Customer

Our customer is a leading vendor of industrial electric machines for multiple industries. Its motors and generators are used at nuclear power plants and mines, they power Arctic icebreakers, railway locomotives, city buses, and subway trains. Several R&D centers ensure continuous product improvement and innovation.

Problem

Hundreds of thousands of dollars and 4-6 months of downtime — this price companies may have to pay when a powerful industrial electric motor fails in a production facility. Critical equipment failures are usually preceded by minor faults and abnormalities in operation. To detect smaller issues, companies are increasingly using condition monitoring systems. Such systems use sensors to continuously check on the health of equipment and can automatically notify personnel of incidents. The most advanced systems use artificial intelligence to accurately predict when a major failure is imminent.

99% of all failures in mechanical equipment are preceded by signs that can be detected (Study by Upsala University).

Solution

Our client designed a sensor to track the health of their electric motors, and Rubius made it intelligent using AI predictive algorithms. We used the equipment’s operational data, collected by our client over time, to train a machine learning model. The model finds analytical correlations between different metrics of the running motor to predict the probability of major failure.

Rubius also developed a web service with a convenient UI to collect data from the sensor and make it available to remote operators and maintenance crews. The sensor, the AI-based application with predictive capabilities, and the web service for remote monitoring form an integrated package that the client calls the AI-Powered Condition Monitoring System (AICMS). AICMS's early detection of defects and AI-based prediction of failures allow planning maintenance in advance and extend the service life of equipment.

We have analyzed over 70 competitor solutions to achieve the best result for our customer and create a system that stands out for its functionality and convenience. We also conducted interviews with the client's team and customers to ensure that AICMS was efficient and user-friendly.

How the AI-Powered Condition Monitoring System works 

The AICMS is based on the Internet of Things (IoT) concept to get all the operational data and alerts remotely.

1. The sensor is attached to the equipment that should be monitored

2. The data acquired by the sensor is relayed to a compact gateway device installed on the equipment or next to it

3. The gateway device uses a wired or Wi-Fi channel to send the data from sensors to the web system

4. The web system processes the input from sensors and presents it to the remote user through an intuitive interface designed for easy logging, visualization and analysis of operational data

How AICMS helps 

Monitors the current state of the engine

To check engine condition in real time, the equipment operator or owner simply launches a mobile app or opens a link. The system helps to track the following engine metrics:

• 3D vibration
• Temperature
• Hours of equipment operation to determine the remaining service life

Automatically detects and reports mechanical faults

Detects the most common mechanical faults and sends alerts about them to users

Reminds the customer when maintenance is due

Regular maintenance is essential to ensure optimum efficiency and extend the service  life of equipment. The system reminds you about maintenance service periods.

Predicts remaining time to failure

Using machine learning-based algorithms, the AICMS predicts the time remaining before the equipment will reach a pre-failure state. 

Result

Client’s experts and Rubius analysts estimate that the AICMS will reduce planned equipment repair costs by 40%. With a new maintenance strategy, the AICMS will help extend the life of electric motors and reduce downtime. Equipment owners can now move from periodic, resource-intensive manual inspections to continuous, data-driven monitoring. In some cases, the customer can now replace typical planned repairs with more cost-effective condition-based maintenance.