Soft-sensing of liquid desiccant concentration based on ELM

This paper presents a soft-sensing method for predicting the liquid desiccant concentration based on the Extreme learning machine (ELM). The soft-sensing method utilizes a blackbox model including eight inputs variables and one output to predict the concentration in real-time and is a better alternative to manual measurement or expensive and complex sensors. The soft-sensing method is verified with the experimental data collected from the Liquid Desiccant Dehumidification System. The testing results show that the proposed method can predict liquid desiccant concentrations accurately with the errors are all within ±10%. The developed method will have wide applications in monitoring, realtime control and operational optimization of Liquid Desiccant Dehumidification Systems.