The setting: A cliff top hotel overlooking Torquay. The challenge: To optimise a water distribution network, using a “Serious Game”. The result: We aren’t as good as hamsters, but that’s because we don’t know how to cheat.
I should probably explain. Every year, each of the Stream universities takes its turn at hosting Challenge Week. This year was the turn of Exeter to facilitate, what is essentially, an opportunity for members of all Stream cohorts to get together and forge a strong network. This is usually done through a mixture of activities, from the more technical serious game challenge to riding around Haldon Forest on segways, to scrambling down coastal paths and throwing ourselves into freezing cold seawater. It all does the job! I won’t go too much further, as my fellow Stream-er Sarah Cotterill has done an excellent job on her blog, so check that out if you want a better run down of the week and some photos.
I thought I ought to focus my blog on to the topic of optimising water distribution systems, this being something I have spent a large chunk of time looking into; from reading academic papers to talking with the teams responsible for this within my sponsor company. My main thoughts from the challenge were “if only it were this easy!” Having said that, even with this simplified gaming challenge there were 15316 possible optimal solutions….which is a lot! In the real world though, there are many complicating factors. Water suppliers are faced by restraints on time and money – in the game we could just click on a pipe and change it (replace, enlarge, reduce, remove), but in reality there are a number of stages to get through.
- Firstly, an extensive planning stage to determine the costs and benefits, and identify risks (such as loss of supply and discoloration.)
- Justification of why a particular scheme should be prioritised over another.
- Approval must be given for the scheme.
- Finally, time and resources must be allocated to complete the physical work.
This is largely why pipe replacement schemes are fairly slow, whilst replacement rates vary, imagine a replacement rate of around 500km per year (this is fairly generous – it’s probably more like 100km per year) when you have a total network length of upwards of 40000km, it would take 80years to fully replace every single pipe. The replacement rate is usually a lot lower than that, and the first pipes you replaced will be reaching the end of their lifespan by the time you have finished that first 80 year cycle. I don’t know about you but that gives me a headache just thinking about it!
Since replacing pipes, takes time and is fairly costly water suppliers often look at other options for optimising the performance of the network assets they have. Leakage may be reduced through pressure management schemes rather than replacing pipes. By dropping the pressure down to a lower level (whilst making sure that every property meets a minimum pressure requirement); background leakage levels are reduced and burst rates are lowered. Pumps and other fittings may be replaced and controlled more effectively to smooth the flow of water in the network as well. This all puts less stress on the assets and can extend asset lifetimes.
I’ll save water age and resilience for another day. I’ll just say, resilience requires building in loops, loops add complexity to a network, and complexity usually results in higher water ages. Sound like a multiple objective optimisation problem to me, OUCH!
Finally, I will share my thoughts on not having a sandcastle building competition though…. absolute disappointment.