Water Journal : Water Journal December 2012
intelligent water networks refereed paper technical features 102 DECEMBER 2012 water Efficiency and network optimisation has long been a challenge for the power industry and can provide a useful starting point to look for methods that provide tangible outcomes. The integration of large numbers of different types of metering from the customer level to generation monitoring provides operators with the ability to optimise and find efficiencies using automated software analysis technologies. This ‘Smart Grid’ concept extends beyond the power industry to other industries where real-time analysis and understanding of a complete network’s operation is paramount. While there are some technical similarities between utilities, water networks do not face the same challenges of requiring second-to-second management and optimisation as seen in electrical networks. Water planning and optimisation tends to be long term, and based on ensuring requirements meet key peak demand conditions, as opposed to finely tuned minute-to-minute optimisation. The much slower transport speeds and ability to store the product means that there has been no historical need for fast-paced data tracking systems. The challenge in ensuring consistently calibrated and accurate metering technologies has also limited the extent of using the information to make such tight decisions. The rollout of individual ‘smart meters’ to individual customers, which provide information directly back to the operators, has already begun in the power space, and ultimately this will be the reality in water networks in the medium to long term (Marney & Sharma, 2012). With all this additional information the challenge is to use it intelligently. Combine this with more reliably accurate network meters and advancements in modelling technologies, and the technical foundations to build IWN solutions are well established. The layers of Technologies at a Water Utility In the water industry, metering and instrumentation technology has been a critical area of development. While the development of meters is quite mature, the interconnectivity of the devices has greatly increased in recent years. This connectivity presents opportunities for improvements in using and analysing this data. Ideally instrumentation is both accurate and connects to telemetry or other systems easily to pass along this information. Low-cost and easily installed meters are readily available, and these provide a base line for any intelligent water networks system. Yarra Valley Water employs a variety of metering device types connected to its SCADA system, from its permanent distribution zone metering to temporary loggers for hydraulic modelling. Remote acoustic logging devices have also been trialled by Yarra Valley Water in cooperation with the Victorian Department of Sustainability and Environment. This trial was to investigate capabilities of analytical software, but the analysis software was segregated from SCADA and thus the other metering technologies. A primary lesson for Yarra Valley Water from this was the importance of unifying available information sources. Isolated sources and technologies that require their own software have costs associated with set-up, retraining and changeover time between them. Instrumentation sources are only as good as the overview and aggregation system set atop them. Modern SCADA systems are suitably advanced, and the packages or add-ons that connect to them (such as OSI-PI) have a well- understood benefit. These systems take incoming instrumentation information and present alarms based on thresholds or pre-configured rules for an operator to respond to. The more information that can be integrated into these systems, the better a system can be monitored and optimised. There are, however, limitations in the current analysis systems, as at their core they still present instrumentation information for operators to interpret. The systems themselves are not able to directly understand or interpret the behaviour. The ability to look forward and predict system behaviour with hydraulic models presents another area of opportunity. Systems from MWSoft, Bentley and Aquis provide the ability to model and predict the behaviour of water networks. This modelling can predict system response to valve configuration changes or new infrastructure and identify optimisation approaches and points (Wu, 2008). However, the challenge remains to have accurate models that represent the realities of the network. Calibration testing is periodically used to verify hydraulic models, and some modelling systems are now developing methods of integrating directly into SCADA to automatically calibrate the model. Ultimately, however, the challenge is still to have information that ‘makes sense’ and presents useful advice or comments that can be acted upon. Industry progress and leaders The next stage is to combine the data from various information sources and make the most of the improvements in the technology layers. The Smart Water Networks forum (SWAN) highlights the benefits of what it refers to as a layer of Data Fusion and analysis (Smart Water Networks Forum, 2012). Technologies in this space are developing quickly, but the field is by no means mature. Broadly speaking there are a few major players such as IBM, Siemens and TaKaDu that are developing solutions to fill this gap. The principle, as outlined by SWAN, is to synthesise raw inputs from a variety of sources for analysis and then enable either automatic response or produce actionable advice to a human operator. The benefit here is nothing new; it is simply performing the work that an experienced operator would do. The difference is the speed and number Figure 2. ‘Inbox style’ event view. Incoming events are shown with a classification and a brief description.
Water Journal February 2013
Water Journal November 2012-1