Engine running state monitoring method and system

Abstract

The invention relates to the technical field of data processing, in particular to an engine running state monitoring method and system, and the method comprises the steps: obtaining a vibration signal and a crankshaft rotating speed signal of an engine in real time, and converting the vibration signal from a time domain to an angle domain by taking the crankshaft rotating speed signal as a reference; performing order spectrum analysis on the vibration signal in the angle domain to generate a real-time order spectrum, and comparing the real-time order spectrum with a reference order spectrum to determine the energy deviation of each order; meanwhile, extracting waveform segments of each order in a plurality of continuous periodic vibration signals, and calculating the similarity between the waveform segments so as to evaluate the morphological dispersion degree and quantify the morphological stability degree; and finally, the energy deviation and the form stability degree are fused, the comprehensive fault risk index of each order is determined, and the operation state of the engine is monitored based on the comprehensive fault risk index of each order. The method improves the accuracy and reliability of engine state monitoring.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more particularly to a method and system for monitoring engine operating conditions. Background Art

[0002] As the core power source for various equipment such as industrial machinery, transportation, and power generation equipment, real-time monitoring of the engine's operating status is the core link to ensure the reliable operation of the entire system and avoid safety risks.

[0003] Vibration signals reflect the mechanical state of an engine and contain a wealth of information about fault characteristics. For example, imbalance in rotating components can cause vibrations of specific frequencies, wear in reciprocating parts can lead to abnormal fluctuations in vibration amplitude, and even early signs of tiny crack growth can be detected through distortion of the vibration waveform. Operating condition monitoring technology based on vibration signal analysis, with its non-invasive measurement advantages, has become a core method for monitoring engine operating conditions and diagnosing faults.

[0004] In the existing technology, spectrum analysis based on fast Fourier transform is the mainstream method for vibration signal processing. This method converts the time domain vibration signal into the frequency domain, first identifies the characteristic frequencies related to the fault, such as the rotation frequency and frequency harmonics of the rotating parts, and judges the fault by comparing the amplitude of the characteristic frequency with the preset threshold.

[0005] However, the traditional fast Fourier transform is based on the assumption of stationarity, which requires that the signal frequency characteristics be constant during the analysis period. However, the engine often faces dynamic speed changes in actual operating conditions, such as acceleration, deceleration and load fluctuations, which will cause the vibration signal to exhibit significant non-stationary characteristics, making the traditional fast Fourier transform method difficult to adapt. When the speed changes, the characteristic frequency that is strongly correlated with the speed will drift over time, resulting in the broadening and attenuation of the characteristic spectrum peak in the fast Fourier transform result, forming a spectrum blur, which cannot capture the true fault characteristics, resulting in missed reports or false reports, and affecting the accuracy of condition monitoring.

[0006] Therefore, there is an urgent need for a method that can adapt to the engine's variable speed and variable load conditions and realize accurate feature extraction of non-stationary vibration signals to accurately monitor the engine's operating status. Summary of the Invention

[0007] To solve the problem in the prior art that when analyzing vibration signals based on fast Fourier transform, the dynamic characteristic frequency cannot be effectively tracked, spectrum ambiguity and characteristic drift are easily generated, resulting in low accuracy of engine operating status monitoring, the present invention proposes an engine operating status monitoring method and system.

