Prediction of fouling in sewage water heat pump for predictive maintenance

Large-scale heat pumps are expected to considerably contribute to future district heating and industrial heat supply. Seawater and sewage water are among the promising heat sources for large-scale heat pumps. However, impurities in such heat sources may lead to fouling on the source side of the evaporator or source heat exchanger.

This study proposes a toolchain for predictive maintenance, including detection of fouling, predicting the further development of fouling and the expected plant behaviour. This prediction of system performance may then be used to optimize the cleaning schedule of the heat source heat exchanger. Different prediction methods for fouling prediction were tested and the effect of fouling on the heat pump performance was studied.

Finally, it was assessed how errors in the prediction of the fouling indicator influence the predicted heat pump performance. The effective thermal resistance of the fouling layer, calculated from measurement data, was used as fouling indicator. Fouling resulted in a reduction of the evaporation pressure and thereby in a reduction of the coefficient of performance and an increase of the overall power uptake as the heat output was fixed.

Fouling was found to appear irregularly for the studied case, which made the degree of fouling difficult to predict. Linear regression and radial base function predicted the further development of fouling with similar errors. Prediction errors did not have a significant influence on the predicted plant performance.

Therefore, the chosen fouling indicator as well as the prediction methods were feasible for predicting fouling in large-scale heat pumps.