Research On Indoor Lighting Quality Evaluation Methods Based On C-Support Vector Classification Algorithm

Although currently lighting energy efficiency has become one of the factors that must be considered in building and interior design with the progress of urbanization in the world,lighting energy saving should not be at the cost of the lighting quality,which means the reduction of building energy consumption without significantly reducing the users' comfort has become the basic principle of building energy efficiency design nowadays.Therefore,how to balance energy consumption and lighting quality becomes the key to non-domestic building interior lighting design,and it is necessary to explore the mapping relationships between illuminant parameters and lighting quality in different types of interior space,which could be the bottom line of lighting energy saving and benefits to the overall revenue of lighting design.Based on the above research and the project of the National Key Research and Development Program of China [Grant No.2016YFB0601700],this study firstly proposed the starting point under the premise of summarizing relevant literature's results on lighting quality,lighting energy efficiency and ways to balance these two goals: The Kruithof's graph should be revised using C-support vector classification(CSVC)algorithm to find the classification boundaries of different visual comfort levels according to certain indoor scene whose visualization results can be used as a reference to balance lighting quality and lighting energy efficiency.On this basis,the results of C-SVC modelling on existing subjective evaluation experiment data of lighting comfort in simulated production environment indicated the feasibility of this method,after which the experimental operation processes and the commonly used methods of data preprocessing and mining with Python to similar experiments were summarized clearly.Next,this method was comprehensively used in the modeling of classroom lighting quality evaluation,which includes following steps:(1)A field survey and an online survey were carried out to find the existing problems of light-environments of blackboard reading,desktop reading and PPT reading in classrooms and provide references to the simulated scene of subjective evaluation experiments;(2)After establishing medium-sized classroom simulation space in the full-size lightenvironment simulation comprehensive laboratory,we conducted a subjective evaluation experiment of classroom lighting visual comfort and collected the experimental data;(3)C-SVC algorithm was used in the modeling on evaluation data of blackboard reading comfort,desktop reading comfort and PPT reading comfort respectively.The classification boundaries of different levels of visual comfort(least comfortable,less comfortable,more comfortable,most comfortable)and results of model verification demonstrated the good performance and practical value of the models.Finally,the research contents,innovation points and further research directions of this paper were summarized,which provides ideas for the subsequent research on lighting quality and lighting energy efficiency.The research result shows that the optimization method based on C-SVC algorithm to the balance between lighting quality and lighting energy efficiency is feasible,and the visualization results of model classification boundaries are convenient for practical use,which is of certain reference value for architectural interior lighting design.