[0008] In a first aspect, the present invention provides a method for monitoring an engine operating state, comprising: Acquire the engine's vibration signal and crankshaft speed signal in real time; using the crankshaft speed signal as a reference, convert the vibration signal from the time domain to the angle domain to obtain the vibration signal in the angle domain; Perform order spectrum analysis on the vibration signal in the angle domain to generate a real-time order spectrum with each order as the horizontal coordinate and the vibration amplitude of each order as the vertical coordinate; compare the real-time order spectrum with the pre-acquired reference order spectrum to determine the energy deviation of each order in the real-time order spectrum; For any order in the real-time order spectrum, all waveform segments corresponding to the order in multiple consecutive engine cycles are extracted as all waveform segments corresponding to the order, the similarity between any two waveform segments is calculated, and the degree of morphological dispersion of all waveform segments corresponding to the order is evaluated based on the statistical characteristics of the similarity between all arbitrary two waveform segments. The degree of morphological dispersion is used as a negative indicator to quantify the degree of morphological stability of all waveform segments corresponding to the order; The energy deviation of each order in the real-time order spectrum and the morphological stability of all waveform segments corresponding to each order are integrated to determine the comprehensive fault risk index of each order. The operating status of the engine is monitored based on the comprehensive fault risk index of each order.

[0009] This technical solution cleverly converts the analysis benchmark from the volatile time dimension to the physical benchmark of the crankshaft angle synchronized with the engine cycle, and uses order analysis to fundamentally convert the non-stationary vibration signal under variable speed conditions into a stable order spectrum, avoiding the problem of inaccurate feature extraction caused by frequency drift and spectrum blurring in traditional fast Fourier transform. In addition, it abandons the fixed alarm threshold and achieves adaptive quantification of the degree of vibration energy deviation by establishing a healthy benchmark model that changes dynamically with real-time output power, reflecting the abnormality of the vibration signal under different load conditions. More importantly, it introduces the analysis of the morphological stability of the vibration signal, which is equivalent to quantifying the periodic stability of the vibration source, thereby effectively distinguishing between highly repeatable deterministic signals caused by real mechanical failures and random interference signals caused by unstable combustion or external shocks. By nonlinearly fusing the energy deviation representing the severity of the fault with the morphological stability representing the diagnostic credibility, a comprehensive fault risk index with both sensitivity and robustness is constructed. This makes the monitoring conclusion no longer a simple judgment of a single dimension. Instead, by nonlinearly fusing the energy deviation index representing the severity of the fault with the cyclic stability index representing the diagnostic credibility, a comprehensive evaluation that couples the energy significance of the fault with its periodic determinism is achieved, which enhances the ability to distinguish between real mechanical fault sources and random interference, and improves the accuracy of engine operating status monitoring.

[0010] Preferably, the vibration signal in the angle domain is determined based on the following method: The crankshaft speed signal is time-integrated to calculate the rotation angle of the crankshaft at each moment. Based on the time synchronization principle of the vibration signal and the crankshaft speed signal, the rotation angle and vibration amplitude corresponding to the vibration signal at each moment are determined. The vibration signal is divided into multiple periodic vibration signals with a rotation angle of 360 degrees as a period. The rotation angle from 0 to 360 degrees is used as the horizontal coordinate, and the average of the vibration amplitudes corresponding to each rotation angle in all periodic vibration signals is used as the vertical coordinate of the rotation angle to obtain the vibration signal in the angle domain.

[0011] This technical solution establishes a direct mapping relationship between the vibration signal and the physical rotation angle of the engine, and synchronously averages multiple engine cycles in the angle domain. This effectively filters out non-periodic random noise interference, such as unstable combustion, and enhances the deterministic mechanical vibration characteristics synchronized with the crankshaft's rotation angle.

[0012] Preferably, the real-time order spectrum is determined based on the following method, including: performing Fourier transform on the vibration signal in the angle domain to decompose it into a series of vibration modes with specific periodic characteristics, each vibration mode corresponds to the vibration characteristics of a specific engine component; the number of repetitions of each vibration mode within each 360-degree rotation of the crankshaft is used as the order of the vibration mode, and each vibration mode is characterized by each order, and the vibration amplitude of the vibration mode represented by each order is used as the vibration amplitude of the order; each order is used as the horizontal coordinate and the vibration amplitude of each order is used as the vertical coordinate to generate a real-time order spectrum.

