Noise Measuring System and Analysis Techniques for Diesel Engines - Lab Report Example

A Report on Noise Measuring System and Analysis Techniques for Diesel Engines Name Course Tutor Date Abstract This report discusses some major aspects concerning the noise measuring system and analysis techniques with regard to a diesel engine. Some of these aspects cut across the measurable parameters for monitoring the condition of a diesel engine, vibration and noise measuring system setting up procedure, time and frequency domain analysis techniques for extracting information that is related to the condition of an engine from the measured noise under different operating conditions, joint time-frequency domain analysis techniques, and the classification algorithms for decision making on the health of a machine. Nomenclature Temperature Temperature can be defined as the ability of a body to transfer its thermal energy to another body. It is measured in Kelvin or Celcius. Sensitivity The sensitivity of a sensor refers to the ratio of its electrical output to that of its mechanical input. In most cases, output is expressed in voltage per acceleration unit. If an instrument generates its own voltage without depending on an external source of voltage power, then the specification of its sensitivity is sufficient. Natural Frequency The frequency at which any undamped system that has a single degree of freedom oscillates on momentary displacement from its position of rest and can be use in determining the range of a vibration measurement that is useful. Amplitude Limit The maximum acceleration range that an accelerometer can measure is what is called an amplitude limit. Table of Contents Abstract 2 Nomenclature 3 Table of Figures 6 Research Objectives 7 Introduction 8 I.Parameters that can be Measured for Monitoring the Condition of Diesel Engines 8 (a)An Outline of Five Parameters that can be Measured and used for Monitoring Diesel Engines 8 (b)Brief Description of Parameters that can be Measured and used for Monitoring Diesel Engines 9 II.Vibration and Noise Measuring System Setting up Procedure for Diesel Engines 12 a.Diesel Engines Vibration and Noise Sources with Focus on In-cylinder Pressure 13 b.Audio Acoustics Measurements’ General Fundamentals and Available Technologies 14 c.Working Principles and Characteristics of Sensors 15 i.Sensitivity 15 ii.Natural Frequency 15 iii.Amplitude Limit 15 iv.Amplitude Linearity 15 v.Frequency Range 15 vi.Phase Shift 15 d.Transducers Testing Procedure and Facilities 15 Test Setup 15 Results and Discussion 16 e.Types of Measurements 20 III.Time and Frequency Domain Analysis Techniques for Extracting Engine’s Condition Related Information from the Measured Noise under Different Operating Conditions 20 Advantages and Disadvantages of Different Types of Analysis Techniques 21 IV.Joint Time-Frequency Domain Analysis Techniques 21 a.Why this Type of Signal Processing 21 b.Commonly Used Techniques for Time-Frequency Domain Analysis Techniques 21 c.The Advantages and Limitations of Different Techniques 21 Advantages 21 Disadvantages 22 d.Classification Algorithms for Decision Making on the Health of a Machinery 22 Conclusion 22 References 23 Table of Figures Figure 1: A Signal Representing an Engine Body Vibration 10 Figure 2: Thermocouple Embedded Positions 11 Figure 3: General arrangement on an electronic measurement system 12 Figure 4: General Layout of a Measurement System 13 Figure 5: Sources of Noise in a Diesel Engine 14 Figure 6: A typical test setup of a transducer 16 Figure 7: Illustration of Period Excitation 17 Figure 8: Measured Acceleration Responses by MEMS Accelerometer and Reference PCB accelerometer for Random Excitation 18 Figure 9: Comparing the Measured Responses by a MEMS accelerometer and PCB accelerometer when the armature of the shaker is excited by hammer impacts 19 Research Objectives The main objectives of this research are: To illustrate the parameters that can be measured for monitoring the condition of diesel engines To Illustrate vibration and noise measuring system set up procedures for a diesel engine To discuss the time and frequency domain analysis techniques while extracting information related to engine condition under different operating conditions To detail time-frequency domain analysis techniques. Introduction One of the most important strategies for monitoring electro-mechanical machinery is condition based maintenance. In order to get more insight into practical issues to be considered while designing and implementing a noise and vibration measuring system of any given machinery, noise and vibration measurement and analysis techniques should be used. For this research, a typical diesel engine will be considered by looking into the strategies and parameters that can be measured in order to monitor it. Further, the nature and mechanisms through which noise and vibrations are generated for such engines will be evaluated, alongside the measuring systems and techniques applied during data analysis. I. Parameters that can be Measured for Monitoring the Condition of Diesel Engines A parameter that has an impact on how a measurand behaves is called a control parameter. Therefore, a parameter is controllable when its value can be maintained at varying sets of measurements. By measuring different parameters, condition monitoring of machinery is possible in determining how bad or good the mechanical condition of a diesel engine is, and