As condition monitoring professionals, we often have to convince the senior management at our plant that a condition monitoring program adds value to the company. To do this, we can show them KPIs of factors such as maintenance activities, mean time between failures, and pre-warning and planning times.
For example, you can show senior management that the measurements of your critical machines are mainly in the green, meaning they are unlikely to fail and are running at a high production rate. You can also show the condition of your B, or slightly less critical, machines, and the C and D machines.
Figure 1: Condition by criticality class.
You can also share the results of maintenance activities by calculating the overall vibration level of the plant. This is done by calculating the overall vibration levels for a number of machines and trending this information over time to assess maintenance activities.
For example, a power plant had a relatively high overall vibration level. They decided to do a balancing campaign and started by balancing all of the fans. The vibration levels started to decrease.
Figure 2: Overall vibration level of a power plant
Then they started an alignment campaign, and the vibration levels decreased further. Then they focused on adjusting the belts. This became a great demonstration of the value of various maintenance activities.
You can also compare overall readings between various production units. In the example shown in Figure 3, there are two very similar plants, and we can compare their vibration levels over two or three years.
Figure 3: Compare overall readings between production units
The measurements are quite similar, but there are differences in some places. These are opportunities for improvement.
You can also compare machine suppliers to calculate the best life cycle cost for a product. For example, Figure 4 shows the vibration levels of four pumps from four different suppliers.
Figure 4: Compare overall readings between suppliers
You can see that the pumps from Suppliers 3 and 4 vibrate quite a bit. They may be less expensive. In that case, the purchasing cost may be lower, but the lifecycle cost would be much higher.
You can also use trending data to assess work done by contractors, for example, on imbalance in fans. This way you can follow the way the values change as well as the performance of different contractors.
One important factor to measure is mean time between failures (MTBF) on different machine types. If you have a machine that experiences a lot of bearing failures, it’s good to have statistics on that.
Figure 5: Mean time between failures
As the MTBF decreases, this KPI can also be used to demonstrate the value of the condition monitoring program.
Figure 6: MTBF by machine type
It’s also a good idea to keep statistics on pre-warning times and planning times. Normally, when you look at trend graphs and see high readings, you detect the fault. You don’t usually write the work order right away. You keep tracking the fault, and after some time, you write the work order to correct the fault. After the fault is corrected, you close the work order.
Figure 7: Remember the pre-warning and planning times
However, don’t discard the statistics on that pre-warning time, meaning the point at which you detected the fault until the point at which you changed the part. This, as well as the planning time, can be useful in the future when it comes to planning. If you have a machine with short pre-warning times, you might consider mounting an online system.
