Combine harvester threshing cylinder harvesting state fault diagnosis method

Abstract

The invention provides a combine harvester threshing cylinder harvesting state fault diagnosis method. The method comprises the following steps of sample collection and preprocessing, training of a network pre-model under a source domain data sample, supervised fine adjustment of a prediction model of a few marked samples in a target domain, and output of a diagnosis result. According to the established prediction model, the weight and the offset value can be updated layer by layer to perform hierarchical expression on the input signal, and correct diagnosis can be more effectively made for the roller fault. According to the method, the weight and the offset value are migrated from the source domain to the target domain to adapt to new target sample recognition, and finally the effect of improving the target domain sample fault recognition accuracy is achieved. By combining the advantages of the transfer learning method in the aspect of solving the samples in different fields, the problem that the sample fault cannot be accurately recognized when the working state samples of the threshing cylinder are insufficient due to many factors is solved, and the target domain sample fault recognition accuracy is improved.

Description

technical field [0001] The invention belongs to the technical field of intelligent agricultural machinery, and in particular relates to a method for diagnosing the harvest state fault of a combine harvester threshing drum, which can be widely used in the fault diagnosis of different working states of the threshing drum. Background technique [0002] At present, the working environment of the combine harvester is harsh and the working conditions are complicated. It is difficult for the driver to distinguish whether the operation sound of the threshing drum is abnormal or not. Therefore, in recent years, there are many fault diagnosis methods for the working state of the threshing drum. The threshing drum is an important part of the combine harvester. It is not only limited by the experience of the driver, but also affected by the terrain and crop posture. It is easy to have faults such as drum blockage, which directly affects the performance of the whole machine; It is very t...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the above-mentioned technical problems, the present invention provides a method for diagnosing the harvest state fault of the threshing drum of a combine harvester. By combining the advantages of the transfer learning method in solving samples in different fields, it overcomes that when the threshing drum has insufficient working state samples due to many factors, it cannot Can accurately identify the problem of sample faults, improving the accuracy of sample fault identification in the target domain

Images