Analysis And Diagnosis Model Of Energy Consumption In Hospital Building

As a typical large-scale public building,the analysis and research of energy consumption data of medical building is the basic work of building energy conservation.Aiming at the disconnection between macro data index method and micro data fine modeling in traditional research,and the contradiction between high cost and low density data modeling of fine energy consumption monitoring system,the macro data and micro data are defined.In this paper,a model framework of combining macro and micro data is proposed.An evaluation model is established to make sure the energy consumption problem subsystem based on the index set of macro data,and then a machine learning diagnosis model is established based on the operation data and energy consumption data of the subsystem.Diagnostic problem variables.After collecting Agna B,The monthly energy consumption data of C three hospitals and the heating and cooling system data of B Hospital were studied by a case study.It was found that as a green hospital,1)B hospital did not show its due advantages of energy saving,and was analyzed by index set.High heating energy consumption is the main cause of high total energy consumption in B hospital.2)A diagnostic standard model for machine learning energy consumption of boiler hot water heating system is established in view of B hospital heating system.First,the operation data with better energy saving characteristics are selected from the collected diagnostic data by clustering or classification algorithm,and then the energy consumption diagnosis model based on regression model is established to diagnose the diagnosed data.The diagnosis found that the primary heating terminal was lacking in the corresponding condition.The energy saving control measures are the main cause of high heating energy consumption in B hospital.3)Based on the diagnostic standard model of energy consumption,the advantages of clustering algorithm in the process of energy saving data screening of standard model are demonstrated through theoretical analysis and diagnosis results.TheQPSO optimized K-METHOIDS algorithm is proposed to improve the clustering process,and the clustering results are superior to those of the usual clustering algorithms,and the effects of PCA dimensionality reduction technique and the number of hidden layers of neural networks on the fitting accuracy and generalization of the regression model are discussed.The regression model is improved by using the LS-SVM algorithm optimized by GA-PSO,and the fitting precision is improved by needles.4)For B hospital central air conditioning system,a diagnosis method of central air conditioning energy saving is put forward.The ANN static model is established by using the operation data under the state of energy saving control,and the data of closing energy saving control are forecasted and compared.It is found by analysis that the energy saving control measures of central air conditioning in Hospital B can save energy 7.6.5)The data verify that the analysis results are correct.Since the diagnosis of energy saving of central air conditioning is essentially a forecasting analysis of cooling load,a NARX-ARMA air conditioning load forecasting model is put forward.Compared with the NARX model,the data test has higher fitting accuracy,higher adaptability and higher performance / price ratio than ANN.6)According to the merits and demerits of ANN and NARX models,this paper presents a dynamic static test model for energy saving characteristics of central air conditioners,which uses NARX to screen the data initially,and then uses ANN to verify the data.It is found that the combined detection rate is 73.6% and 66% respectively.Considering the data quality and model precision,the model has a high testing accuracy and can be used in the energy saving diagnosis of central air conditioning.The combination of macro data and micro data is the advantage of low cost qualitative characteristics of macro data analysis and high precision quantitative analysis of micro data.The actual case data test can be well used in the analysis and research of energy consumption data of hospital buildings.