Fault Monitoring and Condition monitoring
In the field of control engineering, fault detection and isolation and condition monitoring have happened to be among the most important topics. As a result, this paper intends to identify these two phenomena and how they are incorporated in the field of automotive engineering. To begin with, it is important to understand the meanings of the two words before proceeding to the other aspects. Finally, the paper will conclude by offering the possible future development for fault detection and condition monitoring. Fault detection and isolation is defined as a field within control engineering whose role is system monitoring and detecting instances of faults by identifying the fault type and where it is located. To facilitate this, sensors are used as the tools of identification. They work in two ways. To begin with, sensors identify a fault through recognizing fault in the direct pattern. Secondly, the difference between the expected values and the readings by the sensors could be used to identify fault within an automotive engine. A fault would be identified through a residual excess over the threshold (Ricardo 2001).
Fault detection can be classified into two types; The Model Based Fault Detection and Isolation and the Signal Processing Based Fault Detection and isolation. Model based fault detection involves the knowledge of how a system operates and hence using observer based approach, parity-space approach or parameter identification based approach in order to detect the occurrence of a fault. In signal processing based approach, neural networks and mathematical operations are used to detect the occurrence of a fault (Ranky 2002).
On the other hand, condition monitoring involves monitoring the conditional parameter of a given machine so that any chances of a developing failure can be detected. This is one of the integral parts of predictive maintenance. Through this process, a failure is not detected. However, the time of failure is detected. This makes it very appropriate for predictive maintenance. In condition monitoring, two methods are commonly used. These include vibration measurement and analysis and oil and wear particle analysis. Vibration analysis involves the use of seismic or piezo-electric transducers to measure the vibrations within the bearing casings of an automotive engine. In addition, it also involves the use of eddy-current transducers to measure the axial vibrations of the rotating shaft. After having recorded the vibrations, the values are then determined against the historical baseline values. On the other hand, oil and wear particles’ analysis involves using an oil sample from the automotive engine and analyzing the wear particles within the oil or lubricating sample. The quality of the lubricant together with the condition of the engine is included in the report given by the analysis. Using these two methods, it is possible to identify abnormalities and predict a failure within a given duration of time (Ranky 2002).
Considering the two descriptions, it is clear that fault detection involves identifying an existent fault within the automotive engine while condition monitoring does not identify an existent fault. It simply monitors the conditions likely to lead to a fault and also determine the period by which the failure is likely to occur. Therefore, fault detection is used in identification and repairing while condition monitoring is used in predictive prevention. It is used before the fault has occurred. This marks the major difference between fault detection and condition monitoring.
Sensors Typical in Engine Control
In the engine control system, there are several types of sensors which work together to ensure the above mentioned. For instance, there is the Electronic Control Unit (ECU) which uses its sensors to detect any changes in temperature for the engine coolants, to smell the exhaust system so as to ensure that there is sufficient oxygen and to listen to sounds of detonation. Other functions within automotive engines that employ the use of sensors include the output actuators and the input sensors. In Toyota engines, the Electronic Control Unit employs the use of a microcomputer which forms its core. As a result, this microcomputer receives information from the input sensors. Then, the microcomputer uses the programming within it to process the information and make decisions.
As mentioned earlier, the Electronic Control Unit relies upon the information from the input sensors so that it can process information and make decisions. Let’s examine how one of these input sensors works. On-Off Type Throttle Position Sensors are an example of input sensors. This device is a switch device which “either pulls a reference voltage to ground or sends a battery voltage signal to the ECU” (Ibarguengoytia 2009).
One of the on-off sensors used in EFI engines utilizes contacts on a dual system. The contacts major role is to generate a voltage signal which, depending on the information, contacts the IDL or the PSW terminals of the ECU. At the closure of the switch contacts, the switch action raises the signal at the ECU. This causes the appropriate action which then processes the information and use the programmed information to make decisions that are then sent to the output actuators.
Intelligent Sensors and Actuators
An intelligent sensor is a type of sensor used in the field of automotive engineering that do not just gather and transmit data but sensors that are advanced and which are able to analyze, interpret, fuse different data from different sensors and be able to validate data that is collected locally. As a result, intelligent sensors are advanced sensors that can assist in dealing with complex sensing tasks and which can be applied in high level events. An intelligent sensor must have several features. To begin with, it must have a high ability of change detection and response. Using the sensor configuration stage, the sensor should be able to examine the collected data and evaluate its validity. After collection, the data collected should be compared with data from the various sensors so that the accuracy of it is identified (Ibarguengoytia, 2009).
Secondly, an intelligent sensor must have an outstanding information processing ability. The sensor should be able to interpret the data from the sensors and apply signal conditioning and event detection to come up with decisions that will increase the efficiency of the whole system (Ypma 2007).
