There are some problems in the shifting process of hydraulic CVT, such as irregularity, low stability and high failure rate. In this paper, the BP neural network and convolutional neural network are used for fault diagnosis of the HMCVT hydraulic system. Firstly, through experiments, 120 groups of pressure and flow data under normal and four typical fault modes were obtained and preprocessed; they were divided into 80 groups of training samples and 40 groups of test samples via random extraction, using the BP neural network model and convolutional neural network model for fault classification. The results show that compared with BP, PSO-BP and other models, the fault diagnosis rate of the BAS-BP neural network model can reach 92.5%, and the average diagnosis accuracy rate of the convolutional neural network can reach 97.5%, which can be effectively applied to the fault diagnosis of the HMCVT hydraulic system and provide some reference for the shifting reliability of hydraulic CVT.

To improve the driving comfort of continuously variable transmission (CVT) tractors, the shifting quality of hydrostatic power-split transmission with a standard planetary gear set was optimized. Firstly, the powertrain of the CVT and two shift strategies, direct-shift and bridge-shift, were introduced; then, a dynamic model of tractor shifting was constructed, and the models of key components such as wet clutches and proportional pressure valves were experimentally verified. Finally, the control parameters of the above two shifting strategies were optimized, and the acceleration impact and sliding energy loss caused by them were compared. The results showed the following: the minimum peak acceleration of the bridge-shift method was 0.385807 m/s²; the energy consumption of the bridge-shift method was significantly lower than that of the direct-shift method; the sliding friction work of clutches decreased by 14.92% and 75.84%, respectively, while their power loss decreased by 22.82% and 74.48%, respectively.

The high-precision establishment of drive motor models and “pump-motor” system models is crucial for the development of the agricultural machinery powertrain. The research of this paper studied the series combination of electric drive continuously variable transmission devices, hydraulic continuously variable transmission devices, and power shift transmission devices to form a dual continuously variable transmission system. A drive motor efficiency characteristics modeling method combining the improved sine cosine optimization algorithm and BP neural network (ISCA-BPNN) and a hydrostatic transmission device efficiency characteristics modeling method combining the partial least squares method and the idea of sampling without replacement (PLS-SWOR) were proposed. Various binary control strategies for agricultural tractors were designed and compared. The results show that the two proposed modeling methods can effectively establish the efficiency characteristics models of the motor and hydrostatic transmission device. For agricultural machinery equipped with a dual continuously variable transmission system, it is advisable to apply the comprehensive binary control strategy under medium and high loads, and the pure economic binary control strategy under medium and low loads. This study is expected to provide support for the high-level design and intelligent strategy development of continuously variable transmission agricultural machinery in the future.