Predictive maintenance of smart valves

Important results from the project

Somas´ long-term objective is to be able to offer its customers smart valves that provide an early indication of when they require maintenance. The project aim is to investigate and demonstrate the possibilities of using Somas´ collected sensor data for predictive maintenance. To achieve this aim, the following concrete project goals have been established: to identify parameters suitable for an AI system for predictive maintenance, to implement an AI Proof-of-concept with collected data and to create a roadmap for the future work. All three goals are considered to have been achieved.

Expected long term effects

One result of the project is a web-based tool used to analyze large amounts of data. Another result is a first valve health measure, a model which estimates the status of a number of health indicators with which the valve remaining lifetime is predicted. The valve health measure will be further developed and used so that proactive measures can be implemented. The customer can thus increase its valve availability and productivity. Another result is that the planning of future work in a larger perspective has been based on the roadmap developed within the project.

Approach and implementation

The project has been run as a sub-project in a larger pilot project where data collection, system development, business model development also were included. Somas has participated with product expertise and valve life cycle experience and AFRY has participated as AI experts on a consulting basis. The project´s objective was followed up and refined during the project. Desired project deliverables were defined early and facilitated prioritization of activities at project completion.