem¾ A main process compressor suction pressure transmitter drift¾ A main process compressor discharge pressure safety valve

(PSV) passing to flare¾ A turbo expander bypass valve drift open





Advanced

Case Study ResultsPre-Test Results

Ø No average differences between the two groups of operators for: ¾ Number of years experience

as an operator ¾ Number of years experience

as an operator at this company

¾ Number of years experience as a console operator

¾ Number of areas qualified in

Group Comparison Metrics





Case Study ResultsScenario Evaluation Results

Ø Overall, the operators using the Human-centered interface were more proactive, orienting to the problem an average of 4 minutes faster

Ø Anecdotally, for the first scenario with the Traditional console, an alarm rang in which oriented them to the problem faster, but…¾ They didn’t solve the

problem faster! (see Next slide)

Operator performance measure: Time to Orient

Advanced

Total Time(min)

20

10

0





ASMTraditional

Interface Type

2,000

1,500

1,000

500

0

797

Scenario 8 Total Time(Secs)




0

10

20

1 2 3 4 1 2 3 4Scenario

Total Time (min)

Traditional Advanced

Case Study ResultsScenario Evaluation Results

Ø 41% Less time to deal with the event and as a group, were more consistent in doing so!¾ An average of 10.6 minutes

vs. 18.1 minutes

Ø 38% improvement in detecting event before alarm

Ø 26% higher success rate

Operator performance measure: Total Completion Time






15:00

15:30

16:40


¾ Break


Ø Wrap-up





Effective Display Design PrinciplesØ Design Principles

1. Support the Operators Scope of Work

2. Take advantage of Human & Computer strengths

3. Provide Information, Not just Data / Put data in Context

4. Make Job-related Tasks Efficient

Ø Categories

Ø Display TypesØ Display Content

Ø Display StyleØ Display LayoutØ Use of ColorØ Use of Symbols &

Process ConnectionsØ Use of Text & Numbers

Ø NavigationØ Interaction with Displays

SA Stages1 2 3

;

;

;

nana

;

;

;

;

;

;

;

nana

;

;

;

nana





Scope of Work vs. Screen UseØ Typical ‘thinking’ when talking about graphics… single

screen-single display thinking

Typical Console screen use

Supporting SA & Scope of Work





Display Hierarchy—Displays, Mental Models, Navigation

Level 1Alarm Summary

Level 2Control Display



Level 3Detail Display



Level 4Support Display

Level 4Support Display

Level 1SA Trends

Level 1SA Overview

Ø The Display Hierarchy maps to console screen use to support Situation Awareness

¾ Support Proactive Monitoring with Level 1 displays that are always present

¾ Support Control, Troubleshooting, Proactive Panel rounds, with Level 2, 3 and 4 displays





Impact of Colour Schemes

How many abnormal

conditions exist in this process unit?

PI201 18.71

PSI

TC201 280.03

DEGF

FC201 4608.23

BPD

LC204 25 %

LI214 85 %

FC214 1308.29

BPD

FC234 13.29

KBPD

PC106 15.73 PSIG

PC105 18.70 PSIG

TI213 89.03

DEGF

TI223 440.06

F

AI241 320.78 AMPS

D200

T203

T200

E209 E210

C205

P211

P212 P210

P209

TI202 261.53 DEGF TI203

200.09 DEGF

LI224 95 %

NAPHTHA

VRU

FLARE

SOUR WTR

SLOP

CSR

CWS FUEL GAS

FI226 400.312 MSCFD

PI115 10.77 PSIG

FI204 000.00

MSCFD

T200 FRAC OVRHD

T200 TEMPS 1- 266.92 DEGF 2- 254.99 DEGF 3- 255.44 DEGF 4- 266.00 DEGF 5- 271.89 DEGF 6- 280.03 DEGF 7- 281.53 . DEGF

50%

100%

0% 85%

15%

65% emergency priority

high priority

low priority

emergency priority

high priority

low priority

The answer requires:

• perception to scan all of the elements on the

display and discern which are alarms

versus normal shapes (levels)

• cognition to sum them

The answer requires:

• perception to look only for coloured

outlines

• cognition to sum them





Colour PerceptionColour Vision Deficiencies

Green-Yellows & Yellow-Reds are confused

Ø Most common form is dichromatism– About 8 - 10% in males and up to 0.8% in females– Lack the red-green colour vision channel (blue-yellow intact)

Saturated Greens & Reds are confused

Normal

Dichromats





Redundant Visual Coding of Alarms

(b) Coding of alarms with colour, text and shape –more effective technique

(a) Use of shape, orientation and colour –less effective technique

! !!

