Name
Enhance process reliability of Water Treatment System via multiple Machine Learning Based Analytics - ENI
Track
Infrastructure, Water, Transmission & Distribution
Description
Water Treatment System (WTS) is necessary to treat the seawater, as well as the produced water, in order to reach the required specifications allowing the blend of both waters to be injected into the reservoir.
The objective of this paper is to present an integrated digital solution composing monitoring, predictive and advanced analytics tools that have been developed to enhance the performance and the overall availability of the WTS in one Floating Production Storage and Offloading (FPSO) unit. Exploiting the availability of data from our centralized Pi system we developed a suite of solutions including: a predictive models for fouling in Ultra Filtration units and Sulphate Reduction Package allowing the operators to predict the time left until the trains need to be cleaned; forecasting models for Produced Water Coolers heat exchanging coefficients and anomaly detection algorithm for Oil in Water sensor.