Predictive Maintenance. ABB MicroSCADA

14

Common for all the preventive maintenance management programs, is the assumption that machines will degrade within a calculated time frame. For example, if a pump normally runs for 20 months before break down, using preventive maintenance, it would be removed from use and repaired after 19 months of operation. The problem with this is that calculations are not always correct. There are many variables that affect the normal operating life.

Using statistics and calculations to schedule maintenance is often causing unnecessary repairs. The pump in the example could have operated for over 20 months without any problems, so the repairs made were basically a waste of money. Or the pump could have failed in 10 months and it must be repaired with costly run-to-failure techniques. /11/

3.3 Predictive Maintenance

In predictive maintenance the condition of in-service system is monitored in order to predict when something is wrong and maintenance work should be performed.

The main reason and value of predictive maintenance for the customer is to prevent unexpected failures and interruptions in the process. In some cases advantages include increased lifetime of equipment, increased safety and optimized spare parts handling.

The common use of predictive maintenance is regular monitoring of the mechanical condition, efficiency of the operation and other indicators of the operating condition. The indicators could be the vibration of rotating machinery or an infrared image of electrical switchgear, motors and other electrical equipment.

Instead of relying on statistics to schedule repairs and other maintenance activities, like mean-time-to-failure in preventive maintenance, predictive maintenance monitors the real mechanical condition, efficiency and other indicators to define the real mean-time-to-failure.

This data of the actual mechanical condition provide actual data for scheduling maintenance activities. A predictive maintenance program can minimize unscheduled breakdowns and ensure that the condition of repaired equipment is

15 acceptable. The program can also identify problems before they become critical.

Most problems can be minimized by early detection and repairing. Major problems can be usually prevented, if the problem is detected early.

When predictive maintenance is used for a computer system, such as

MicroSCADA system, the indicators are slightly different but the main principle is the same. All components with moving parts are the first one to break e.g. cooling fans and hard disks. But computer systems have also non-physical indicators, like excessive memory or CPU usage that can be monitored to predict oncoming problems. /11/

By interviewing the people involved in maintenance work, the indicators in the system attributes can be found out that could be used as signs of an oncoming problem in the MicroSCADA system.