Optimization Of Central Air Conditioning Cold Source System Based On Data Learning Model

The large air-conditioning cold source system operates under partial load chronically and possesses energy saving space by optimizing operation,while large air-conditioning systems are mostly operated according to rated conditions.In the existing energy-saving research of cold-source systems,cold source system model is often applied to optimize system operation,but traditional air-conditioning system model need some unobtainable equipment structural parameters,and the model can’t reflect the attenuation of equipment performance with the running time.Meanwhile,most system operation optimizations are limited to univariate or partial system optimizations.In order to optimize air-conditioning system globally during the whole life and guide the system to energy-saving operation,this paper constructs a central airconditioning cold source system model based on data learning theory and research the global optimization of central air-conditioning cold source system for saving the system operation consumption.The BP neural network model and parameter identification model of central air conditioning cold source system are respectively established to predict the system operation energy consumption based on data learning theory in this paper,and models are compared in the prediction accuracy of cold source system operation energy consumption.Considering the different accuracy prediction when these two models predict the system operation energy consumption with different model input parameters,the hybrid model is proposed to combine the high accuracy prediction interval of BP neural network model and parameter identification model by applying the K-means clustering method.Compared with BP neural network model and parameter identification model,the cold source system operation energy consumption prediction accuracy of hybrid model increases by about 30%,and the system operation energy consumption prediction relative error of hybrid model is within 8%.Considering the current situation that operation adjustments of cold source system equipment are mostly limited to start-stop regulation,the optimization adjustment strategies and operational influencing factors of multiple chillers,pumps and cooling towers in parallel operation are researched based on the system model in this paper.The optimization adjustment strategy is proposed to optimize the load distribution of multiple chillers in parallel operation by applying genetic algorithm.According to the frequency conversion operation principle of pumps,optimization adjustment strategy for parallel operation of multiple pumps is obtained based on the limit of frequency conversion interval and pump rated power,and the average energy saving rate under optimization adjustment strategy is about 53.5%.By analyzing the energy consumption of parallel cooling of multiple cooling towers under the same heat exchange amount and cooling tower efficiency,optimization adjustment strategy for parallel operation of multiple towers is obtained based on the limit of frequency conversion interval of fan,and the saving energy increases with the number of parallel cooling towers with meeting the frequency conversion interval of fan.The optimization objective function of the central air-conditioning cold source system is established for seeking the lowest cold source system operation energy consumption,and optimization variables of objective function are the chilled water supply temperature,the chilled water supply and back temperature difference,the cooling water inlet temperature and the cooling water inlet and outlet temperature difference.With the restrictions of safe operation and energy conservation,the cold source system operating parameters are optimized globally under different system load rates,and the variation trend of optimized system operating parameters are analyzed with different system load rates.Different cold source system operation models are compared with global optimization operation mode in the cold source operation energy consumption.There are three kind of cold source system operation model,such as the fixed flow operation mode,the chilled water side optimization operation mode and the cooling water side optimization operation mode.Compared with the cold source operation energy consumption in above three operation modes,the energy consumption in the global optimization operation mode averagely decreases by 14.10%,4.74% and 7.62% when the cold source system load rate is between 40% and 100%.Considering the actual central airconditioning cold source system of a subway station,this paper applies the above global optimization method to simulate the system optimization operation energy consumption at a cooling day,the energy saving rate after global optimization of the cold source system is about 13.2% compared with the actual operation energy consumption of the cold source system.