Predictive Diagnostics for Crushing Equipment: Petrozavodsk Team Develops an AI-Based Monitoring Solution
Researchers and students at Petrozavodsk State University, working with Karelia-based mining equipment manufacturer Sverus, are developing a hardware and software system that detects early signs of failure in crushing equipment. The project has won funding through a grant competition organised by the Head of the Republic of Karelia's Grants Foundation.
The hardware and software package combines compact wireless sensors with an autonomous data logger. The equipment captures vibration data from crushers, vibrating screens, and other industrial machinery. Artificial intelligence algorithms then analyse the signals to identify early indicators of equipment faults before they develop into major failures.
Predict Rather Than Repair
The move toward predictive diagnostics reflects a broader shift in industrial maintenance philosophy. Traditional maintenance has largely followed a reactive model: repair equipment after it fails. Detecting emerging defects such as bearing wear, rotor imbalance, or loosened fasteners before they lead to serious breakdowns is steadily gaining acceptance across industry. Vibration analysis is one of the most sensitive diagnostic methods, making it possible to detect developing problems at a stage when no visible signs of deterioration have yet appeared.
The Petrozavodsk State University team has focused on addressing one of the main barriers to wider adoption: affordability. Commercial predictive maintenance systems are often too expensive for small and medium-sized enterprises. The Karelia-based developers have instead chosen a lightweight, cost-effective architecture built around compact wireless sensors and an autonomous data logger. Straightforward installation is another important advantage of the system.
A Dedicated Team for a Specialized Challenge
Project partner Sverus manufactures vibrating screens as well as cone and jaw crushers. Until recently, equipment of this type in Russia was rarely fitted with integrated vibration sensors, and its vibration characteristics under normal operating conditions had received limited study. As a result, the development team adapted its algorithms to the low-frequency vibration range typical of large industrial machines and their bearing assemblies. That approach is expected to improve diagnostic accuracy specifically for this class of mining equipment.
According to industry experts, collaboration between a university and an industrial manufacturer is essential for ensuring that the final product addresses practical operational needs. Although the concept of using sensors and analytics to detect developing faults is well established, the technology continues to evolve. Vibration and acoustic diagnostic techniques began emerging in the mid-20th century, and industrial users have gradually progressed from standalone instruments to automated systems with computer-based analysis. Today, students and industry engineers are refining those methods further by incorporating the realities of modern production environments.
Equipment Monitoring: Past, Present, and Future
Russia's mining industry already offers successful examples of similar technologies in operation. Sputnik-Service, for example, equipped a vertical shaft impact crusher at a crushing and screening plant with a bearing condition monitoring system. The sensors tracked vibration and temperature, while operating data was displayed on HMI panels and remote monitoring systems based on Simple-Scada software. The solution also enabled continuous collection of motor load, temperature, and vibration data for subsequent intelligent analysis to predict equipment failures.
In 2022, Russian company NPP TIK, which specialises in vibration diagnostics and industrial equipment protection systems, expanded production of domestically manufactured monitoring instruments to replace imported products.
A year later, Karelia launched the Digital IT Camp 4.0 project, where students addressed artificial intelligence challenges submitted by regional enterprises, while the strongest participants earned industrial internships. Among the priority application areas identified were equipment condition monitoring, predictive maintenance, and production process optimisation.
In 2024, Rostekh introduced a Russian hardware and software platform for predictive equipment diagnostics. The system combines proprietary sensors with artificial intelligence-enabled software to assess machine condition and forecast remaining service life.
From Theory to Industrial Practice
The Petrozavodsk project combines Industrial Internet of Things technologies, wireless sensors, vibration signal processing, and machine learning into a solution designed specifically for the needs of a Russian mining equipment manufacturer. Under the grant programme, the team will conduct its first field trials at Sverus production sites. Once the system's effectiveness has been confirmed, the next stage will be to scale the technology and adapt it for other categories of rotating and vibrating equipment, including pumps, electric motors, gearboxes, conveyors, fans, and machine tools.
The project is particularly valuable for students because it gives them an opportunity to apply their knowledge in real industrial settings. Their jointly developed solutions, created together with practicing engineers, are tested under actual operating conditions. That experience allows them to make a tangible contribution to strengthening the technological capabilities of Russian industry.
