Method for modeling thermal error model of gear grinding machine based on bidirectional LSTM (Long Short Term Memory) network

Abstract

The invention discloses a method for modeling a gear grinding machine thermal error model based on a bidirectional LSTM network. The method comprises the following steps: 1) preprocessing original thermal error data; 2) randomly generating a whale population, and judging whether the initial position of the whale population exceeds a preset range or not; if yes, changing the initial position of the whale population into a boundary; if not, keeping the position of the whale population; 3) establishing a Bi-LSTM neural network; 4) mapping the position of the whale population into the batch size of the Bi-LSTM neural network and the number of neurons of a hidden layer; 5) inputting the preprocessed original thermal error data into the Bi-LSTM neural network, and taking MAE as the fitness of the whale optimization algorithm; 6) judging whether the MAE meets a preset requirement or not, if not, updating the position of the whale population, and if the fitness after updating is smaller than the optimal fitness before updating, determining the optimal position X * before updating; 7) judging whether the number of iterations reaches the maximum value or not, and if yes, terminating iteration; if not, making t to be equal to t + 1, and cycling the step 4) and the step 7); and 8) outputting the MAE.

Description

technical field [0001] The invention belongs to the technical field of mechanical error analysis, in particular to a modeling method for a thermal error model of a gear grinding machine based on a bidirectional LSTM network. Background technique [0002] Tooth profile gear grinding machines have made great contributions to economic development and national defense security, and are key equipment to achieve high-efficiency and high-precision grinding of high-performance gears. On the one hand, the grinding accuracy of the gear forming gear grinding machine needs to be improved urgently to improve the geometric accuracy of high-performance gears. However, thermal errors inevitably occur during the grinding process, which hinders the accuracy improvement of high-performance gears. On the other hand, with the improvement of production efficiency requirements, the grinding efficiency of gear tooth profile grinding machines is also getting higher and higher. According to the dif...

AI Technical Summary

Problems solved by technology

However, WOA has disadvantages such as poor optimization effect and poor convergence speed when dealing with complex optimization problems.

Moreover, like other swarm intelligence algorithms, WOA is prone to fall into local optimum

Images