Mechanical equipment condition monitoring and fault diagnosis are two different levels of the same discipline, they are related yet distinct. For convenience, they are collectively referred to as mechanical equipment fault diagnosis.
Mechanical equipment fault diagnosis is a comprehensive applied science and technology that identifies the operating state of mechanical equipment (machines or units). It mainly studies how changes in the operating state of mechanical equipment are reflected in diagnostic information. Specifically, it involves measuring the operational state signals of the equipment and processing, analyzing, and extracting features from the measured signals in conjunction with its historical conditions, thus quantitatively diagnosing (identifying) the operating state (normal, abnormal, fault) of the mechanical equipment and its components, and further predicting the future operating state of the equipment to ultimately determine what necessary measures should be taken to ensure optimal operational performance of the mechanical equipment.
The main content includes monitoring, diagnosing (recognizing), and predicting the operational state of mechanical equipment.
Among them, state monitoring, also known as simple diagnosis, generally checks the operating state of the equipment by measuring certain relatively singular characteristic parameters (such as vibration, temperature, pressure, etc.), and determines whether the equipment is currently in a normal, abnormal, or fault state based on the relationship between the characteristic parameter values and threshold values.
If regular or continuous state monitoring is conducted on the equipment, trends and patterns in the changes of the equipment’s operating state can be obtained, which can then forecast the future operational development trends of the equipment, commonly referred to as trend analysis. Diagnosis (recognition) not only requires an understanding of the equipment’s operating state and development trends but is more importantly about finding the causes of faults, identifying and judging the severity of faults, and directing scientific repairs.
This is what people often refer to as precise diagnosis; the monitoring of equipment state and fault diagnosis can be understood from the following two aspects.
1. Basic Links of Equipment State Monitoring and Fault Diagnosis
The basic links can be divided into state monitoring, analysis diagnosis, and governance prevention.
The shaft, bearing, and gear are the three major components that make up the equipment drive chain. When using vibration methods for state monitoring and fault diagnosis, electric eddy current displacement sensors can be used to measure the radial vibration of the shaft in two perpendicular directions as close as possible to the bearing’s plane, and measure its axial displacement near the end face of the shaft; acceleration sensors or velocity sensors can also be used to measure the horizontal, vertical, and axial vibrations of the casing (bearing seat) at the bottom or side of the bearing seat. By monitoring, collecting, recording, analyzing, processing characteristic signals, and displaying data or graphs, comparing with judgment standards and reference graphs, the state of the mechanical equipment can be identified, and diagnostic conclusions can be made.
Based on this conclusion, choose routine inspection, monitoring operation, or shutdown repair. For diagnosed causes, locations, and severity of equipment faults, measures such as adjustment, repair, or replacement can be taken to prevent similar faults from occurring again.
2. Methods and Techniques for Monitoring and Diagnosis
1. Clear Concepts
Abnormal vibrations often indicate faults. To identify faults based on extracted vibration signal characteristics, it is necessary to analyze around parameters such as time, frequency (rotational speed), period, phase, and amplitude (energy).
In engineering, most encountered signals are dynamic, with amplitudes varying over time. Dynamic signals can be classified into deterministic periodic signals (simple harmonic signals, complex periodic signals) and non-periodic signals (quasi-periodic signals, transient signals), non-deterministic stationary random signals (various state experienced signals, non-various state experienced signals), and non-stationary random signals. These signals correspond to vibrations and are closely related to faults.
There are four classification methods for vibrations. The first is the dynamic classification method based on the essence of vibration, which can be divided into forced vibrations, transient vibrations, self-excited vibrations, and parameter-variable vibrations; the second is a rough estimation of fault locations based on the frequency of vibration; the third is a classification method that distinguishes various vibration characteristics based on signal features and provides theoretical explanations for spectrum characteristics; the fourth is a classification method based on the characteristics of the vibration system (linear vibration, nonlinear vibration).
For example, rotor imbalance and misalignment faults produce simple harmonic vibrations, which are forced vibrations under centrifugal force; when several rotor imbalance or misalignment faults exist simultaneously in a gearbox or unit, complex periodic vibrations are generated; rolling bearing element damage faults lead to decaying vibrations, which are quasi-periodic vibrations; gear tooth breakage, rotor friction, and other faults produce impact transient vibrations.
2. Clear Thinking
(1)Understand the principles, structures, and operational conditions of the test objects. Before monitoring and diagnosing, confirm whether the object being measured is a rotating machine or a reciprocating machine, whether it is a sliding bearing or a rolling bearing, the composition of the unit drive chain, and whether external excitations exist; understand the operational conditions and fault history of the object being measured, estimate possible faults and their locations, and identify which parameter changes are most sensitive.
(2)Estimate the type of vibration, vibration level, and possible minimum and maximum frequencies for the object being tested, calculate the fault frequency, and determine the analysis frequency band; select sensors and instruments or monitoring and diagnostic systems based on the unit structure and environmental conditions. For example, a non-contact electric eddy current displacement sensor can be used for sliding bearings, while velocity or acceleration sensors can be used for rolling bearings.
