Prediction method for printing quality of solder paste printer

The invention relates to the field of solder paste printing machine printing fault prediction, and discloses a solder paste printing machine printing quality prediction method comprising the following steps: collecting production data when a printing machine prints a PCB; sequentially counting the number of the PCBs with the tin shortage fault in each m PCBs according to the production data; carrying out normalization processing on the production data, and dividing the production data into a training sample set, a verification sample set and a test sample set; constructing a CNN-BiLSTM deep learning model through the training sample set, inputting the test sample set into the CNN-BiLSTM deep learning model, and predicting the number of tin shortage faults of the PCB; by adopting a one-dimensional convolution mode, the characteristic change condition of the printing quality in the solder paste printing process can be effectively grabbed; and the printing quality condition of the solder paste printer in a next