Ø Presentation of alarm information is critical to the operator’s understanding of the underlying process disturbances and impacts





Potential Immediate HMI Improvements

Ø Integrated Trending¾ Dedicated overview trend screen¾ Automatic Detail trend

Ø Creation information hierarchy¾ Make important information stand-out

Ø Create display hierarchy¾ Include “at-a-glance” overview display

Ø On-screen navigation¾ Supports Situation Awareness mental model

Ø Educate Users & Display designers in human information processing






15:00

15:30

16:40


¾ Break

Ø HCI Effective Practices¾ Operator Interface¾ Alarm Management¾ Control Room & Console Design¾ Operating Procedures

Ø Wrap-up





Role for Alarms

Ø Good alarm systems signal the operator to intervene before:¾ Significant escalation from minor disturbances¾ Potentially dangerous situations escalate to

plant trip or plant incident

Ø Alarms are necessary for Operators when: ¾ Workload levels become too high to effectively

monitor all points and conditions on their own¾ Attention is focused elsewhere





Common Problems with Alarms

Ø Standing AlarmsØ Nuisance / Chattering AlarmsØ Alarm FloodsØ Alarm System Design

¾ Multiple alarms for single event





Multiple Alarms for Single Action“Belt and Suspenders”

Ø Potential Causes¾ Alarms identified from HazOp or Engineering Design ¾ Alarms added following HazOp or incident reviews

Ø Potential Effects¾ Delayed operator action

� “I’ll wait for the next alarm”¾ Alarm limits are set within

normal operating region

Ø Leading Remedies¾ Develop and apply Alarm Philosophy

� One alarm for one action¾ Alarm Rationalization





Alarm Improvement PlanningØ Goal:

¾ Minimize the size and frequency of alarm floods (principal requirement)¾ Minimize the average alarm rate (secondary requirement)¾ Provide consistent information and guidance to the operator on the

potential consequence of an abnormal event and the time available to respond

¾ Have no negative impact to running plant

Ø Multi-step approach1. Benchmark Current Alarm Performance2. Apply Immediate Improvements3. Strategic Improvements





Step 1: Benchmarking

Ø Main Activities:¾ Review alarm philosophy¾ Review DCS implementation¾ Establish current alarm performance (normal and peak alarm

rates)� Compare to industry metrics (average alarms, peak alarms, time in

upset, etc.)¾ Setup interim performance monitoring Reports





Step 2: Immediate Improvements

Ø Main Activities:¾ Apply “Alarm Tuning” settings to immediately reduce alarm

rates:� On-Delay, Off-Delay, Alarm Deadbands, PV Filter

¾ Setup weekly “Worst Actor” review meeting at the console� Address top 10 alarms from previous week

¾ Use alarm benchmarking tools to track performance changes¾ Implement Low-stress alarm sounds (Patterson Protocol)





Alarm Tuning

Ø ASM Consortium study1 showed that application of alarm tuning can significantly reduce alarm rates.

©2017 Human Centered Solutions, LLP; All rights reserved.

Plant B-1 Plant B-2 Plant D-1 Plant D-1

Ten min. peak reduction 62% 80% 25% 23%

Hourly peak reduction 34% 71% 16% 19%

Total original alarms 1320 4410 210 332

Total alarms after debounce 986 2185 180 275

1. Foslien, W (2006). ASM® Research Note on Alarm Flood Analysis. September 2006.

42




Alarm SoundsToday’s Typical Auditory DCS Alarms

Ø In multi-console control rooms, operators can have difficulty determining whose console is generating the alarm

Ø “Standard” alarms are abrupt, often loud¾ Can generate a startle-reaction and ¾ Can contribute to heightened operator stress levels

Ø Continuous, loud, high-pitched sounds can interrupt the operator’s ability¾ To communicate with other people and ¾ To maintain their concentration

Ø Example of a ‘out-of-the-box’ WAV file:





Alarm SoundsAviation’s Approach

Ø The aviation industry has developed a protocol for addressing many of these challenges (the Patterson Protocol)

Ø Uses multiple acoustic properties¾ To alert people to the urgency of the situation¾ To attract attention without being aversive to the individuals

responding an abnormal or emergency situation ¾ To easily identify the console with the alarm

Ø Example Alarm Sounds via Patterson Protocol

¾ Priority: Low High Urgent

Patterson, R. D. 1982. Guidelines for auditory warning systems on civil aircraft. (Civil Aviation Authority paper 82017). London: civil Aviation Authority.





Step 3: Strategic Improvements

Ø Main Activities:¾ Alarm Rationalization

� Preparation� Rationalization� Sign-off� Training� Implementation

¾ Implement dynamic alarming� Identified during rationalization and weekly worst actor reviews





Alarm Improvement Impact

Alarm Rationalization

& Tuning


Target: less than 1 alarm / 10 min (144 /day) per console




Potential Immediate Alarm Improvements

Ø Establish alarm performance benchmarkØ Apply “Alarm Tuning” settings

¾ On-Delay, Off-Delay, Alarm Deadbands, PV Filter

Ø Setup weekly “Worst Actor” review meeting at the console

Ø Use alarm benchmarking tools to track and report performance changes

Ø Implement Patterson’s alarm sounds






15:00

15:30

16:40


¾ Break


Ø Wrap-up





Work EnvironmentØ The work environment enhances operations team situation

awareness by supporting:¾ operator alertness¾ efficient work practices¾ collaborative interactions (including with other disciplines)¾ abnormal situation prevention and response.

