What Dynamic Simulation Can Add to Water Utility Risk ... · What Dynamic Simulation Can Add to Water Utility Risk Assessment Mohamed A. Hamouda, Jared Best, William B. Anderson,

What Dynamic Simulation Can Add to Water Utility Risk Assessment

Mohamed A. Hamouda, Jared Best, William B. Anderson, and Peter M. Huck

2014 OWWA/OMWA Joint Annual Conference &

OWWEA Trade Show May 4th - 7th, 2014

Ensuring Product Quality

Product Quality Control (QC) monitors compliance with standards • QC tells us that something has gone wrong after

it has happened

Process Quality Assurance (QA) uses risk

management • QA tries to stop that something before it goes

wrong

2

Preparation1. Assemble the WSP Team

System Assessment2. Describe the water supply system3. Identify hazards, hazardous events, and assess risks4. Determine and validate control measures, reassess and prioritize risks5. Develop, implement, and maintain an improvement / upgrade plan

Operational Monitoring6. Define monitoring of the control measures

7. Verify the effectiveness of the WSP

Management and Communication8. Prepare management procedures9. Develop supporting programs

UpgradeInvestment required for major system modifications

Feedback11. Revise the WSP following an incident

10. Plan and carryout periodic review of the WSP

Incident (emergency)

Water Safety Plan

3

Risk Assessment: State-of-the-art

Define Scope

• Know the system: flow, quality, and operating parameters • Define failure: quality, quantity, economic, environmental impact,

and/or customer trust

Identify Hazards

• An existing or possible/probable event that could cause failure • Methods: FTA, FMEA, Checklists, HAZOP, HACCP (e.g. CWWA) etc.

Estimate Risk

• Qualitative: ordinal scale likelihood x consequence (e.g. AB DWSP) • Quantitative: stochastic/probabilistic analysis

Evaluate Risk

• Define thresholds or tolerability criteria, or ranking method • Identify risk reduction methods and evaluate/rank them

4

Score N/A Insignificant Minor Moderate Severe Catastrophic

N/A 0 1 2 4 8 16 Most Unlikely 1 1 2 4 8 16

Unlikely 2 2 4 8 16 32 Medium 4 4 8 16 32 64 Probable 8 8 16 32 64 128

Almost Certain 16 16 32 64 128 256

Qualitative Risk Assessment

A

D

E

C

B

5

Acceptable Risk ALARP (As Low As Reasonably Possible)

Unacceptable Risk

Identify and rank target events causing unacceptable risks

Identify and rank risk-reduction options

Quantitative Risk Assessment: e.g. QMRA

QMRA

6

1. Hazard Assessment: System description, indicators, failures

2. Exposure Assessment: Pathogen concentrations (% infectious),

Treatment (Log reduction), Consumption (L/d)

3. Dose-response Analysis: Impact of Exposure (Pinf, Pill,

#Illnesses/Yr, DALY*)

4. Risk Characterization and Alleviation: Compare to

acceptable risk levels and outline risk reduction measures

Risk Assessment

Qualitative Quantitative

Relatively easy and fast Cumbersome

Easy to communicate, decision is clear, common

Easily comparable

“Comprehensive” “Rigorous” and specific

Abstracting/obscuring data

Data input is explicit

High subjectivity, overlooking combined effects

Less subjective and transparent (for experts)

7

Dynamic Simulation

In dynamic simulation models, events occur sequentially over time. Specialized software is required

Time series data are used explicitly and the output is also displayed in time series

In a dynamic simulation, stochastic variables may be discrete or continuous

Different from static simulation (simple Monte Carlo)

8

Dynamic Simulation for Risk Assessment

Objectives: Review existing water treatment dynamic

simulation platforms Select a water treatment simulation platform to

be used in conducting risk assessment Outline changes to the risk assessment

framework resulting from the use of dynamic simulation

Demonstrate the use of dynamic simulation in hazard identification

9

Dynamic Simulation: State-of-the-art

Water treatment simulators have existed since the 1990s: • OTTER (WRc) • Stimela (TU-Delft) • Metrex (TU-Duisburg) • WTP (US EPA) • WatPro (Hydromantis)

10

Dynamic Simulation: State-of-the-art

However they have had poor uptake • Design and performance evaluation is mainly

driven by pilot trials and rules of thumb. • Extensive data required for calibration, and

models have restricted validity outside calibration range, why?

• Focus on building an accurate simulator rather than tailoring it to a particular use.

• One more thing to learn! Bugs are frustrating!

11

SimEAU

SimEAU was developed as part of the EU Fifth Framework project, TECHNEAU

To replace WRc’s OTTER & TU Delft’s Stimela Greater emphasis on calibration using routine

data, and theory based process models Modelling process dependence on downstream or

upstream processes and consequences on performance

Open framework designed to simplify adding new water treatment process models

12

How SimEAU Compares?

Process METREX STIMELA OTTER SimEAU Coagulation + Flocculation

Combined Sedimentation Rapid Filtration Biological Filtration Ozonation Adsorption Softening Chlorination Membrane Filtration

13

Risk Assessment Framework + Dynamic Simulation

Define Scope

• - Limitations set by data and simulation model

Identify Hazards

• + Simulations help identify hazards • + Hazard impact scenarios (what if…)

Estimate Risk

• + Accepts discrete and continuous parameters • - To be comprehensive dramatically increases complexity

Evaluate Risk

• + Probability of failure: non compliance, health risk, etc. • + Risk alleviation measures scenarios

14

SimEAU Demonstration : Hazard Identification Baseline simulation

15

SimEAU Demonstration : Hazard Identification

16

SimEAU Demonstration : Hazard Identification

17

SimEAU Demonstration : Hazard Identification

Filter to waste

18

Filter to waste

SimEAU Demonstration : Hazard Identification Hazard what-if scenarios: e.g. flocs of lower mass

density increased filter-to-waste period

Filter to waste19

Filter to waste

Benefits of Dynamic Simulation

More explicitly represent the timing and sequencing of events, their impact, and response

Data is used rather than assuming distributions Directly calculate the impact of variations of raw

water quality, operating parameters and process performance on the plant model

Capable of capturing complex interdependencies System success criteria is more realistic as opposed

to conservative criteria used in less rigorous assessments

Risk alleviation measures can be directly simulated

20

Conclusions

Dynamic simulation can potentially enhance risk assessment by introducing transparent, explicit, and real use of raw data

Water simulators can be tailored for use in risk

assessment, to increase uptake

Dynamic simulation requires extensive data, and can help shape the future of data collection and monitoring in WTP

21

Future Work

Developing simulation models to assess the risk of producing non-compliant water for certain contaminants of concern

Integrating QMRA in SimEAU, the analysis will then rely on simulated performance of the treatment plant rather than estimating performance based on previous studies

Exploring other uses of dynamic simulation

22

Acknowledgements

Special thanks to Jeremy Dudley, WRc UK, for providing a copy of SimEAU and for answering questions regarding the use of SimEAU and identified bugs in the program.

23

Thank you to our NSERC Chair partners

24