Price: 0
Number of applications: 7
05.08.26 (inclusive)
By agreement
MVP
R&D Tasks
Light industry
Intelligent control systems
Software/ IS
Maintenance of the equipment is carried out according to a fixed calendar schedule and upon the breakdown. Both approaches are ineffective. Scheduled work is performed regardless of the actual condition of the nodes: serviceable parts are replaced prematurely, resources are wasted, and some failures still occur between routine maintenance. Unscheduled stops are the main source of losses. The failure is detected at the moment when the equipment is no longer working: the production cycle is interrupted, urgent repairs are many times more expensive than planned, the necessary spare part is often out of stock and the delivery time is stretched. Data on the condition of equipment is either not collected at all, or remains inside local automation systems and is not stored in historical depth. Readings are taken manually, recorded in logs and tables, and are not available for analysis. As a result, it is impossible to determine which nodes fail more often, which operating modes accelerate wear and how the resource is actually consumed.
Reducing unplanned equipment downtime by approximately 30-40% by detecting failures at an early stage. Reduction of repair costs: the transition from emergency to scheduled work, performed in a convenient technological window. Prolongation of the actual service life of the nodes — replacement according to the condition instead of replacement according to the calendar, reducing the consumption of spare parts by 15-25%.
Vladislav Antsiferov
Purpose and description of task (project)
To develop a hardware and software complex that collects data from industrial equipment, detects early signs of node degradation and predicts failures before they occur, shifting maintenance from scheduled to state-oriented. System composition: Data collection. Modules for removing telemetry from vibration, temperature, current, pressure and RPM sensors; connection to existing automation systems using industrial protocols; data buffering in case of loss of communication and reloading after recovery. Storage and processing. A time series server with a high recording frequency, normalization and purification of signals, binding readings to a specific piece of equipment and operating mode. The analytical core. Models for detecting anomalies and predicting the remaining resource of nodes, trained on historical operational data and recorded failures. Gradual retraining as data accumulates in a particular enterprise. Service management. Automatic generation of repair requests indicating the node, the nature of the deviation and the recommended time; accounting for the repair history; calculation of the need for spare parts based on the forecast. Interfaces. A web-based equipment fleet status panel with a color indication of risk, a card for each unit with graphs of indicators, mobile access for staff at the site, and notifications of critical deviations.
Whatsapp
