Research On Data-driven Cooling Load Prediction And Energy Saving Optimization Method For Central Air Conditioning

With the continuous improvement of the national economy and people’s living standards,the application of central air conditioning in commercial buildings and large residential buildings is becoming increasingly widespread.While creating a more comfortable and healthy living environment for people,it also brings about huge energy consumption issues.Therefore,the research on energy conservation in central air conditioning is of great significance.As it can not only reduce building operating costs,but also reduce energy consumption and carbon emissions,thereby promoting environmental protection and sustainable development.This dissertation mainly conducted the following work:(1)Aiming at the problems of insufficient feature extraction,low prediction accuracy,and sensitivity to noise in central air conditioning cooling load prediction,a deep learning hybrid model named WTD-CNN-LSTM is proposed to achieve accurate prediction of cooling load,which combines wavelet threshold denoising(WTD),convolutional neural network(CNN),and long short-term memory network(LSTM).By evaluating the effectiveness of the model,the coefficient of determination R~2 of the proposed prediction model is 0.9365,which is 0.0221-0.0572 higher than other comparison models.The results show that the proposed prediction model can perform nonlinear modeling of cold load and has good prediction accuracy.(2)Aiming at the problems of slow convergence speed,insufficient optimization accuracy,and easy to fall into local optima in the sparrow search algorithm,a multi-strategy improved sparrow search algorithm(MISSA)is proposed by integrating multiple strategies.Firstly,the Circle chaos initialization population is adopted to lay the foundation for global optimization.Then,the golden sine cosine strategy is introduced in the discoverer position update to enhance the local development and global exploration capabilities of algorithm,and improve the convergence ability of algorithm.Finally,the Cauchy mutation strategy is introduced in the joiner position to improve the algorithm’s ability to jump out of local optima.The improved sparrow search algorithm is verified by benchmark test functions in terms of convergence speed and optimization accuracy.The results show that the improved sparrow search algorithm has significantly enhanced convergence speed and optimization accuracy,and exhibits good robustness.(3)Aiming at the high energy consumption caused by the nonlinear,large lag,and strong coupling characteristics of the central air conditioning water system,an energy-saving optimization method based on WTD-CNN-LSTM and MISSA is proposed.Firstly,the energy consumption of the chiller,chilled water pump,cooling water pump,and cooling tower are analyzed,and their energy consumption models are established.Based on historical operation data,the Levenberg-Marquardt algorithm is used to identify unknown parameters.Then,an energy-saving optimization model of the central air conditioning water system is established.By analyzing the constraints of each equipment in the central air conditioning system and using the cold load prediction results of WTD-CNN-LSTM as one of the constraints,the air conditioning operation is adjusted in advance according to the cold load demand in the future time to achieve on-demand cooling and energy-saving.Finally,the MISSA optimization algorithm is applied to the actual air conditioning system to optimize the system operation parameters,and has better results compared to other optimization algorithms.