[0013] Preferably, the energy deviation of each order in the real-time order spectrum is determined based on the following method: according to the real-time load condition of the engine, all reference order spectra under the load condition are obtained; for any order in the real-time order spectrum, the mean and standard deviation of the vibration amplitude of the order in all reference order spectra are used as the reference amplitude of the order and the fluctuation value of the reference amplitude of the order, respectively; the absolute difference between the vibration amplitude of the order and the reference amplitude, as well as the cumulative value of the reference amplitude and the fluctuation value of the reference amplitude are calculated; the ratio of the absolute difference and the cumulative value is determined as the normalized deviation rate of the vibration amplitude of the order; the ratio of the vibration amplitude of the order to the reference amplitude is determined as the energy gain factor of the order; the energy deviation of the order is determined by fusing the normalized deviation rate of the vibration amplitude of the order and the energy gain factor of the order by multiplication.

[0014] This technical solution implements a collaborative amplification mechanism. For small early deviations, the indicator shows a near-linear response. For significant fault signs, the energy gain factor acts as a powerful nonlinear amplifier, allowing the final energy deviation indicator to grow exponentially, thereby greatly improving the monitoring system's quantitative sensitivity to fault initiation and development.

[0015] Preferably, the reference order spectrum is determined based on the following method: operating a healthy engine under multiple different load conditions; under each load condition, collecting multiple vibration signals and crankshaft speed signals as reference vibration signals and reference crankshaft speed signals, respectively; determining a reference order spectrum based on each reference vibration signal and the corresponding reference crankshaft speed signal, and obtaining multiple reference order spectra under the load condition; wherein, the method for determining each reference order spectrum is consistent with the method for determining the real-time order spectrum.

[0016] Preferably, the similarity between any two waveform segments is determined based on the following method: for any two waveform segments among all waveform segments corresponding to any order, they are respectively recorded as the first waveform segment and the second waveform segment; the first waveform segment and the second waveform segment are subjected to de-averaging processing to eliminate DC offset and retain only waveform morphological features; the normalized mutual correlation coefficients of the first waveform segment and the second waveform segment after de-averaging processing are calculated at different time delays; and the peak value in the normalized mutual correlation coefficient is used as the similarity between the first waveform segment and the second waveform segment.

[0017] Preferably, the degree of morphological discreteness is based on the following method: determining the morphological difference between the first waveform segment and the second waveform segment based on the similarity between the first waveform segment and the second waveform segment, and the morphological difference is negatively correlated with the similarity; taking the first waveform segment and the second waveform segment as a combination, dividing the sum of the squares of the morphological differences corresponding to the combination by the total number of combinations to obtain an average morphological difference; and using the average morphological difference as the degree of morphological discreteness of all waveform segments corresponding to this order.

[0018] This technical solution constructs a statistical evaluation framework for pairwise comparison of all waveform morphologies of the target order in continuous cycles. By squared summing and normalizing the morphological differences, a discrete sum that depends on the total number of cycles is converted into a standardized, horizontally comparable average morphological difference, thereby providing an objective quantitative indicator for quantifying whether the vibration of this order originates from the periodic reproducibility of deterministic mechanical faults or from the instability of random interference.

[0019] Preferably, the comprehensive failure risk index of each order is determined based on the following method: For any order, the morphological stability of all waveform segments corresponding to the order is used as input, and a gating weight is generated through a nonlinear activation function transformation; the energy deviation of the order and the gating weight are weightedly fused to obtain the comprehensive fault risk index of the order.

[0020] This technical solution physically simulates expert diagnostic logic by introducing a nonlinear gated fusion mechanism. This design implements a confidence-weighted approach to fault severity assessment. Energy deviations representing fault severity are fully factored into the final risk assessment only when the system has a high degree of confidence that the vibration is caused by a persistent mechanical fault. This enables the monitoring method to distinguish between true faults and incidental interference, improving the robustness and reliability of diagnostic conclusions.

