University of Twente Student Theses

Login

Design of a dynamic weigh system for conveyors

Timmer Arends, J.H. (2016) Design of a dynamic weigh system for conveyors.

[img]
Preview
PDF
8MB
Abstract:Miedema and Dewulf are merged companies in the field of providing agricultural machines, especially for potato growers. They both attach great importance to innovation. A current field of research is about dynamic belt weighing systems. Miedema would like to have such a system on their conveyors and Dewulf would like to have one on their harvesters. The basic idea is similar for both applications: they have to weigh the product dynamically as accurate as possible and cope with all the present factors. The most important factors are the belt tension, sticking soil, angles, belt velocity, vibrations, non-uniform loading and environmental conditions. For carrying out measurements there are also some parameters that have to be determined like filtering, resolution and calibration. For the Miedema conveyor the goal was set to get a maximum error of 1%. The goal for these dynamic weigh systems is to optimize the process for the used machine parameters. It makes it easy to see if the belt is over loaded. An even more important goal is to get information of the yield per hectare or per hour. For the harvester this information can be coupled to the GPS signal in order to get yield maps of every field and the conveyor can load trucks exactly to the maximum allowed weight. For both applications it is chosen to use a strain gauge load cell. When this cell is loaded the strain causes a difference in resistance and this can be coupled to a mass. Some initial tests and a set-up was already made for the harvester and the goal was to gain more knowledge about the weigh principle and test some of mentioned factors. Before the actual testing literature and theory was covered to understand the test results. The same was done for the conveyor. Miedema already designed a weigh frame and the goal was to test this principle and give recommendations regarding further improvements. From the literature it was clear that the weigh system should be close to the tail, the velocity measurement should be done at the tail pulley, the frame should be long and stiff, the belt tension must be constant, there should be uniform contact pressure with the idlers and sticking soil should be minimised. From a theoretical approximation it became clear that a correction factor is needed to get to the correct value. For the harvester this value is 3.4665 and for the conveyor it is 1.3127. This factor comes from the configuration of the idlers which are connected to the load cells. The theory also shows the great influence of the belt tension and the angles. From testing it appears that the calculated correction factor is very close to the real life situation. For the harvester to real factor is 3.5060 and for the conveyor it is 1.3408. For the harvester the belt shows very large vibrations so a double notch filter and a low pass filter is designed which is velocity dependent. For the conveyor only a low pass filter was necessary to get rid of most of the noise. It is chosen to measure at specific length steps so the velocity has no influence on the measurement. Tests show that a zero calibration is highly desirable so a function is made which measures the value during two belt lengths and then subtracts the mean of this zero calibration from every new measurement. All tests with the conveyor in a real life application resulted in an error lower than 1%. Recommendations towards further research for the Dewulf harvester are to test the influence of the angle which the machine can make, the influence of the belt tension, the influence of measuring with time steps, how to reset the factor when the belt is installed in the machine and a redesign for the steel connection of the belt. For the Miedema conveyor testing should be done over a large time period, a function has to be implemented to change the factor, place the weigh frame more to the tail and rotate it 180 degrees, do not measure if the machine operates at more than 5 degrees in length direction and enclose the gap between the weigh frame and the total frame.
Item Type:Internship Report (Master)
Clients:
Miedema Landbouwwerktuigenfabriek B.V., the Netherlands
Faculty:ET: Engineering Technology
Subject:52 mechanical engineering
Programme:Mechanical Engineering MSc (60439)
Keywords:Dynamic weighing, load cell, strain gauge, mass measurement, calibration, filtering data
Link to this item:http://purl.utwente.nl/essays/69531
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page