Parameter Optimization of an Absorption Heat Exchanger with Large Temperature Difference

The absorption heat exchanger with a large temperature difference has a higher heat transfer superiority than the other heat exchangers (including plate heat exchanger), which is more suitable for long-distance heating. To improve its system performance, parameter collaborative optimization (including building accurate predictive models) has become an effective method because it does not require too much investment. In this study, a heat exchange station was chosen as a case study, and a model of a long short-term memory (LSTM) neural network was used to predict the temperatures of primary return water and secondary return water. Accordingly, the reliability of the fitting result based on the model was confirmed through a contrastive analysis with the prediction results of a support vector machine (SVM) model, a random forest (RF) model, and an extreme gradient boosting (XGBoost) model. In addition, the algorithm of particle swarm optimization was used to optimize the flow rate of primary supply water. The results showed that the temperature of primary-side return water decreased from 29.6 °C to 28.2 °C, the temperature of secondary-side return water decreased from 39.8 °C to 38.6 °C, and the flow rate of primary-side supply water decreased from 39 t/h to 35.2 t/h after the optimization of the flow rate of primary supply water. The sensibility assessment emerged that the secondary-side flow rate to the secondary-side supply water temperature was about 7 times more sensitive than the primary-side supply water temperature, and concretely, the lower the temperature, the higher the sensibility. In summary, the accuracy of the proposed prediction model was validated and the optimization direction was pointed out, which can be used to provide guidance for designing and planning absorption heat exchange stations with large temperature differences.