[0021] Preferably, the operating state of the engine is monitored based on the comprehensive fault risk index of each order, including: pre-acquiring the correlation between the vibration mode corresponding to each order and the specific component of the engine, as well as the normal range of the comprehensive fault risk index of each order; if the comprehensive fault risk index of a certain order exceeds the normal range of the comprehensive fault risk index of the order, it is determined that the operating state of the specific component of the engine associated with the vibration mode corresponding to the order is abnormal; if the comprehensive fault risk index of a certain order does not exceed the normal range of the comprehensive fault risk index of the order, it is determined that the operating state of the specific component of the engine associated with the vibration mode corresponding to the order is not abnormal.

[0022] In a second aspect, the present invention also provides an engine operating status monitoring system, which includes a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of any one of the engine operating status monitoring methods.

[0023] The present invention has the following effects: The present invention converts non-stationary vibration signals into stable order spectra by applying order spectrum analysis, thus overcoming the defect of spectrum distortion caused by the stationarity assumption of fast Fourier transform. On this basis, through the analysis of health benchmark and waveform stability dynamically associated with real-time load, the energy deviation and credibility of fault severity can be adaptively evaluated, thereby improving the accuracy and reliability of engine status monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic flow chart of the method of the present invention; Figure 2 is a schematic diagram of the vibration signal of the engine in the present invention; Figure 3 Schematic diagram of the crankshaft speed signal of the engine in the present invention; Figure 4 This is the spectrum diagram obtained by using traditional FFT analysis of vibration signals in the present invention; Figure 5 is a schematic diagram of a vibration signal in the angle domain of the present invention; Figure 6is a schematic diagram of the real-time order spectrum in the present invention; Figure 7 3. It is a schematic diagram comparing the morphologies of the second-order vibration mode in multiple working cycles in the present invention; Figure 8 It is a schematic diagram of the comprehensive fault risk index monitoring result finally generated in the present invention. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] The present invention provides an engine operating status monitoring method that achieves accurate and robust health status assessment of variable speed engines by combining order analysis, operating condition adaptation and cyclic statistics. Figure 1 The specific process of this method includes the following steps: S1: Acquire the engine vibration signal and the engine crankshaft speed signal in real time.

[0027] This step forms the data foundation for the entire monitoring and analysis process. To comprehensively and accurately characterize the engine's status, it is essential to simultaneously acquire multiple physical quantities reflecting its mechanical vibration, motion, and workload. The vibration signal is the core basis for diagnosing mechanical faults, the crankshaft speed signal is an essential benchmark for eliminating non-stationary effects and converting to the order domain, and output power reflects workload, enabling adaptive operation and establishing a dynamic health baseline. High-frequency accelerometers are placed on key load-bearing components of the engine, such as the cylinder head or engine block, to collect vibration signals that reflect internal mechanical shock. A photoelectric encoder or Hall effect sensor is also installed on the engine crankshaft to generate high-resolution angular pulse signals. Real-time calculation of these pulse signals determines the instantaneous rotational speed of the crankshaft, thereby generating a crankshaft speed signal. An external power sensor is used to obtain real-time engine power output, which directly reflects the engine load intensity. Furthermore, information related to the engine cycle is obtained via the on-board CAN (Controller Area Network) bus. The CAN bus is a serial communication bus designed specifically for automobiles. It allows communication between various electronic control units and sensors within the vehicle without the need for a central host, thus reducing complex wiring.

[0028] The definition of an engine cycle is based on the crankshaft rotation angle. A four-stroke engine cycle corresponds to 720° crankshaft rotation, and a two-stroke engine cycle corresponds to 360° crankshaft rotation. This angular period is fixed. When the crankshaft speed changes, the time required to complete the same angle will inevitably change. The higher the speed, the shorter the time of a single engine cycle, and the lower the speed, the longer the time.

