GSP serves over 200.000 inhabitants in the Province of Belluno. GSP operates 6.500 km of drinking water pipes and sewers. GSP has embraced the ambition to reduce the non-revenue-water ration with 35% in three years’ time. This goal will be realised by creating DMAs, by increasing the amount of water meters, by reducing pressure and by replacing the worst pipes in the drinking water distribution networ
Approach and solution
To identify the pipes in the network with the highest chance of failure, Spatial Insight applied risk model SI-Rehab that includes probability models SI-Cluster and SI-Regression. Spatial Insight (SI) has received data from GSP, and SI has analyzed the data and concluded that sufficient data of sufficient quality was available for the first runs with SI-Regression, SI-Cluster and SI-Rehab.
SI-Cluster, the analysis that cluster groups of pipes in a smart way, has identified areas in the network that have a high impact on the performance. To align with the project size and available budget of GSP the cluster method has been extended with SI-Hotspot. Within each cluster the area is identified where the most leakages occurred. By doing so the 0,6% of the network has been identified that will cause 9% of the leakages, and the 5% of the network that will cause 48% of the leakages.
SI-Regression predicts the performance of the network over time. This analysis heavily depends on the presence of the year of installation for each pipe in the network. SI has explored to work with an ‘best guess average year of construction’ for the pipes with lacking year of construction. The results made good sense, and therefore Spatial Insight is confident that these regression curves can be applied successfully.
The regression curves will be used in SI-Rehab and form a base for the implementation of BestNet.
For SI-Rehab the chance of failure is combined with the effect of failure which results in the network rehabilitation score. The calculation for effect is split into internal and external effect. The internal effect assigns a value to the consequences for the network performance of an outage at a certain location. The calculation relies upon a network analysis. Spatial Insight has made an additional filtering to reduce the total length of highest priority pipes to meet the available pipe replacement budget of GSP. Within the worst clusters, Spatial Insight has selected the parts that have shown most leakages. As published in the npj Clean water journal, historic leakages are the most important predictor of new leakages.
Spatial Insight has calculated the internal and effect score for pipe segments of circa 10m. The internal effect has been calculated by applying SI-Network, that determines which sections will be affected if leakages occurs and a section needs to be closed, and the impact if during the closing of a section, a valve could not be closed.
Contribution to the organisation’s strategy
Spatial Insight managed to identify the 0,6% of the network that he results of the calculations allow GSP to invest the available budget for pipe replacement in the way that will be most significantly contribute to the reduction of the current water losses. At the same time, the calculations reveal the value and importance of asset data and leakage registration. With better and more complete data, the results will be even more detailed and useful.
Customer review
Marco Bacchin, General Director of GSP: “The analysis is important for our company. We need accountable modelling results for our investment plan. Spatial Insight delivered useful results in a short period of time and identified data improvement opportunities in a very detailed way. We have also appreciated the short delivery time, the relevance of the results, and the room for further improvements as well.”
CONTACT
"We have integrated SI-Regression and SI-Cluster in our ultimate risk model. Let's apply SI-Rehab on your data."