Gear pitting fault diagnosis with mixed operating conditions based on adaptive 1D separable convolution with residual connection

•The proposed method can effectively detect the faults of different pitting degrees of gears under mixed conditions.•The proposed method can reduce the model parameters by approximately 50% in comparison to the traditional 1D CNN, while maintaining excellent diagnostic performance.•The proposed method uses the grid search and random search algorithms to optimize the hyperparameters of the model .•The proposed method uses directly the raw vibration signals for training without pre-processing, reduces the manual operation workload effectively, and expands the scope of the model for gear pitting fault diagnosis.•The proposed method passes the features through the residual connection and can effectively solve the representational bottleneck problem of the features in the model.