[0029] To monitor the engine's operating status in real time, we capture the 100 engine cycles preceding any given moment. This provides sufficient historical data to reduce random interference while also reflecting recent trends. The sample size is determined based on historical experience. Vibration signals and crankshaft speed signals corresponding to these 100 engine cycles are collected and analyzed to monitor the engine's operating status. The vibration and crankshaft speed signals are collected over the same timeframe.

[0030] In a four-stroke engine, the speed of the camshaft that controls the opening and closing of the valves is half the speed of the crankshaft. That is, the crankshaft rotates twice for every camshaft rotation. For the most common four-stroke engine, one engine cycle is equal to two crankshaft rotations. Therefore, the vibration signal and crankshaft speed signal corresponding to 100 engine cycles are equal to the vibration signal and crankshaft speed signal corresponding to 200 crankshaft rotations. The vibration signal and crankshaft rotation speed signal can be used as the data basis for subsequent operations.

[0031] like Figure 2 and Figure 3 As shown, Figure 2 The waveform of the collected vibration signal becomes increasingly dense over time, showing a typical non-stationary characteristic. Figure 3 The crankshaft speed signal, acquired synchronously, rises smoothly over time with slight fluctuations, truly reflecting the engine's dynamic operation. These two synchronous, non-stationary signals are the starting point for all subsequent analysis.

[0032] S2: Convert the vibration signal from the time domain to the angle domain and generate a real-time order spectrum.

[0033] This step is the prerequisite for eliminating the influence of speed fluctuations and achieving accurate operating status monitoring. Essentially, it stems from the strong coupling between the frequency components of the vibration signal and the crankshaft motion, as well as the non-stationary nature of the vibration signal.

[0034] From a mechanical vibration perspective, almost all major engine vibrations are directly related to crankshaft rotation: The centrifugal force caused by crankshaft imbalance has a frequency that is strictly equal to the crankshaft's rotational speed. The reciprocating motion of the piston in a four-stroke engine, because it completes a power cycle every two revolutions, generates an inertial force vibration frequency that is twice the crankshaft's rotational speed. In a gear transmission system, the meshing frequency is the product of the number of gear teeth and the crankshaft's rotational speed. Even the impact vibration of valve opening and closing is proportional to the camshaft's rotational speed. The frequencies of these vibration sources do not exist in isolation; instead, they closely follow the crankshaft's rotational speed.

[0035] When the crankshaft's rotational speed is stable, the instantaneous rotational speed is constant, and the aforementioned vibration frequencies also remain constant. At this point, the vibration signal's power spectral density (which reflects the statistical characteristics of the frequency component distribution) exhibits a stable peak, consistent with the characteristics of a stationary signal (one whose statistical characteristics do not change over time). However, when the speed changes (such as during acceleration, deceleration, or idle speed fluctuations), the instantaneous rotational speed becomes a function of time, causing all vibration frequencies to become synchronous functions of time. For example, when the crankshaft speed increases from 1000 to 3000 rpm, the centrifugal force vibration frequency increases from 16.7 Hz to 50 Hz, and the inertial force vibration frequency increases from 33.3 Hz to 100 Hz.

[0036] This continuous shift in frequency components over time causes the vibration signal's statistical characteristics, such as peak position and energy distribution, to change over time. This disrupts the periodicity of the waveform, and the vibration pattern within the same time interval no longer repeats. Consequently, changes in crankshaft rotational speed alter the core frequency components of the vibration signal, causing its statistical characteristics to lose temporal stability and exhibit non-stationary characteristics. This non-stationary characteristic is not interference, but rather an inevitable consequence of the engine's mechanical motion under dynamic operating conditions.

[0037] This non-stationary characteristic introduces errors into traditional Fast Fourier Transform (FFT) analysis methods, which rely on frequency stability. When faced with frequency-drifting signals, the FFT method cannot produce sharp, clear fault spectrum peaks. Instead, it produces a phenomenon known as spectral dispersion, smearing the fault energy into a fuzzy, broadband energy blob. The direct consequence is that key fault features are diluted, the signal-to-noise ratio drops dramatically, and they are ultimately drowned out by background noise, leading to serious missed diagnoses and misjudgments.