the results obtained can be used in conjunction with predictive maintenance. (a) An Outline of Five Parameters that can be Measured and used for Monitoring Diesel Engines Temperature Temperature can be defined as the ability of a body to transfer its thermal energy to another body. It is measured in Kelvin or Celcius. Vibration Vibration parameter and it varies from one point to another and from one engine to another. However, similar components have got similar vibration signals. Oil Pressure Pressure has a direct impact on the valves and the diesel engine cylinder walls. The quality of combustion can also be affected by the kind of engine pressure. Oil Debris Oil debris also hit on the cylinder walls, thus accelerating wear, corrosion, and abrasion. Electronic Parameters These parameters are determined by use of electronic measurement systems. (b) Brief Description of Parameters that can be Measured and used for Monitoring Diesel Engines Vibration Vibration parameter is crucial in condition monitoring of a diesel engine. In this case, the condition of the engine is determined by way of analyzing the vibration signals generated. In fact, every machine produces its unique form of vibration during its operation, and theoretically, the vibration for most machinery is understood, or can easily be predicted through instrumentation, such as the use of wide band analyzers and transducers. Some of the methods used for processing and storing vibration signals for analysis include the peak to peak or RMS value of the signal for determining the mechanical condition of the diesel engine, and spectrum analysis for accepting a signal and breaking it down into individual frequencies by use of Fourier processes. Further, we have the method of envelope analysis for limiting signals’ required frequencies’ condition monitoring of a given system, whereas cepstrum analysis provides details about repeating components in a spectrum that has same frequencies. Envelope analysis suppresses the unwanted background vibrations and builds an envelope to surround the signal being analyzed. This eliminates troublesome low frequencies from other forms of vibrations. However, the use of vibration monitoring in diesel engines is limited due to the fact that some signals involved are complex, for instance the signal shown in Figure 1.3 Figure 1: A Signal Representing an Engine Body Vibration Measurement of Vibration The vibrations that exceed 5000 Hz occur when metallic materials come into contact. Therefore, if high frequency parts of vibration are measured and monitored when bearing metal and shaft come into contact, then condition monitoring of the bearing becomes feasible. Temperature Temperature can be measured and its numerical values computed by temperature-sensing devices. The devices translate temperature into some reference current, voltage or resistance. The main temperature-sensing devices include: Thermocouples, thermistor, temperature-transducing ICs, and resistance temperature detectors. Temperature sensors are useful in standard condition monitoring of the main bearings in the diesel engines. The temperature changes are due to friction on the bearing, when the shaft and bearing come into contact. Measurement of Temperature It is noted that the sensitivity of the thermocouple applied for measuring of temperature depends largely on its diameter. Therefore, if three thermocouples are used, each with diameter 0.5mm, and the thermocouples positioned at different three positions, sensitivity can be compared since it depends on the embedded depth. Figure 2 bellow illustrates how the thermocouples are embedded on different positions on the bearing of an engine. Figure 2: Thermocouple Embedded Positions Oil Pressure or Cylinder Pressure Information on in-cylinder pressure provides a lot of information that can reflect the condition of an engine and its combustion efficiency (Albarbar, Gu, Fengshou & Ball, 2010). It is mainly based on the application of front flame pressure transducer, among other techniques, such as strain on the cylinder head bolts, acoustic emission and in-structure-borne sound. From the information on in-cylinder pressure, faults such as wear, incorrect injection timing and sticking rings can be diagnosed and detected. Since there is a strong relationship between power distribution and cylinder pressure, monitoring of in-cylinder pressures can be used to detect early problems related to combustion. Oil Debris The content of the lubricating oil of a diesel engine is a good indicator in terms of the condition of the engine with regard to corrosion, wear, and combustion of products in the combustion chamber (Albarbar, Gu, Ball & Starr, 2007). The analysis of the oil content can be useful, especially in circumstances where vibrations analysis has proved difficult. There are several techniques that can be used when analyzing oil. For example a simple test on viscosity can be carried out in order to determine change in the chemical structure of oil or the flashpoint can be determined to indicate if the oil fuel is contaminated. The other technique is that of using a magnetic chip detector where, a magnetic plug gets inserted into a lubricating system and the indication of wear is determined by the size of the attracted particles and the rate at which the particles are attracted. In terms of spectrography technique, the concentration of different particles within the oil is determined whereas the