Civil Engineering Departmental Seminar - Time, models, choice and transient events

The unachieved gold standard for water supply systems is for them to be durable, resilient, robust, long-lived, adaptable and economical.  Many events threaten this ideal including extreme transient events – whether unusually high and/or low pressures and flows.  But these same transient events – particularly when carefully controlled – present an opportunity to extract key system characteristics.  This two-fold threat-opportunity nature often leads to notably different criteria for evaluating the performance and quality of simulation models.    To deduce system characteristics the modeller is primarily concerned with accuracy and fidelity – does the model accurately capture the essential system attributes, attributes such as pipe and device parameters or faults such as leaks and blockages?    The evaluation of threats and extreme events intrinsically puts the modeller in a different domain, one of risk management and the assessment of not only current conditions but what the future will bring.    And the future is uncertain territory.

Traditionally, transient modellers face threats by attempting to be “fail-safe” in design, that is deducing worst case responses and trying to mitigate that reality.   But that is not enough.   Our water systems do fail and the challenge is thus to be “safe-fail” when designing protection strategies, a challenge with quite different modelling requirements.  This talk will overview key alternatives to risk management and will distinguish between the modelling approaches required for assessing operational concerns with those associated with long-term capital investment and risk assessment.   One approach is fail-safe; the other must be safe-fail.