[0038] like Figure 4 As shown in the figure, as a comparative example, a traditional fast Fourier transform (FFT) of the vibration signal reveals that the characteristic frequency associated with the speed drifts over time due to the speed change. The FFT result fails to form a clear, sharp spectral peak, but instead smears the fault energy into a fuzzy, broadband energy blob, a phenomenon known as spectrum blurring. This dilutes key fault characteristics, sharply reduces the signal-to-noise ratio, and makes accurate condition judgment impossible.

[0039] Therefore, angle domain resampling is necessary to convert the non-stationary vibration signal from the time domain to the angle domain, converting it into a pseudo-stationary signal with equal angle intervals and unaffected by speed. Only in this way can effective order analysis be performed and the true characteristics of the fault be accurately captured.

[0040] Specifically, this step realizes the dimensional conversion of the analysis benchmark through order tracking technology. First, the instantaneous rotation angle of the crankshaft is calculated, and then the angle domain is resampled to obtain the vibration signal in the angle domain. Finally, the angle domain signal is fast Fourier transformed to obtain the real-time order spectrum.

[0041] Specifically include: S21: Establish a precise mapping relationship between time and angle.

[0042] The crankshaft speed signal is time-integrated to calculate the rotation angle of the crankshaft at each moment. Based on the time synchronization principle of the vibration signal and the crankshaft speed signal, the rotation angle and vibration amplitude corresponding to the vibration signal at each moment are determined.

[0043] Since the instantaneous rotational speed of the crankshaft is the rate of change of the instantaneous rotational angle with respect to time (angular velocity), by integrating the instantaneous rotational speed over time, the total cumulative angle of rotation of the crankshaft at any moment can be obtained as the instantaneous rotational angle of the crankshaft at that moment.

[0044] For example, the crankshaft Instantaneous rotation angle at the moment for:

[0045] In this formula, It is the integral symbol, which represents the process of accumulation. is the starting time, and the integral formula represents the time from the starting time to During this period of time, each instantaneous small rotation angle is accumulated to obtain the crankshaft Instantaneous rotation angle at the moment . Indicates the speed, is the integral variable, representing the time from the start to Moment Any instant within this time period. Indicates The crankshaft speed at this moment is measured in revolutions per second. is pi, 2π is a key conversion factor here. In mathematics and physics, the standard unit of angle is radian, and one full circle corresponds to radians. The unit is revolutions per second, multiply it by After that, it is converted from "how many revolutions per second" to "how many radians per second". is the differential element of the integral.

[0046] In short, by performing definite integration of the instantaneous angular velocity over the time interval, we can achieve the The precise accumulation of tiny angular displacements occurring within the time domain. Considering that the final result of this integration is the total angle of rotation of the crankshaft from the start to any moment, a precise and continuous function mapping from the time domain to the angle domain is constructed, obtaining the crankshaft rotation angle at each moment. The instantaneous rotation angle calculated by this method is in radians, the standard unit of angle in scientific computing. For engineering practice or visualization purposes, it can be converted to degrees using the angle-to-radian conversion formula.

[0047] Then, since the vibration signal and the crankshaft speed signal are collected synchronously in time, at each moment, the vibration amplitude of the vibration signal at that moment can be obtained, and the rotation angle of the crankshaft speed signal at that moment can also be obtained. Therefore, the rotation angle and vibration amplitude corresponding to the vibration signal at each moment can be obtained.

[0048] S22: Resample to obtain a vibration signal in the angle domain.

