Accuracy Control of Fiber Cable’s Outer Diameter with Algorithms of Filtration, Prediction and PID Controller

The qualification ratio of fiber cables is highly related to the accuracy and evenness of its outer diameter. Due to external influences such as fluctuations of current and voltage, there is a high probability that the outer diameter of a fiber cable is out of the expected interval, which can significantly decrease its quality and qualification ratio. To improve the accuracy and evenness of cable’s outer diameter, several algorithms are integrated into our approach. An online anomaly detection and mitigation algorithm is used to eliminate anomaly outer diameters collected by the laser outer diameter scanner. To further enhance the accuracy of outer diameters, the back-propagated neural network is used to train a prediction model and predict subsequent outer diameters one step ahead. Based on the deviation between the predicted outer diameter and the expected outer diameter, a proportional–integral–differential (PID) controller is applied to tune the screw rotation speed of plastic extruder to improve the accuracy of fiber cable’s outer diameter as well as the qualification ratio. To improve the applicability of the PID controller for various fiber cables, a nonlinear regression algorithm is used to optimize the coefficients of the PID. Experimental results show that our approach can significantly improve the accuracy and evenness of fiber cable’s outer diameter and can be applied to circular fiber cable production lines.