� Solution Elements: ¾ Collaboration between supervision,

console and field¾ Lighting, Acoustics & Climate Ergonomics





Starting from the Console Chair

Devices:• Video Monitors

• DCS Monitors

• Large Monitors

• Standard Keyboards

• Hardwired Alarms & Switches

• Radio, Phone & Intercom

• PC Workstation

• Desktop

• Task Lighting

• Team Meeting Table

Ø Users Supported¾ Primary Operator¾ Secondary Operator¾ Engineer or Technician

� Plus Eng. Support Room





Control Suite DesignHCI / Console Design – Viewing Angles

Ø Vertical View

Ø Horizontal View

27 in.

95th

%

56.1

in.

5th %

40

.6 in

.

Eye Heights

27 in.

B

B

©2017 Human Centered Solutions, LLP; All rights reserved.




Environmental Design – Lighting Ø Ambient Light levels should be 400-600 lux to support visual acuity

¾ Task lighting required when levels are below ~300-400 lux

Ø Challenge to Ambient Light Levels with Black or White Background Displays ¾ Luminance Contrast causes eye strain & fatigue ✔¾ Occurs when ambient light is high and

background illumination is low or vice versa : :





Environmental Design – AcousticsØ Ambient noise levels should be ~50-55 dB

¾ The three largest sources of ambient noise are typically � HVAC vibrations propagating through ductwork� Control room equipment fans� Plant equipment noise

Ø ‘High activity’ noise levels should be < 60 dBØ Abatement strategies

¾ Sound-absorbing materials on surfaces ¾ Remoting CPUs outside of the control

room (also helps with Climate!)¾ Directional speakers for electronic ‘noise’ ¾ Effective arrangement of Consoles

Ø Console-specific alarm tones





Environmental Design – Climate Ø As a guide, ‘comfortable’ means—

¾ Temperature should be between 18.0 C and 22.0 C¾ Relative humidity from 40 percent to 65 percent ¾ Airflow between 0.10 and 0.15 m/s

Ø Workplace environment controls should be adjustable at each console¾ Personal heaters, vented air, …

Ø Ensure HVAC diffusers are not located directly above operator sitting positions ¾ i.e., avoid uncomfortable drafts

Ø Control room & equipment rooms should be on separately controlled zones





Potential Immediate Control Room Improvements

Ø Provide adequate number of workstations and screensØ Address control room environmental issues to support

situation awareness ¾ sound levels¾ light levels¾ vibration¾ other

Ø Provide sit-stand consoleØ Setup field operator work areaØ Manage/ minimize traffic flow around the console






15:00

15:30

16:40


¾ Break


Ø Wrap-up





Texas City 2005 – Final Report1

Ø “During the startup, operations personnel pumped flammable liquid hydrocarbons into the tower for over three hours without any liquid being removed, which was contrary to startup procedure instructions.”

Ø “Other key safety preparations listed in the startup procedures were omitted or ineffectively carried out.”

Ø “Starting, but then stopping, the unit was unusual and not covered in the startup procedures, which only addressed one continuous startup.”

Ø “In the previous five years, most of the 19 startups had deviated from written procedures.”

Ø “Management had allowed operators to make procedural changes without performing proper Management of Change (MOC) analysis, thereby encouraging operators to make unplanned (and potentially unsafe) deviations during startup”

Ø 15 Killed and 180 Injured


1. Investigation Report – Refinery Explosion and Fire. U.S. Chemical Safety and Hazard Investigation Board. Texas City, March, 2005




Use and Maintenance of Procedures

Ø Key challenges with the use and maintenance of procedures:¾ High costs associated with the maintenance of procedures.¾ Ease of access to up-to-date procedures¾ Procedures tend to be used infrequently¾ Low compliance with policy in using and referencing procedures





Procedure Clarity & Accuracy

Ø “In addition to being readily available, procedures must be clear. … If the user does not understand a procedure, or does not have confidence in its accuracy, the procedure will most likely not be used or it will be used incorrectly.”3


3. Center for Chemical Process Safety. “Guidelines for Writing Effective Operating and Maintenance Procedures”, American Institute of Chemical Engineers, 1996; ISBN 0-8169-0658-0




Potential Immediate Procedure Improvements

Ø Setup industry specific Philosophy & “Style Guide”¾ Including procedure classification to build-in compliance

Ø Train Procedure Writers, Reviewers & UsersØ Initial focus on Emergency / Startup / Shutdown

procedures






15:00

15:30

16:40


¾ Break


Ø Wrap-up


effective hci practices workshop - canadian biomass · wpac safety conference effective hci...

Documents