Research On Intelligent Control Strategy Of Air Conditioning Based On Deep Learning

As one of the important infrastructures in contemporary transportation,China’s high-speed railway has made unprecedented achievements in comfort and speed experience.However,with the promotion of powerful transportation countries,"economical" and "intensive" development has been put on the agenda.The road ahead of China’s high-speed railway is not only the improvement of speed and the increase of quantity,but more importantly,the implementation optimization and supporting development of stations along the route.As high-speed railway station,which undertakes the task of passenger transfer in the construction of high-speed railway.How to improve the comfortable experience of waiting in the station while saving energy consumption is an important issue to be studied in this paper.In this paper,taking X high-speed railway station as an example to construct intelligent air-conditioning control strategy based on deep learning.Firstly,the main microclimate factors affecting indoor comfort were collected through field investigation and sensing equipment,and the thermal environment of the station was modeled by Energy Plus energy consumption simulation software.The nine operating conditions were designed to simulate the building energy consumption of high-speed railway station throughout the year,and675,378 experimental data were obtained.Secondly,in order to achieve the goal of "high comfort" and "low energy consumption",designing the thermal comfort level as the decision-making basis of air conditioning in X high-speed railway station and put forward a thermal comfort level prediction model based on the improved deep forest algorithm.Also,using two deep learning algorithms and support vector machine as comparative experiments.The improved deep forest showed an accuracy performance of 0.97079,which has a good prediction.Finally,in order to verify the energy saving of the intelligent control strategy of air conditioning based on thermal comfort level,the half-open air conditioning strategy and the full-open air conditioning strategy are compared,and five algorithms and three multi-time models are used to forecast indoor temperature and humidity,energy consumption and water flow respectively.Compared with single time,GRU algorithm was selected to build short-term prediction models of environmental state under three strategies to simulate the annual operation status and energy consumption of air conditioners.The experimental results show that the intelligent air-conditioning control strategy based on deep learning proposed in this paper has an energy saving rate of nearly 35.88% compared with the full-on air-conditioning strategy,which can improve the building energy saving and comfort of X high-speed railway station.The intelligent air-conditioning control strategy proposed in this paper can provide reference and decision support for other air-conditioning control in high-speed railway station.