ferrography technique is useful is separating the particles with regard to their sizes. The size and shape of the particles can be analyzed to find out the failure mechanism such as fatigue, corrosion and abrasion. Electronic Parameters These parameters are determined by use of electronic measurement systems. Figure 3 below illustrates the general arrangement of an electronic measuring system. Figure 3: General arrangement on an electronic measurement system From the general arrangement above, the signal is detected from the physical phenomena by a sensing device and then filtered and amplified before being transmitted to a screen in form of an alternative current (a/c). The signal data is displayed on the screen where it can be processed. The devices used in digital measurement have several components that interface a digital device with the analog aspects. The common interfacing components include the analog-to-digital converter and the digital-to-analog converter used in data acquisition systems. II. Vibration and Noise Measuring System Setting up Procedure for Diesel Engines A general measurement system is as shown in Figure 4 bellow. The role of the transducer is to convert the sensed information into a signal that can be detected. Figure 4: General Layout of a Measurement System From the diagram, it can be seen that the sensor is placed within the process so as to pick a signal for the transducer stage. The transducer converts the information that has been sensed into a detectable signal. The signal is then amplified and filtered before it is relayed for control stage. a. Diesel Engines Vibration and Noise Sources with Focus on In-cylinder Pressure Noise out of a diesel engine is made up of several components as shown in Figure 5. The sources include mechanical noise, combustion noise, or a combination of combustion and mechanical noise (Albarbar, 2012). The noise is excited by different forces and mechanisms, and transmitted through some paths. For instance, combustion noise emanates from the rapid rate at which the cylinder pressure is increasing. The pressure also leads to mechanical vibrations, and excites the resonances within the gases in the combustion chamber. The excitation force or pressure, in most cases, distorts the structure of the engine, causing it to vibrate like damped free oscillations. Figure 5: Sources of Noise in a Diesel Engine The table below illustrates excitation forces, vibration transmission, and the associated noise emitters. Source of Excitation Force applied to structure Vibration Transmissions Noise Emitters Combustion Pulses due to rapid change of cylinder pressure Piston, cylinder head, and connecting mechanisms ICE Block Manifolds covers Mechanical Mechanical impacts, bearings, fuel pumps, injection, and piston slaps Cylinder walls and piston connections Timing cover and ICE Block Sump b. Audio Acoustics Measurements’ General Fundamentals and Available Technologies Audio acoustic measurements are generally based on the application of microphones to convert the quickly changing sound pressure signals into electric signals whose voltage variations are analogous to sound pressure. When carrying out acoustic measurements, the engine’s operating environment is controlled from outside noise by closing all the doors, and switching off all surrounding machines in order to minimize the contamination of acoustic signal (Albarbar, 2012). The use of microphones is one of the technologies available for measurement of sound. In order to get accurate results, the microphone should be stable, sensitive, respond well to frequency, and be capable of operating in a wide range of sound fluctuations (Skobts, Izotov & Tuzov, 1966). c. Working Principles and Characteristics of Sensors There are several working principles and characteristics of sensors. They include sensitivity, amplitude limit, natural frequency, resolution, frequency range, amplitude linearity, and phase shift. i. Sensitivity The sensitivity of a sensor refers to the ratio of its electrical output to that of its mechanical input. In most cases, output is expressed in voltage per acceleration unit. If an instrument generates its own voltage without depending on an external source of voltage power, then the specification of its sensitivity is sufficient. ii. Natural Frequency The frequency at which any undamped system that has a single degree of freedom oscillates on momentary displacement from its position of rest and can be use in determining the range of a vibration measurement that is useful. iii. Amplitude Limit The maximum acceleration range that an accelerometer can measure is what is called an amplitude limit. iv. Amplitude Linearity When an accelerometer is excited from the smallest acceleration level that can be detected to the highest, it reports its acceleration output in terms of voltage, and the level of accuracy at which it reports is known as amplitude linearity. v. Frequency Range The range beyond which the sensitivity of a transducer does not change more than its stated percentage is a frequency range, and it is limited by mechanical or electrical characteristics of a given transducer or by the auxiliary equipment associated to it. vi. Phase Shift Phase shift refers to time delay between the electrical output signals and its corresponding mechanical input. d. Transducers Testing Procedure and Facilities Test Setup Figure 6 bellow shows a typical test setup of a transducer. A small shaker