Study of the control of crankshaft fillet rolling machines´ pressure fluctuation based on neural network

This article states the advanced processing craft of automobile engine crankshaft´ rolling alignment and strengthening treatment of circular bead. For the problem of pressure fluctuation in pressure control which happened in the JAW of Roll clamp, we analyze the factors and overcome the pressure perturbation by method of internal model control. This method uses discriminator which composed of ANN OF PID without saturation characteristics and clap delay line to establish a proportional hydraulic system model based on ANN internal model control system framework. Accordingly, it deduces the pulse transfer function which is equivalent to ANN OF PID without saturation characteristics and achieve inverse model. Practice has proved that the method used in the pressure control of automotive crankshaft fillet rolling system is Successful. Furthermore, the article also pointed out the key problem that we should focus on to obtain valid data when modeling Hydraulic system.