(3)Select appropriate monitoring and diagnostic parameters. Displacement d is very sensitive to low-frequency signals (below 100Hz), reflecting the degree of change in the working position of particles and can monitor potential damage to the equipment; velocity V is very sensitive to mid-frequency signals (below 1kHz), reflecting the speed of particle motion and can monitor kinetic energy damage to the equipment; acceleration a is very sensitive to high-frequency signals (above 1kHz), reflecting changes in force on particles and can monitor the impact force on the equipment. Although mathematically, the three quantities can be converted into each other, in practical testing, they should be selected reasonably based on each one’s suitable frequency.
(4)Draw a schematic diagram of the measurement points of the unit and a test report form.
(5)Check the testing system and determine the sensor installation positions.
(6)For the measured vibration characteristic data and graphs, repeatedly compare and analyze with judgment standards or reference graphs, eliminating factors unrelated to faults one by one from the perspective of vibration mechanisms, and make diagnostic conclusions based on field sensory diagnostic experience.
(7)Adopt corresponding governance and prevention measures to eliminate hidden dangers or faults.
3. Appropriate Methods
If there are simple diagnostic instruments that can only display the magnitude of vibration amplitude (or energy), the optimal installation method should be selected based on the frequency response range of the sensor, and measurement points should be chosen and marked. It is necessary to identify which parameter (acceleration, velocity, displacement) the displayed vibration amplitude corresponds to, whether there are range divisions (low-pass, high-pass, band-pass) and their purposes, and whether it can display dimensional (or non-dimensional) diagnostic parameter values or decibel values.
Using such instruments, diagnostic methods can be selected based on amplitude values (effective value Xrms peak value Xr, mean value), trend graphs, waveform factors (peak value Xp, mean value X), peak factors (peak value Xp, effective value), non-dimensional parameters, impulse pulses, etc., to make qualitative simple diagnostics for the unit.
If there is a rotating (reciprocating) mechanical data acquisition and processing system (precision diagnostic instrument) with multiple analysis functions, it can simultaneously test the acceleration parameter signals of four measurement points and one photoelectric signal, or simultaneously test the displacement parameter signals of four channels and one key phase pulse signal. After collecting and storing vibration signals, the instrument can provide acceleration a, velocity V, and displacement d spectrum graphs in the time domain, amplitude domain, and frequency domain, greatly facilitating the diagnostician. The analysis spectrum graphs in the three domains have different characteristics and purposes.
The time domain waveform is mainly used in the following situations:
(1) When the machine speed is below 100r/min;
(2) When the fault is relatively serious and there are obvious impacts;
(3) When the signal exhibits frequency modulation or amplitude modulation phenomena;
(4) When it is necessary to determine the exact vibration amplitude;
(5) When it is necessary to determine whether there are external interferences (random or fixed) in the signal. From the time domain waveform, the main frequency components can be roughly judged. For example, when a rotating machine has a serious imbalance fault, there are obvious periodic components in the signal characterized by the rotational frequency; when the shaft is misaligned, the second harmonic component of the rotational frequency is significantly increased within one cycle. The autocorrelation function of the time domain can extract periodic components mixed in random noise. Based on its shape, the nature of the original signal can be judged; the vibration signal of normally operating equipment often has an autocorrelation function similar to that of wideband random noise (no periodicity), while when equipment has faults, especially periodic impact faults, there will be larger peaks in the autocorrelation function at lags that are integer multiples of its period.
When performing frequency spectrum analysis in the frequency domain, the following aspects can be considered:
(1)Analyze according to the high, medium, and low frequency bands of the spectrum to understand the main fault occurrence locations; analyze according to characteristic frequencies, power frequency, superharmonics, and subharmonics to determine the range of rotor faults.
(2)Analyze the sources of main vibration components, such as a prominent power frequency component often caused by imbalance, the second harmonic is often due to misalignment or transverse cracks in the shaft, and the amplitude of the compressor surge is significantly larger than normal, with surge frequencies generally ranging from 0.5 to 50Hz.
(3)Conduct spectrum comparisons to discover faults.In amplitude domain analysis, it is necessary to combine dimensional and non-dimensional diagnostic parameters and kurtosis indicators, margin indicators, impulse indicators with stable waveform indicators and root mean square values (effective values) to balance sensitivity and stability.
4. Diligent Practice
Practice is the only standard for testing effectiveness. As a research department, it should continuously develop new applicable diagnostic methods from practice; as manufacturers of monitoring and diagnostic instruments or systems, they should develop stable performance, fully functional, easy-to-install or portable monitoring and diagnostic instruments or systems based on production site needs; engineering and technical personnel engaged in state monitoring and fault diagnosis should strive for strong support from leadership, appropriately equip fully functional precision diagnostic instruments or systems, and with tenacity and hard work, practice to organically combine monitoring and diagnostic theories, techniques, and methods with production practice to achieve results.
(Data compiled by: Hong Dachun)
—ReprintedAuthor=Meide,Share and Like=Motivation—
If you find this helpful, please share it with your friends.
Click【Read the OriginalOnline Learning