Experimental study on the performance consistency diagnosis of split-type air conditioner based double colour ball matching rule.

With the continuous improvement of consumer product quality and the increasing demand for energy conservation, the quality requirements for air conditioning products have evolved beyond merely meeting energy efficiency standards. There are now stringent demands for performance consistency among products of the same model but from different batches. However, in the actual sampling process, there are various sources of performance deviation. Therefore, there is an urgent need for a diagnostic mechanism and method to provide personalized optimization strategies for machines exhibiting performance deviations, to restore their annual energy consumption efficiency (APF) levels to normal. Based on comparative experiments, a database consisting of 6 sets of APF deviation prototypes and 18 sets of internal and external machine combination samples is constructed. A personalized diagnosis system for performance consistency issues of split air conditioners is established by combining the two colour ball parameter matching strategy and BP neural network algorithm. Twelve relevant parameters are selected and the deviation level value is set as the diagnostic basis for the expert system. After comprehensive working conditions, the diagnostic problems and problem weight proportion of the prototype are given, and the component optimization strategy for the prototype is proposed. The verification experiment shows that after optimizing the corresponding components, the APF of the two prototypes increased by 1.2 % and 1.3 % respectively, reaching the APF qualification standard. And the optimization results verify the technical feasibility of the expert system.