Fault Diagnosis and Prediction Techniques for Advanced Measurement Systems in Flexible Production of Metrology and Inspection

Measurement and verification play a crucial role in flexible production, and with the development of technology, advanced measurement systems in flexible production systems gradually integrate fault diagnosis and prediction techniques to improve production efficiency. In this paper, a deep confidence neural network model, combined with the ISSA-VMD feature fusion model, is used to model fault diagnosis and prediction in flexible production of power systems. The training effect, prediction performance, feature extraction and fault diagnosis of this paper’s model in flexible production are evaluated and analysed through simulation experiments. The Loss value of this paper’s model converges to about 0.05 after 15 rounds of training, and has a good fitting effect on the training and test sets. The RMSE, MAE and R² of the model in this paper are 0.613, 0.371 and 0.988, respectively, which show good prediction performance. And the prediction results in the measurement system of power generation in flexible production are also more close to the real results. In addition, the DBN model incorporating ISSA-VMD feature fusion can completely separate the five fault signals, and the overall fault identification accuracy reaches 98.53% for the fault test set selected in this paper, which has strong diagnostic effect. This study provides more scientific and effective technical support for metrological verification in flexible production.