PNIPU: Industrial Equipment Defect Detection Appendix

The main articles are:

• Neural networks (neural networks)
• Video analytics (terms, applications, technologies)

2024: Presentation of Industrial Equipment Defect Detection Application

Scientists of the Perm Polytechnic are developing an application for detecting defects in industrial equipment. The university announced this on June 5, 2024.

Industrial enterprises are faced with the need to quickly and accurately diagnose equipment to prevent downtime, accidents and improve the quality of their products. For example, 70% of industrial electricity comes from electric motors, of which 20% undergo major repairs during the year. The cost of it can be 80% of the cost of the annual production of such an engine. In addition, large foreign companies for diagnostics and production of equipment left the Russian market. PNIPU scientists create an application based on neural networks for visual flaw detection, which determines the presence or absence of defects with high accuracy. The development will ensure the reliability and safety of production processes, improve the quality and productivity of work in the enterprise.

In many industries around the world, including Russia, visual flaw detection is traditionally carried out manually with the help of specialized inspectors or technicians. This process often requires significant time and financial costs.

Polytechnians are developing a prototype in a form mobile application that will help companies conduct visual flaw detection of production facilities for the presence or absence of damage from images taken with the phone. The product is written in the language programming Java for. operating system Android The application analyzes the image and indicates the presence or absence of defects on the site. The results can be presented in a convenient format - a text description or a graphic designation in the image.

Almost any structure subject to destruction and/or degradation can be diagnosed: machines, machines, any equipment parts, rocket, aircraft engines, household items. For example, if pipelines are damaged, flaw detection will help detect cracks, ruptures, assaults, scratches, nicks and dents. For electric motors, these are defects in the housing, cracks and chips in the bearings.

The development will be used at the enterprises of engine and mechanical engineering, oil and gas, rocket and space, aviation and manufacturing industries. The customer will have an additional opportunity to carry out current or overhaul of production facilities with the least losses, which means not to spend huge funds in case of their failure during the period of active work.

The library of neural networks, independently developed by PNIPU scientists, allows you to fully control the program code. The author's algorithm makes it possible to greatly simplify and speed up the procedure for obtaining a damage diagnosis system by 2-3 times or more. For example, engine flaw detection for June 2024 takes 20-30 minutes, and there will be five.

Our application uses neural network technologies for effective visual flaw detection. This provides higher accuracy compared to traditional methods. Thanks to the mobile platform, users can conduct it anytime and anywhere, which increases their performance and responsiveness to problems. Anyone can download the application for free and check its operation on a test neural network, but with limited functionality and trained on a small amount of data. The estimated cost of full access to all neural networks and the functionality of the application is 2000 rubles/month, - said Ivan Pashkov, a student of the Faculty of Electrical Engineering of PNIPU.

Flaw detection will be carried out without the need to manually adjust parameters or perform complex operations. This will greatly simplify the process and reduce processing time. An intuitive interface and automated analysis process make the product available even to users without special technical knowledge and additional training. Our solution offers Russian companies an innovative tool developed domestically, which reduces dependence on imports and strengthens national industry and technological sovereignty, "added Grigory Kilin, Candidate of Technical Sciences, Associate Professor of the Department of Electrical Engineering and Electromechanics at PNIPCU.

The development of scientists from the Perm Polytechnic will become an important tool for ensuring the reliability and safety of production processes. It will assist in effective maintenance of equipment, accident prevention and safety improvement, monitoring and diagnostics of electrical equipment. The application will improve the quality and productivity of work in the enterprise, as well as contribute to the development of the domestic industry in the country.