IOA – Implementation of Automation
The aim of the project is to implement on-line simulation to district heating net and boilers where the difference between simulated data and measured is used as input to a BN.
Description of the project
The aim of the project is to implement on-line simulation to district heating net and boilers where the difference between simulated data and measured is used as input to a BN. The deviation trends is then used to determine different process disturbances and sensor faults as a probability for something being correct or faulty. The information shall be presented to operators and process engineers in an easy to understand way. The results are discussed on a regular basis with operators, maintenance staff and process engineers to get feed-back on actual faults.
Objectives and justification of novelty
The main objective of the project is to implement what has been run during a few weeks earlier on-line at Malarenergi for boiler 5. As a first step the system will be used for the district heating grid to compare existing load models to also include measurement of temperature in actual buildings, to determine how well sub stations are working as well. This will later be followed by implementation on boilers in both Västerås and Eskilstuna.
The system also will be improved so that we don´t need to operate the simulator through a Matlab-SIMULINK, excel interface, but can work directly between Modelica simulation model and the process data base. This has been developed in the STREAM project in co-operation with Linköping Unirversity. Also Hugin BN will be connected directly to the data base.
The novelty is primarily to get a system that is robust enough to actually be operational not only for a day, but continuously, so that both Malarenergi and Eskilstuna Strängnäs Energi och Miljö get useful outcome of the project.
Input to the simulator first will be temperature, flows and pressures for the district heating system for one area in Västerås and one in Hallstahammar, but also will be NIR measurements that already is installed in Eskilstuna, and on its way at Malarenergi boiler 6. To measure both moisture and biomass composition with respect to lignin and HHV, plastic like PVC in boiler 6 etc predictions also will be made on emissions like CO, NOx and dioxine.
The focus of the project is to actually get the system running in a robust way but also to get a good working procedure involving process operators. To do this we plan frequent follow up and discussion meetings at the power plants. At the agenda we then will discuss both what we get out of the system but also what process problems there have been recently, that would be of interest to include in the BN. In this way we will develop the decision support system (BN) continuously with the purpose to create a “learnings system”. This will also include discussions around different specific measurements and signals to sort “normal signals” from “faulty signals” in an automatic way.