Research On Adaptive Air-Conditioner Control Method Based On Thermal Comfort

In modern society,people spend most of time in a day indoors.With the improvements of living standards,people require better indoor environment comfort and home intelligence.However,existing air-conditioning control systems are mainly based on constant temperature controls,set by personal experience of users.This control method faces the problem of unreasonable temperature settings.Accordingly,air conditioning control systems using this method cannot adaptively adjust temperatures in response to the changes of the external environment.Therefore,it is becoming difficult for traditional air-conditioning control systems to meet people's requirements for high-quality life of comfort,intelligence and energy-saving.With respect to this issue,this article sets thermal comfort as the controlling target of air conditioning control systems,and uses the Raspberry Pi as controlling core to realize air conditioning control system with the controlling target of thermal comfort.In control systems that take thermal comfort as the target of air conditioning control, firstly,it is needed to determine the evaluation index of thermal comfort.After analyzing a variety of thermal comfort evaluation indexes,this paper selects the predicted mean vote(PMV)with a more mature theoretical basis and a wider application,as a thermal comfort indicator for thermal comfort control.The factors that affect PMV include four environmental factors and two personal factors.The influencing factors are sequential in time.In view of this characteristic,in the algorithm,this paper selects an LSTM model that can process sequence features.In order to use historical data to make more accurate predictions on PMV values,this article improves the existing LSTM model and trains the collected PMV data sets on the model to finally obtain the PMV prediction model.Experiments show that compared with the traditional LSTM and multi-layer perceptron in PMV prediction,the improved LSTM is superior to the other two of neural networks in terms of mean square error(MSE),mean absolute error(MAE)and determination coefficient(R~2)indicators.After the modeling of thermal comfort index PMV,the PMV prediction model is transplanted to the air conditioning control system with the controlling core of the Raspberry Pi.Compared with other embedded devices and PCs,the system using the Raspberry Pi as control core has higher performance utilization,lower cost,and simpler operation.Additionally,from analysis,the direct control method of thermal comfort outperforms indirect control method.Also,the experiment shows that air temperature and gas flow rate have greater impacts on PMV.Therefore,temperature and wind speed,which are directly controlled by thermal comfort,are set as the control variables of the air conditioning control system.Taking the Raspberry Pi as the control core,this article completes the design of the data collection module,display module and setting module.Data collection mainly focuses on the measurement of the indoor environment;the function of the input module is to set the PMV value and air conditioning wind speed;the function of the display module is human-computer interaction.Experiments show that the maximum error between the predicted value of the system and the theoretical value is0.5?,which verifies the feasibility of the design scheme and lays a foundation for the subsequent practical application of the system.