Thirdly, intelligent sensors must be in position to adapt to changes within its immediate environment and have some self-adjusting mechanism that could be useful in the effort to avoid statistical mistakes that could be caused by ageing. Finally, a good intelligent sensor must be able to communicate information that is reliable. The information must be validated and considered relevant and accurate before it is transmitted to other systems engaged in supervision. This means that the intelligent sensors must have features that enable it to self-validate information before communicating it (Patra et al 2005).
Engine Control System Sensors
One of the examples of the intelligent sensors is the Vane Air How Meter Electrical Circuit which is classified under the input sensors that are necessary in the spark calculation and in basic injection. This sensor is placed on the measuring plate and is attached in a way that it rides on a resistor which is connected to the voltage input and the ground. This type of intelligent sensor contains two types where one results into low voltage with high air volumes and high voltage with low air volumes while the second type results into high voltage with high air volumes and low voltage with low air volumes (Boltryk et al 2005).
Another intelligent sensor is the manifold absolute pressure sensor. This sensor, sometimes known as the vacuum sensor is commonly applicable in D type EFI engines. This sensor is fixed on the bulkhead and contains a vacuum line joined to the intake manifold. The main role of this sensor is measuring the volume of air taken in. this is achieved through detection of manifold absolute pressure. The manifold absolute pressure sensor is made up of a piezoresistive silicon chip and an IC with the two sides of the sensor containing a vacuum on one side and manifold pressure on the other side. As a result, pressure intake changes leads to a flexed silicon chip. This, similarly, leads to signal voltage change at the pressure intake manifold. This eventually assists the engine to maintain a given pressure. When the sensors detect these changes, they send a message to the ECU which then processes the information and makes a decision that is passed to the actuators (Powner & Yalsincaya 1994; Sachenko et al 2003).
An example of an actuator is the rotary actuator which has strong parts that are resistant to heat and which is constructed to deal with the harshest of conditions possible.
Future Development
As indicated earlier, prevention is far much better than cure. It is therefore important that automotive engines are prevented from faults rather than being cured from the same. With fault detection and condition monitoring, it is clear that the technological advances are becoming a very integral aspect of this endeavor. It is therefore very important that more effective ways are developed. The more accurate the results from fault detection are, the higher the chances of getting an appropriate solution.
However, there are likely challenges that must be addressed if good results are to be expected. For instance, the sensitivity of these sensors and actuators must be given high priority. Little tempering caused by environmental factors and could lead to poor results. It is therefore important that better sensor gadgets are developed that will be able to adjust to the environmental factors faster. For example, acoustic devices of electrostatic nature usually become faulty when exposed to moisture. Considering the fact that the sensor is very likely to be used outdoors, there are high chances that the data provided will be wrong. It is therefore important that the sensors are made to be self-adjusting so that incase of interference by moisture, the sensor indicates that for a couple of minutes, it will be out of service. After the moisture has been dried from the dielectric plates and the whole system starts working normally, the device should again indicate that it has resumed. With devices that can self detect and adapt to the environment around, the field of fault detection and condition monitoring will be improved greatly (Andrew 2003).
List of References
Andrew J., 2003. Integrity-Based Self-Validation Test Scheduling. IEEE Transactions on Reliability, 52(2): , pp.162-167.
Boltryk P, Harris, C & White, M., 2005. An Algorithmic Approach to the Optimal Extraction of Signals from Intelligent Sensors. Proc. Nanotechnology Conference and Trade Show, May 8-12, 2005, Anaheim,California, U.S.A.
Ibarguengoytia, P, Sucar, L. & Vadera S., 2001. Real time intelligent sensor validation. [Journal Paper] IEEE Transactions on Power Systems, 16(4) Nov: pp.770-5.
Patra, C. Kot, C. & Panda, G., 2000. An Intelligent Pressure Sensor Using Neural Networks. IEEE Trans on Instrumentation and Measurement, 49(4): pp, 829-835.
Powner, T. & Yalcinkaya, F., 1995. Sensor Review. From basic sensors to intelligent sensors: definitions and examples. 15(4): pp19-22
Ranky, G., 2002. Sensor Review. Smart sensors. 22(4): pp, 312-318.
Ricardo. G., 2001. Intelligent Sensor Systems, New York: Wright State University
Robin R.,1996. Biological and Cognitive Foundations of Intelligent Sensor Fusion. IEEE Trans. On Systems, Man and Cybernetics Part A: Systems and Humans, 26(1): pp. 42-51.
Sachenko, A. & Turchenko, V., 2003. Instrumentation for Data Gathering. IEEE Instrumentation & Measurement Magazine, 46(3): pp. 34-41.
Ypma, A., 2001. “Learning methods for machine vibration analysis and health Monitoring”, PhD thesis, Delft University of Technology,