Anomaly detection for predictive maintenance and energy saving

Anomaly detection for predictive maintenance and energy saving

Situation

A factory is focused on production of high-performance brake calipers and electronic parking brakes for leading automotive manufacturers worldwide. The business user of the solution is the maintenance department of partial production plant. The factory performed the maintenance of the production lines was done regularly after a specific time period, or after breakage. During the maintenance the production line couldn’t produce, therefore every breakage was very costly. On the other hand, not correctly maintained production lines could lead to leakage of the air which was used in production.

The goal of our solution was to decrease the costs related to energy consumption, caused by leakage of air or increase of the electricity consumption on the testing stations used in production lines. 

Decrease of the energy consumption could be achieved by using predictive maintenance for early removal of the air leakage or decrease of electricity usage

Solution

Our solution should retroactively detect anomalies in energy consumption by calculating dynamic limits. The calculation of the limits also considers the information from the assembly process, like speed of the production, workshift, product type etc. After a predefined number of anomalous consumptions, solutions will deem that the production line should be checked. Our solution would send an alert to the maintenance department with information about the production line, how many anomalous consumptions were detected and at which production line. 

Final solution is to be deployed on the corporate cloud. Results would be viewable in clients inhouse visualization tools.

Result

Solution would grant the maintenance department the possibility to review energy consumption outliers for specific production lines regularly (one a day/week/month/…)). All of the information would be in automatically generated reports, which will be used by experts in the maintenance department to decide if preliminary maintenance checks should be done or not. Using these more targeted checks, the maintenance department should achieve a decrease of the energy consumption and therefore decrease the costs of the production. 

The project data and utilities were also utilized in a follow up project (also by Cognexa), this time aimed at production quality. Knowledge transfer to the internal data team was also ensured.

 

Client

Production factory of a major Tier 2 automotive supplier

Description

An AI system that helps the process engineers identify anomalies in energy consumption and review energy consumption outliers for specific production lines. That leads not only to a decrease of the energy consumption, but also to prevention of machine failure incidents.

Intererested in free AI consultation?

Drop us a line and we will help you find out
how could your company benefit from using AI.


Marek Šebo
Founder & Business Architect

Daniel Šemnický
Business development

    Subscribe to our newsletter

    Empower your services to find the best growth opportunities

    Cognexa.com