Research On Key Technologies Of Indoor Microbial Pollution Level Prediction

Indoor microorganisms have an impact on human health,and even cause serious casualties when they are involved in pathogenic microorganisms.In order to better control the level of indoor microbial pollution,online monitoring and early warning of microbial pollution is an ideal choice.However,there are the following problems.The acquisition time of routine detection results of microbial contamination is 24⁴8 hours,which requires sampling,cultivation,observation,counting and other links so that indoor microorganisms can not be monitored in real time.Existing methods,such as using UV-APS,can detect readings in real time,but they are expensive and have no specific recognition.Therefore,in order to monitor indoor microbial contamination concentration in real time and realize on-line control of indoor microbial contamination concentration on this basis,a more economical and convenient indoor microbial concentration monitoring method is needed.To this end,this study first investigated the existing indoor microbial pollution level standards at home and abroad,and determined the target microorganisms and concentration limits.Through research and development,it is found that the relevant standards in different countries are not uniform,and there are different limits for the number of bacteria,fungi or total bacteria.Because of the lack of exposure-response data,China and most countries mainly adopt research methods to determine the standards.Through research and development,it is found that indoor microbial contamination level is related to many factors.Current studies generally use temperature and humidity,particle weight concentration and counting concentration,number of people and ventilation as independent variables to establish indoor microbial concentration prediction model,the form of which is multi-linear regression model.According to the above research results,the research scheme is designed,and the basic theory of multiple linear regression analysis and non-linear regression analysis is introduced.The modeling steps,evaluation criteria and selection conditions of multiple regression model are given.On the basis of a large number of field investigation data,through data collation and theoretical analysis,it is found that the indoor bacterial and fungal concentrations in Nanjing are similar in summer,which are concentrated between 200 and 800 CFU/m³,which can provide reference for the formulation of standards in Nanjing.Indoor bacterial concentration was mainly distributed in the fifth stage（1.1-2.1μm）of Anderson sampler.Indoor and outdoor fungal concentration and outdoor bacterial concentration were mainly distributed in the fourth（2.1-3.3μm）and the fifth（1.1-2.1μm）of Anderson sampler.In the case of window closure sampling,the size distribution of indoor and outdoor bacteria is different,and there is no significant correlation between the two.It can be inferred that indoor bacteria are not affected by outdoor bacteria.The size distribution of indoor and outdoor fungi is similar to that of outdoor bacteria.It can be inferred that indoor fungi originate from outdoor.According to the steps of establishing the multiple regression model,the multiple linear regression model and the non-linear regression model are established respectively.In the process of modeling,the initial multiple linear regression model does not satisfy the homogeneity condition of variance,so the weighted least squares method is used to eliminate the problem of variance heterogeneity.In this study,the piecewise relationship between PM_2.5 concentration and indoor air bacterial concentration was found in data analysis.Combining this characteristic,a multi-nonlinear piecewise function model was established.Finally,the model is compared and validated.It is concluded that the fitting effect and prediction accuracy of the multi-nonlinear piecewise function established in this study are better than those of the multi-linear regression model.