Predictive maintenance is a useful technique that aims to predict the future failure point of a machine or component, so that parts can be replaced, based on a plan, just before it is expected to fail. This means that equipment downtime is minimized and that component lifetime is maximized.
Predictive Maintenance Organization
For this technique to be successful measurement of various parameters checks need to be carried out that show a connection with the components life cycle. Examples of such parameters are:
- Bearings or component vibration
- Temperature of the electrical connections
- Insulation resistance of the motor coil
The first step of predictive maintenance is to develop a historical perspective on the relation between the selected variable and the component life. This is accomplished by taking data readings (eg. the vibration of a ball bearing) at regular intervals until the component fails. This information can then be used to predict how close that component is to failure. In this example subsequent bearings should be replaced when the vibration reaches near the established failure rate. Manufacturers of software for predictive maintenance may recommend ranges and values to replace the components of most equipment; this historical analysis makes it unnecessary in most applications for you to run the tests yourself.
Once the information to practice predictive maintenance on a machine has been quantified the next step is to determine how often variables that are indicative of the machine failure should be measured. The goal of this process is to review the techniques commonly used in monitoring the machine so that procedures can be put in place for required testing. The purpose of monitoring is to obtain an indication of the condition (mechanical) of the machine so that it can be operated and maintained safely and economically. Testing too often means resources are wasted and downtime is increased, testing too rarely can result in missing the need for a replacement.
The inspection of machines can be broken into three segments:
- Monitoring of machines: The purpose is to indicate when a failure becomes likely. You must distinguish between good and bad condition, and if it is bad you should indicate how bad it is compared to the expected failure point.
- Protection of machines: The aim of this is to prevent catastrophic failures. A machine is protected, when the values that indicate their status reach values considered dangerous, the machine stops being used until parts are replaced and its levels fall back to acceptable ranges.
- Failure diagnosis: When a machine foes fail then processes need to be in place to indicated why the failure happened. Was monitoring not frequent enough to notice the problem? Was it a random failure point etc.
In recent times many businesses have switched to predictive maintenance systems. The use of vibration analysis, used oil analysis, wear control etc. has allowed companies to more accurately predict when a failure is imminent allowing for maximized uptime and minimized downtime.