Research on load excitation identification method of multi-connected air conditioning compressor based on RBF network with multi-strategy fusion SSA

The load excitation of the multi-connected air conditioning compressor is the major source of vibration in the entire air conditioning pipeline system, which imposes a direct impact on the vibration noise and reliability of the overall system. Accurate identification of the compressor's load excitation is a primary condition to ensure its normal operation and is an essential part of the comprehensive testing and evaluation of the system's dynamic performance. In this paper, aiming at the problem of the inability to directly identify the time-varying load excitation of the multi-circuit air conditioning compressor, a predictive model is built on the basis of the Radial Basis Function (RBF) Neural Network. Meanwhile, to ameliorate the overall network model’s identification performance, a Multi-Strategy Fused Sparrow Search Algorithm (ISSA) is adopted for optimization regarding the center, the basic functions’ variance, and the weights from the hidden layer to the input one in RBF. This algorithm not only improves the joiners, discoverers, and scouts positioning in the Sparrow Search Algorithm (SSA) but also adds the Firefly Algorithm (FA) to disturb the positions of all sparrows, thereby preventing the algorithm from being trapped in local optima and enhancing the global search ability. The results of practical testing and simulation experiments show that the Firefly Algorithm improved Multi-Strategy Fused Sparrow Search Algorithm-Radial Basis Function (FAISSA-RBF) algorithm reduces the convergence to below 10^–2, and the R² is improved by approximately 12.7% compared to the SSA-RBF. This ensures the normal and stable operation of the air conditioning compressor system under time-varying conditions and the effective identification of load excitation.