Steel manufacturers use huge cooling chambers in order to cool the output of their casters to a temperature needed to handle it. The caster is operational 365 days a year with 8 maintenance dates based on a fixed interval. A failure of the cooling chamber results in a downtime of the caster and is classified as a plant failure.
The current maintenance strategy for this type of cooling chambers is based on fixed intervals and the data generated by already mounted vibration sensors. On every maintenance date all the necessary work (repairing and cleaning) is performed to these chambers. The external company which is in charge of the cleaning task has a lead time of several days. The consequences of this fixed maintenance schedule is a suboptimal maintenance efficiency caused by:
- Interruptions before the actual maintenance date (maintenance too late)
- Maintenance work being performed to a perfectly working plant (maintenance too early).
In order to increase the efficiency of the maintenance process this company asked us to develop a predictive model which is capable of predicting the real maintenance needs with a prediction horizon of one month. Our inCARE platform gathers and analyses all the available sensor data from the cooling chamber for patterns in the data which indicate the remaining life time. Due to our platform, the company is able to switch from a rigid and inflexible maintenance strategy to a flexible and future condition based maintenance strategy. Additionally, the inCARE platform offers tools for a root cause analysis which enable the company to optimize their own plants.
The result of the utilisation of the inCARE platform is a strong increase in plant availability as well as a strong decrease in manufacturing costs.
Due to its adaptive models and machine learning capabilities the inCARE platform can be used on a wide variety of technical machinery providing the same advantages.
Please contact us via Email if you need further information or a live demonstration of our application.