Water companies certainly have no shortage of data (real time and historical) and everyone today knows that starting from data, valuable information can be obtained for the decision-making phases.
But in practice, how can operators in the Water&Wastewater sector achieve significant results? And above all, what advantages can a correct and in-depth data analysis bring to the management of an aqueduct?
Thanks to GE Digital technologies, thousands of customers around the world (more than 3.500 aqueducts on 5 continents) are now able to:
- Reduce costs
- Improve efficiency
- Improve the accuracy of water demand planning
- Reduce unplanned downtime
- Improve chemical management
Data management, the creation of models and algorithms are often considered the realm of Data Scientists, with the result that there are very few analysis projects that could bring much broader benefits for an integrated water service. So far, most technicians and operators have not yet fully exploited the analysis software, but the blame does not lie entirely with them, but rather with the complexity of the Advanced Analytics platforms.
Increase success with Analytics
Industrial Analytics is becoming more accessible for Utilities, thanks to "plug&play" solutions, such as Proficy CSense by GE Digital. Using an easily configurable user interface, technicians can combine data from disparate sources to quickly identify issues, discover the root causes, predict future performance and automate (retro)actions to continuously improve the quality and performance of the service offered.
In a recent example, GE-Digital helped a medium-sized water company a predict pump failures up to 16 days in advance, using Proficy CSense. This result was achieved swithout writing a single line of code and translated into a graph (a trend) that helps technicians identify faults.
The component identified as causing the pump failure was a critical bolt that was subject to corrosion, but was difficult to visually inspect due to its location. As the bolt eroded, its threads loosened and lost contact, allowing the impeller to wobble. The extra vibrations created by this movement would have caused further damage to the engine and its coupling. Eventually, the bolt head separated, causing the impeller to fall out of the housing and cause a catastrophic failure. The result is that this cheap bolt knocked out this very expensive pump for weeks.
It was an expensive one "point of failure" which had to be eliminated.
Through the use of analytics and a trained data model, patterns and changes in vibration signals are now monitored and future failures can be identified. The result is that the water company has two weeks to schedule preventative maintenance instead of wasting resources on unplanned downtime, resulting in an outage of just one day instead of weeks.
All this was possible without writing a single line of code and using already available historical data. The magic happened when GE Digital Water's algorithms were generated based on the information they already had.
In this case, paper maintenance records were entered manually. Analysis of these recordings identified two cases of pump failure and related vibration changes that had previously been overlooked. Using this insight, the data was cleaned and a statistical method known as “principal component analysis” was used to find the minimum number of tags needed to accurately predict failures, reducing extraneous and unnecessary data noise. There was no need for hundreds of points or dozens of data sources, you just needed to look at the useful ones from the right perspective.
This new information was then combined with already available learning models to train the algorithm for optimal, repeatable performance. The resulting data visualization easily identifies anomalies that can be incorporated into preventative maintenance planning.
Explore the possibilities with Analytics
The example of how GE Digital and this water utility were able to predict anomalies demonstrates that it is possible to pump the exact amount of water required, where it is required, according to specifications, while keeping operating costs as low as possible.
If you are collecting and storing data from your operations, the opportunities are endless. To spark your imagination, here are some use cases where analytics can have a big impact:
- Optimization of the use of chemicals (e.g. ammonia)
- Reduction of energy costs based on the use of assets in processes
- Improved water demand accuracy and flow optimization
- Equipment failure prediction
Now that you understand that analytics is at your fingertips, where do you want to take it?
For more information on this topic, watch the video: Predicting Asset Failure and Process Changes with Self-Serve Analytics.