This paper describes the application of the Kalman Filter to real-time weight measurement of materials stored in large industrial silos. One of the problems with the real-time weight measurement is the readings are subject to diurnal ambient variations due to the effects of temperature, incident sunlight, wind, rain and cloud. This paper describes how this problem was solved.

A product called SiloWeigh, which used strain gauges attached to the legs of silos, was developed and manufactured in the late 1980’s by Scaletron, a local Montreal company. In the late 1990’s a second generation of the product was developed, in order to improve the product’s performance, using the latest technological advancements. The SiloWeigh hardware and software were used to obtain the measurements for this paper.

One of the areas where the product required improvement was the measurement accuracy. The required accuracy was +/- 2%; however, the measurement accuracy of the first version of the product, in some circumstances, could be in the order of +/- 20%, due to the varying diurnal ambient conditions. The problem could be solved with an expensive retrofit using load cells located under the legs of the silo. However the lower cost strain gauge approach better suited the market’s needs.

This paper describes how the problem was solved using Kalman Filtering. A number of examples are described showing the results obtained from different types of silos located in various parts of North America. The paper also describes the Kalman Filter and how it was implemented in this application, together with insights that were gained during the research and development process.