Case Study: Maintenance Module

Maximize maintenance efficiency and minimize downtime

Challenge

Technical facilities employ large numbers of sensors for monitoring and controlling purposes. By applying Cepel technology to the collected data, relations and interdependencies between specific values and component failures or system disruptions can be automatically detected. This approach results in very detailed and far-reaching information about the current condition of the entire facility.

The example below shows sensor data from a pump station. One sensor showed irregular drops of pressure. Cepel was able to correlate this information with maintenance schedules of the facility and could determine the actual reason for the behavior: a defective return valve in another part of the system.

Datenmeer Solution

Cepel detected the change of the probability distribution of the pressure values (Fig. 1, 2) of the sensor. By comparing the date of the events to maintenance schedules it identified the defective valve.


Fig. 1 Distribution of pressure values with functional return valve

Fig. 2 Distribution of pressure values with defective return valve

Based on its precise calculation Cepel is able to detect even the smallest deviations (Fig. 3) in the probability distributions. These deviations can be indicators for system failure or other significant events. At the same time Cepel is very robust and resistant against noise errors within the data.


Fig. 3 Cepel detects the relevant pressure change and anticipates the pending system failure

Your Benefit

Using Cepel to thoroughly inspect sensor data helps reducing maintenance costs and preventing facility downtime.

Datenmeer