[0049] Taking a rotation angle of 360 degrees as one cycle, which is exactly one crankshaft rotation, the vibration signal is divided into multiple periodic vibration signals. The vibration signal in step S1 is 100 engine cycles, which is a total of 200 crankshaft rotation cycles, that is, the crankshaft rotates 200 times; The rotation angle from 0 to 360 degrees (equally spaced at 1 degree) is used as the horizontal coordinate, and the average of the vibration amplitudes corresponding to each rotation angle in all periodic vibration signals is used as the vertical coordinate to obtain the vibration signal in the angle domain.

[0050] like Figure 5 As shown, the time-domain vibration signal is converted into an angle-domain vibration signal. The inherent periodicity of the vibration signal is restored, resulting in a stable form. This operation not only eliminates waveform distortion caused by speed fluctuations but also forms a stable and reliable data foundation that accurately characterizes the core mechanical health of the engine, providing the prerequisite for subsequent accurate order spectrum analysis.

[0051] This operation is a key step in transforming non-stationary time-domain signals into stable feature analysis. Its core logic lies in solving the challenge of analyzing vibration signals under variable engine speed conditions through angle resampling and synchronous averaging. First, by forcibly converting the analysis basis from the volatile time axis to the crankshaft rotation angle, which is consistent with the physics of mechanical motion, it ensures that deterministic vibration events caused by mechanical structure imbalance, wear, and other factors always recur at a fixed angular position regardless of speed fluctuations, thus creating a stable analysis domain.

[0052] On this basis, the synchronous averaging link, that is, averaging the vibration amplitudes corresponding to each rotation angle in all cycles, realizes the function of a high-efficiency statistical filter and can utilize the statistical characteristics of the signal: for periodic deterministic signals that are strongly correlated with mechanical faults, they are significantly enhanced through coherent superposition due to the strict alignment of phases (rotation angles), while for non-synchronous random noises such as random fluctuations in combustion and background shocks, they are offset by incoherent superposition in the process of obtaining the average due to the scattered phases (rotation angles).

[0053] Therefore, this operation ultimately results in a vibration signal in the angular domain with a high signal-to-noise ratio. This not only eliminates the waveform distortion and spectrum blurring caused by speed fluctuations, but also forms a stable and reliable data foundation that can accurately characterize the core mechanical health status of the engine, providing a prerequisite guarantee for subsequent accurate and correct order spectrum analysis.

[0054] S23: Perform Fourier transform on the angle domain signal to generate a real-time order spectrum.

[0055] Performing Fourier transform on the angle domain signal to obtain the real-time order spectrum fundamentally eliminates the analysis deviation caused by non-stationarity and provides a stable and clear feature carrier for subsequent fault feature extraction.

[0056] Specifically, a Fourier transform is performed on the vibration signal in the angle domain to decompose it into a series of vibration modes with specific periodic characteristics. Each vibration mode corresponds to the vibration characteristics of a specific engine component. The number of repetitions of each vibration mode within every 360-degree rotation of the crankshaft is used as the order of the vibration mode, and each vibration mode is characterized by each order, and the vibration amplitude of the vibration mode represented by each order is used as the vibration amplitude of the order. A real-time order spectrum is generated with each order as the horizontal coordinate and the vibration amplitude of each order as the vertical coordinate. Each order corresponds to a vibration mode, and the vibration amplitude of each order reflects the energy intensity of the vibration mode.

[0057] In summary, the core of performing a Fourier transform on an angular domain vibration signal is to shift the analysis basis of the vibration signal from the time domain to the order domain, thereby revealing the inherent correlation between vibration energy and the engine's rotational motion. The Fourier transform decomposes the angular domain vibration signal, which varies periodically with the crankshaft angle (e.g., 360°), into a linear superposition of a series of orders (harmonic components). The result is an energy distribution spectrum with order as the horizontal axis, rather than the frequency-energy distribution used in traditional frequency-domain analysis. Order is defined as the number of vibrations within a mechanical rotation period and is a normalized physical quantity that is independent of speed.

