Kalman filter for dynamic weighing system

Abstract

In the area of mass production, products are weighed using load cell based dynamic weighing systems. A load cell is an uncontrollable weighing device and the value of weight, for the passing product, is estimated by filtering the electrical signal from a load cell. Improvement in filtering increases the speed of weighing and enhances the measurement accuracy. In this paper a Kalman filter is proposed as a weight filter for the dynamic weighing system. Furthermore, the paper includes mathematical models of the load cell and forcing functions. These models are used to examine the suitability of the proposed Kalman filter approach. Since this approach is based on the accurate model of the system in question, the exact model of the load cell based dynamic weighing system has been derived and presented. For one particular value of the weight, the parameters of the model are time-varying due to the product coming onto the weigh-table and due to the product length. Changing the measurement from one value of the weight to another, causes even greater changes in the values of the model parameters and introduces a non-linearity in the system. Thereforc, an adaptivity approach has been considered and a solution proposed. The simulation and experimental results are presented and compared. The results achieved, show that the Kalman filter may provide effective alternative to the conventional method especially when the system is nonlinear and low frequency noise is incorporated in the bandwidth of the useful signal.

Application of Analog Adaptive Filters for Dynamic Sensor Compensation

“…An artificial neural network has been proposed for dynamic measurement which needs a learning phase [4]. Other methods, such as employing a Kalman filter [5] and estimation with a recursive least square (RLS) procedure [6], have also been applied for dynamic weighing systems. Almost all the above reported methods are based on digital signal processing techniques which need analog-to-digital convertors and powerful signal processors.…”

An Improved ADSE Algorithm for Dynamic Weighing System

“…The noise that comes from the machine includes mechanical vibrations, variation in the environment and etc. To extract desired part of measurement data, a Kalman filter is used to filter the noise that comes from the weighing system [13]. The Kalman filter is able to cope with changes of the noise frequency characteristics and give both fast transient response and a stable weighing result [14].…”

Study on Dynamic Weighing System base on intelligent algorithm

“…Cj,k < f(t), Oj,k(t) >= f(t)Yjk (t)dt (1) ldi k =< f(t), Vj,k(t) >= f(t)Vjkk(tdt (2) Where scale function Oj k (t) and wavelet function Vjk (t) is determined by the selection of a particular mother wavelet q(t) and the following equations: [10]. It's not sensitive to the initial value and a global optimum can be obtained when the number of iteration is enough.…”