is connected to the shaker power amplifier, the signal generator, and the computer for collecting and storing data. Signal processing by the computer uses MATLAB. On the armature that is attached to the shaker, a conventional accelerator and three MEMS accelerometers are attached back to back (Albarbar, 2012). Figure 6: A typical test setup of a transducer The specifications of the accelerometers are listed on the table below: Table showing Accelerometer Technical Specifications Results and Discussion The setup shown in Figure 6 was used and the results compared. Four accelerometers were engaged and simultaneously connected to a BNX shielded cable and each output connected to a response filter to eliminate interference, noise and antialiasing. Data collection was done at a sample frequency of 9kHz, and both averaged & Hanning windowing were employed at all the tests. (i) Period Excitation Two frequencies, 53Hz and 95Hz, were chosen for application of sinusoidal signals to the shaker, away from 50Hz line frequency & its harmonics, and several experiments performed using different amplitude levels. The responses obtained were measured from all the accelerometers at the same time. Figure 7 bellow illustrate some of the responses. Figure 7: Illustration of Period Excitation These responses illustrate there is no distortion is the responses measured by the accelerometer but there is a clear change in sensitivity and shift in comparison to reference accelerometer. (ii) Random Excitation This excitation is similar to sinusoidal tests in the sense that the shaker is excited at random excitation in a frequency range of 10Hz-1.2kHz using different amplitudes. The results for the responses are as shown in Figure 8. The responses of the accelerometers seem identical in time-frequency domains, but the estimated sensitivity is close to that of the design value but more different from the estimated one during the sinusoidal tests. Figure 8: Measured Acceleration Responses by MEMS Accelerometer and Reference PCB accelerometer for Random Excitation (iii) Impulsive Excitation Impact excitation is provided at the armature’s centre through the application of a soft tip hammer within a frequency range of 400-500Hz and the obtained responses are as shown in Figure 9. Figure 9: Comparing the Measured Responses by a MEMS accelerometer and PCB accelerometer when the armature of the shaker is excited by hammer impacts The responses measured represent decay responses as expected of an impact excitation. In this case, sensitivity estimate is 155.9mv/g, different from earlier estimates. e. Types of Measurements Measurement of sound can be done under two forms of environments, indoors (reverberant) and outdoors (free field). A free field is uniform, undisturbed by other sound sources and is free from boundaries. Examples of free fields include the outdoors that is well-above-the-ground and anechoic chambers. The sounds given out by the source in a free field propagate away from the source and are never reflected back. Indoor acoustic environments have got boundaries that reflect sound. When all the incident sound is completely reflected without being absorbed, the resulting sound is said to be perfectly reverberant, and energy flow is equal in all directions. III. Time and Frequency Domain Analysis Techniques for Extracting Engine’s Condition Related Information from the Measured Noise under Different Operating Conditions Airborne acoustic measurements are described in different ways and, time domain and frequency domain are important methods to use for condition monitoring (Albarbar, 2012). Time Domain In condition monitoring, time period that is of concern is every small, few milliseconds. The output amplitude out of a microphone is normally shown on screen as a function of time, and this is useful in identification of some of the noise generating events and mechanisms within a machine. For the case of internal combustion engine, the firing sequence and some of the changes that occur in some mechanism can be seen on a time domain of an acoustic data. Some of the parameters that can be calculated from a time domain measurement include the standard deviation, mean value, kurtosis and skewness, and can be used for indicating the characteristic of an acoustic signal. Frequency Domain Multiple periodic waves of different frequencies can be obtained by breaking down different acoustic signals. In a frequency domain, a signal is represented by a spectrum. For a spectrum to be plotted, the signal is first transformed to a function of frequency from a function of time by use of Fourier transform, and the energy of a given signal is computed by finding the square of the modulus of the Fourier transform or the signal. Advantage Features related to frequency domain are more consistent in detecting damages than the time domain parameters. Disadvantage The disadvantage of using frequency domain analysis is that there is a low resolution and energy leakages. Windowing techniques such as Hamming and Hanning are resourceful in reducing or eliminating energy leakage. Further, this analysis provides only information related to the frequency components of the signal that has been measured. Advantages and Disadvantages of Different Types of Analysis Techniques IV. Joint Time-Frequency Domain Analysis Techniques Joint time-frequency domain analysis applies well in a phenomenon of machinery that is highly transient, such as in condition