Because of a computer’s ability to handle vast amounts of data, it can be employed to monitor the details of a system’s behavior and alert the user to when the system needs his or her attention. Thus, the computer is acting as a tool to reduce a vast amount of raw data and turn it into useful, actionable information.
Consider the following simplified example. Note that while based on real experience, this is a notional example for illustrative purposes.
Everyone is familiar with the oil light in a car. Traditionally, this light is linked to a pressure switch downstream of the oil pump. When the oil pressure drops below a fixed preset limit, the switch trips, illuminating the oil light and alerting the driver.
While this method may be acceptable for protecting the engine, it still can result in a road breakdown. Typically, when the light comes on, the engine is in imminent danger of damage. The driver should, therefore, pull over immediately and seek outside assistance to bring the vehicle in for repair.
PHM strives to improve upon the limited switch methodology by exploiting two phenomena. First, chances are that the oil pump didn’t suddenly “drop out.” The amount of pressure it produced probably dropped gradually from normal to the point of being insufficient.
Similar behavior can be expected for most, but not all, wear-out failures in electro-mechanical systems.
Modern engines are heavily instrumented with sensors - not switches - for oil pressure, oil temperature, engine speed, engine load, etc. If this data is collected while the vehicle is being operated, the oil pressure and engine speed sensor values might appear as shown in Figure 1.
Figure 2 shows how the two sensors relate to each other by plotting the simultaneous values of engine speed and oil pressure against each other. Assuming this data was collected when the vehicle was relatively new, the information about this relationship can be captured and stored by the PHM software for later use.
The data from the same sensors at a later time are shown in Figure 3. As can be seen, the oil pressure is still considerably above the limit set for the oil pressure switch.
However, Figure 4 illustrates that the oil pressure, while still well within a safe range, is lower for a given engine speed. The PHM software is able to recognize this change in the relationship and provide an early alert for proactive maintenance decisions.
This example is simplified for illustrative purposes. In reality, the PHM tools look at how all the sensors relate to each other, simultaneously.
This important feature allows the software to capture much more complex relationships than could be easily visualized by a person. This makes PHM technologies powerful decision-making tools for equipment maintenance.
The applications of this technology within a heavy truck obviously extend well beyond the engine. The driveline, cooling, electrical, HVAC, suspension and exhaust after-treatment systems can all be monitored through the same PHM tool, provided their performance-related data is available.
As mentioned previously, conditioned based maintenance is a concept for maintaining vehicles based on the needs of each specific vehicle. In an ideal CBM program, vehicles are only taken out of service for maintenance when required by the condition of their subsystems.
The most generalized goal of CBM is to reduce operational costs by performing maintenance only when truly necessary.
A successful CBM program, compared to a preventive maintenance (PM) program will reduce unnecessary parts and supplies replacement. It will also allow vehicles to remain in service for longer periods, while still reducing or eliminating unscheduled maintenance and road breakdowns, and maintaining overall vehicle useful life.
It is obvious that these last two outcomes are in competition with the first two, and these competing requirements are the major challenge to implementing CBM. Overcoming this challenge requires greater knowledge of the current condition of any particular vehicle.
Four potential issues