AvtoMash introduces a machine vision system to optimize the quality of van assembly

2024: Pilot Project on Implementation of Machine Vision System

On January 17, 2024, the MCD company announced that the AvtoMash plant was introducing a machine vision system to optimize the quality of van assembly.

As reported, the company was faced with the task of increasing the efficiency of assembly work and optimizing the traceability of parts in production. To do this, representatives of AvtoMash LLC, together with experts from Modeling and Digital Twins JSC (MCD JSC), launched a pilot project to introduce a machine vision system at the enterprise.

One of the problems that had to be solved together with AvtoMash LLC was the insufficient level of digitalization of processes at the enterprise. The material entering the warehouses was registered in the accounting system, but its further use was not fully tracked. Only finished products that entered the assembly workshop were registered. In addition, tasks for workers were formed in paper form, and completion reports were filled out manually after each shift. There was also a problem with the variety of details. They are very similar to each other, which made their identification difficult for employees of the enterprise.

The system developed by MCD JSC made it possible to solve these problems. The company's specialists used it to introduce digital magazines at the AvtoMash plant, as well as a post with a video camera for part recognition. According to experts of MCD JSC, the introduction of a machine vision system will increase the efficiency of the enterprise by 25%. In addition, AvtoMash specialists noted a decrease in the process time (runway) by 28.3% and the indicator of unfinished production in the stream by 50%, as well as an increase in production per person by 13.1%.

As part of the pilot project, a version was created for the enterprise that allows employees to independently fill out part catalogs and learn how to use the system. In the future, this will make it possible to exclude the influence of the human factor at the AvtoMash plant, as well as reduce the number of errors.

source = MCD

Work on the system began after the necessary technical inspection and pre-design inspection of the enterprise, as well as the collection of the necessary data - a catalog of parts, 3D models, photo and video materials. At the same time, there were a number of difficulties. Thus, the model developed by the specialists of MCD JSC is not typical, since it distinguishes not versatile, but similar objects. In addition, during model training, a large amount of manual markup is required to correctly identify parts. The pilot project allowed to test the capabilities of the system. If difficulties arose, another scenario was used, including the application of other algorithms, including machine vision. The implementation of the pilot project took 1.5 months.

The productivity and quality of manufactured products is considered an important requirement in today's enterprise. Automation of production is, first of all, a gradual set of measures providing for the replacement of manual labor-intensive operations with modern technological solutions. The introduction of the machine vision system at the production of AvtoMash LLC made it possible to raise the overall level of digitalization of the enterprise, as well as optimize the efficiency of production processes, reducing the impact of the human factor. Thanks to the addition of automated workplaces (AWS), many employees can not only receive appropriate working conditions, but also the opportunity to improve their qualifications, which, in turn, will ensure the digital growth of the company and development in general. The introduction of AWS is necessary for the traceability of the process of obtaining parts from the procurement stage to the finished product in real time. The development of digital logs led to a more streamlined workflow, allowing better utilization of resources and optimized performance. In addition, this facilitated the exchange of data on the stages of manufacturing parts between different production areas, contributing to the creation of a more centralized and information working environment. The result of this system was positive transformations that optimized the transparency of production processes, the efficiency and accuracy of order fulfillment, with the exception of parts re-sorting. commented Ilya Pugachev, Technical Director of Machine Vision, MCD JSC

The key task that was to be solved when introducing digital technologies was to increase the traceability of the process of obtaining parts. For this, certain metrics are used - for example, the management can see the metrics of the production of parts, determine at what stage their production is, and also understand how much is already manufactured.

source = MCD

Operators can track the manufacturing stage of a part. Each of them has its own route, consisting of laser cutting for blanks, bending metal, shot blasting, painting and removing parts. At the last stage, the part article is determined using a camera and a machine vision system.

All information regarding the manufacturing steps of the part and the degree of its readiness is recorded in digital logs. Logs are stored in digital databases and can be filtered, exported, and used to generate reports. At the end, a general journal called InfoCenter is used - it records completely finished parts, allowing you to determine their quantity.

The system is also aimed at helping employees of the enterprise who are not familiar with the denominations of all products produced by the company. Using this solution, they will be able to identify the right products and ship to the right place in the warehouse. Thus, the system allows you to add comments to part routes if it goes according to special criteria. These tips help employees and streamline their work.

In the process of mapping the current state on the sandwich van manufacturing stream, a number of problems were identified in the painting, bending and laser cutting areas of the metal. First, the lack of traceability between sites. Secondly, the extra time spent on finding parts after painting. And the latter - the lack of operational and reliable information from the site managers about the release of components in real time. To solve certain problems, together with the team from MCD JSC, a database of components was created, digital logs with tasks were introduced, automated workstations (AWS) were equipped, a stationary post for detecting parts after painting was organized, an event log and a joint information center for managers were created for operators. All assigned tasks within the framework of the pilot project were solved. Our next step is to replicate the project within the enterprise. told Smolyaninov Roman Dmitrievich, Development Director of AvtoMash LLC

In the future, companies expect to launch a major digitalization project. As of January 2024, the system operates as a digital log for each individual AWS. In the future, they are expected to be linked through a shift task. AvtoMash LLC wants to use the system on a larger scale. The company plans to automate routine processes, minimizing the impact of the human factor. Each operator will have a tablet with an installed system, which will receive a replacement task, where indicating the operation to be performed.