monitoring of diesel engines. a. Why this Type of Signal Processing Joint time-frequency domain analysis technique is useful because a number of engine airborne signals relate to events such as valve operations and combustions which occur at fixed times that are determined by an engine’s crank mechanism. Therefore, if a time-frequency analysis is performed, these events can be determined depending on their occurrence in terms of a combination of time and frequency. b. Commonly Used Techniques for Time-Frequency Domain Analysis Techniques Some of the commonly used methods of time-frequency domain analysis include Short-time Fourier Transform, Wavelet Transform, Bilinear Time Frequency Distributions and Linear Time Frequency Distributions (Yildirim, Erkaya, Eski & Uzmay, 2009). Both Wavelet transform and short-time Fourier Transform are classified as linear time-frequency transforms because they map a signal into a time-frequency plane, and remain very sensitive to signals that are transient. Bilinear time-frequency distributions apply to analysis of transient signal and normally produce large ripples on a component’s envelope. c. The Advantages and Limitations of Different Techniques Advantages Both bilinear time-frequency transforms and linear time-frequency transform posses some superiority over the conventional frequency analysis based on Fourier and produce better time-frequency locations and resolutions. Bilinear time-frequency transforms has the limitation that when applied to analysis of transient signals, large ripples are produced on the component’s envelope, and this leads to loss of some important information required for condition monitoring. Wavelet is useful when it comes to the analysis of non-stationary signals, such as the ones used in reciprocating machines (International Conference on Wavelet Analysis and its Applications, 2003). Further, wavelet is better than other time-frequency techniques because it produces a continuous wavelet transform, whose properties are better for analyzing the impulsive acoustic signals, for instance the signals form diesel engines. CWT offers better detection of signal and extraction of feature than discrete wavelet transform (DWT). For CWT, is produces non-linear frequency content representation and thus making it appropriate for detection of small transient events. It can be implemented on different frequency scales. Disadvantages The limitation with linear time-frequency distribution in frequency analysis is that some temporal information get obscured, thus called for the use of short-time windows to overcome this limitation. Further, a high resolution in time & frequency cannot be obtained simultaneously, unlike with the wavelet analysis. d. Classification Algorithms for Decision Making on the Health of a Machinery All machinery requires some level of monitoring in order to enhance safety and availability. Therefore, the overhauls and maintenance are based on measured conditions. With proper monitoring of a machine’s condition, the machinery’s maintenance costs are lowered whereas its life is prolonged (Yuan, Yan, Li, Zhang, Sheng & Zhao, 2013). Different algorithms can be used to diagnose engine faults, the main ones being Least Mean Square (LMS) and Recursive Least Squares (RLS). The use of LMS is simple and effective. Because it is simple, it offers a standard for comparison of all the other adaptive filtering algorithms. Generally, LMS is based on stochastic gradient which a descent approach based on the filtering process that involves error estimation, and computation of a linear filter output. Conclusion In general, different parameters such as temperature, pressure, oil debris, and electrical measurements form basic part of a measured monitoring system of a diesel engine. Further, the article evaluates the vibration and noise measuring systems, transducer testing procedure, details on time and frequency domain analysis techniques, and the algorithms used in decision making process on the health of machinery. References ALBARBAR, A., GU, F., BALL, A., & STARR, A. (2007). Internal combustion engine lubricating oil condition monitoring based on vibro-acoustic measurements. INSIGHT – WIGSTON THEN NORTHAMPTON-. 49, 715-718. ALBARBAR, A.-H. (2012). Vibration and noise of diesel engines: Theory, measurement and analysis. Colne, LAP Lambert Academic Publishing GmbH & Co. KG, ISBN: 978-3-659-11039-9. ALBARBAR, A., GU, FENGSHOU, & BALL, ANDREW. (2010). Diesel engine fuel injection monitoring using acoustic measurements and independent component analysis. Elsevier. http://eprints.hud.ac.uk/8384/1/Gu.pdf. INTERNATIONAL CONFERENCE ON WAVELET ANALYSIS AND ITS APPLICATIONS. (2003). Proceedings of the third International Conference on Wavelet Analysis and Its Applications (WAA) Chongqing, PR China, 29-31 May 2003. [River Edge], N.J., World Scientific. http://search.ebscohost.com/login.aspx?direct=true&scope=site&db=nlebk&db=nlabk&AN=235707 YILDIRIM, A., ERKAYA, S., ESKI, A., & UZMAY, A. (2009). Noise and Vibration Analysis of Car Engines using Proposed Neural Network. Journal of Vibration and Control.  15, 133-156. YUAN, C., YAN, X., LI, Z., ZHANG, Y., SHENG, C., & ZHAO, J. (2013). Remote Fault Diagnosis System for Marine Power Machinery System. SKOBTs OV, E. A., IZOTOV, A. D., & TUZOV, L. V. (1966). Methods of reducing vibration and noise in diesel engines. Boston Spa (Yorks.), National Lending Library for Science and Technology. Read More