The Comparison of Methods for Mass Measurement on the Continuing Movement Conveyor using Digital Processor

Abstract

the comparison of the different types of time domain filter for dynamic weighting system is presented in this paper. The three popular filters consist of FIR filter, IIR filter and Kalman filter, which are implemented to evaluate its performances both accuracy and computation time. The conveyor belt type checkweigher is also established, comprising of an in-feed, a weighting and an out-feed conveyor. The load cell is mounted on weighting conveyor. Furthermore, two photoelectric sensors are installed at the front and the end point of weighting conveyor. As the result, the total time of object movement on weighting conveyor, which can be detected. The load cell signal is then passed to all filters at a time, utilizing to compensate the error caused by the movement of conveyor. The suitable measurement algorithm is designed to program on a commercial digital processor of Teensy4.0. Experimental results showed that the Kalman filter has ability to eliminate the noise with the smallest error of ±10g similar to IIR filter. However, IIR filter is slowest converged compared with all filters, but it shows the fast computation of 40µs. Meanwhile, the FIR filter is lowest performances both accuracy and computation time.