Tank сar fault prediction system
Condition analysis, fault prediction, fault cause identification, and operation recommendations
- Oil and gas
- Chemical industry
- Artificial intelligence
- IBM SPSS
Accidents on trade routes cost the global economy billions of dollars. Even an insignificant en-route failure may disrupt the supply schedule, entail a traffic congestion, or even cause fire or toxic emissions. Rubius has developed artificial intelligence that predicts – and helps prevent – tank car faults on railroads.
The system was ordered by a petrochemical plant located in Siberia. The company owns thousands of tank cars in which it transfers end products across Russia and to other countries. Our AI knows everything about those cars, including model, capacity and volume, number of fault-repair cases, what was transferred, where, and when. All information is parsed from the process control system, ERP, CRM, and Russian Railways systems.
What the system does:
- analyzes the tank’s wear percentage and future trip before it goes out on the track: departure time, route, cargo receiver, and the team that will load the car
- uses the received data to predict fault probability, causes, and possible occurrence time
- provides operation and repair recommendations to prevent potential faults
- gathers car fleet statistics: counts the total number of cars, available tanks, repair costs, and outage
- notifies about critical wear and upcoming maintenance
Artificial intelligence helped the customer reduce the number of en-route faults by 30%, increase the share of scheduled repairs by 83%, and turn repair and maintenance costs 9 times. Moreover, AI can establish the dependence of faults on specific employees and cargo receivers, define the optimal tank filling level for maximum service life, and identify the railroads where most faults occur.