Investigation into systematic errors in continuous, in motion, conveyor belt weighing systems for bulk materials

<p dir="ltr">Historically there has been a lot of research into the accuracy of belt weighers and most of the sources of belt scale error have been understood. This has not in itself led to greater confidence in belt weighers. They are still considered by the majority of users as unreliable and in need of constant recalibration. This situation comes about because companies that market belt scales do not necessarily have a good scientific or engineering basis for their product. Add to this price competition and inexpert customers and the scene is set for confusion and disappointment. Systematic error, that is, the calibration error which is always there, can be as much as 10% in the extreme, is most often in the order of 1% and is desired to be in the order of 0.1% by the user. If the belt weigher is calibrated with real live material, systematic error can be almost eliminated. But since 95% of all belt weighers are calibrated using a simulated load such as a calibration chain or static mass, large systematic errors are common. The main sources of systematic error in belt scales is studied in this work. Belt weighers have two obvious measurement sub systems, a weigh frame which senses weight on the belt and a tachometer which measures belt travel. Both subsystems are equally important, both relatively easily verified. The almost forgotten third measurement sub-system is the Weigh Length, which is the weight sensitive length of the weighing equipment fitted into the conveyor whose role it is to facilitate the measurement of the linear density of the material running on the conveyor belt (mass per unit length along the conveyor). The weight sensitive zone occupied by the belt scale, known as the Weigh Length, summarises in a single number, a complex and invisible sensing process. The Weigh Length number used in almost all belt scales is based on a theoretical construct and is thought to be the main cause of unknown systematic error in belt scales. It is the Weigh Length which has attracted the largest amount of interest from academic endeavor because its contribution to the process is mysterious and not obvious to the naked eye. The driver for the work is the prevalence of significant systematic error encountered in real belt scale applications. A well conducted live load test can establish calibration thus eliminating systematic error, and, can establish the repeatability of the belt scale if it involves multiple tests, but live load tests are not possible in most real life situations. Even then, such tests do not readily identify which measurement sub-system is at fault. In the absence of live load testing, calibration of the subsystems must be relied upon. The methods available mostly do not involve calibration while in normal use and are simulations of the passage of real materials. This report references many years of personal experience in belt weighing and the sorts of errors which have been observed. You will note in this report that the terms 'Belt Weigher', 'Belt Scale' and 'Weightometer' are used interchangeably, but predominantly the term Belt Scale will be used to describe the device more properly termed a 'Gravimetric Conveyor Belt Weigher'.</p>