[0058] The advantage of this transformation is that it avoids the influence of speed changes on signal analysis. For example, the vibration mode corresponding to the first order corresponds exactly to the physical event that occurs once per crankshaft rotation. The order represents the synchronization of the frequency of this vibration mode with the crankshaft's rotational frequency. Regardless of engine speed, this event remains constant, occurring once per crankshaft rotation. Therefore, in the order spectrum, this vibration mode will always appear as a stable peak at order 1, completely eliminating the characteristic frequency drift and spectral blurring caused by speed variations in traditional Fourier transforms. Similarly, the vibration mode corresponding to order 2 precisely corresponds to a physical event that occurs twice per crankshaft rotation; its frequency is twice the crankshaft's rotational frequency, and so on.

[0059] Ultimately, the Fourier transform automatically identifies various vibration modes (such as ignition, piston movement, gear meshing, etc.) that are proportional to the engine speed, and represents these vibration modes through orders. Since order characteristics are independent of speed, during engine acceleration or deceleration, although the crankshaft rotational frequency and the vibration mode frequency are changing, their ratio (order) remains constant. This allows features that would drift and blur in the spectrum of the fast Fourier transform results to form clear, sharp, and easily identifiable spectral peaks at fixed order positions on the order spectrum. This avoids the frequency drift and blurred spectral peaks encountered by traditional fast Fourier transform spectrum analysis under variable operating conditions, ensuring the stability and reliability of the diagnostic features.

[0060] like Figure 6 As shown in the figure, the real-time order spectrum is generated after Fourier transform is performed on the signal in the angle domain. The horizontal axis is the order, which represents the multiple relationship between the vibration frequency and the crankshaft speed. The vertical axis is the amplitude, which represents the energy intensity of the order vibration mode (reflected by the vibration amplitude). The real-time order spectrum is the "vibration fingerprint" of the engine under the current working condition. Figure 6 It can be clearly seen that there is a significant energy peak at the 2nd order, while the energy at other orders is lower. The real-time order spectrum provides a clear and stable characteristic carrier for the subsequent quantification of energy deviation.

[0061] When a component fails, the energy of the corresponding order or harmonic order is significantly increased. Due to the stability of the spectral peak position, the vibration amplitude changes of specific orders can be accurately tracked even during dynamic engine acceleration or deceleration, clearly identifying the source of abnormal vibration.

[0062] In summary, the real-time order spectrum generated by Fourier transforming the angle-domain signal is essentially a fingerprint of the vibration energy distribution by order. It provides a stable, clear, and physically meaningful feature basis for engine condition monitoring and fault diagnosis under non-stationary operating conditions.

[0063] S3: Determine the energy deviation of each order in the real-time order spectrum.

[0064] Because engine vibration is not constant—for example, a healthy engine may vibrate more intensely when fully loaded (high load) than when idling (low load)—using a fixed standard to determine whether vibration is abnormal can easily lead to errors.

[0065] Therefore, to accurately and dynamically assess the engine's operating status, it's crucial to account for the variability in vibration characteristics under different operating conditions. This requires systematically operating the engine under various typical load conditions and collecting vibration signals, once the engine is confirmed to be healthy (e.g., after leaving the factory or after an overhaul). By performing order analysis on these signals, we can construct multiple baseline order spectra covering the engine's primary operating range. These baseline order spectra define the normal range of amplitudes for each order under varying engine loads—the "healthy template."

[0066] In subsequent real-time monitoring, the system accurately compares the real-time order spectrum under the current working conditions (load conditions at the current moment) with the pre-acquired reference order spectrum under the same working conditions (same load conditions), and can determine the energy deviation of each order in the real-time order spectrum, thereby providing reliable data support for accurately identifying abnormal vibrations and judging the type and degree of faults.

[0067] Specifically, they include: S31: Determine a reference order spectrum based on the engine in a healthy state.

[0068] Run a healthy engine under multiple load conditions, for example, output power of 10kW, 20kW, and 30kW, to ensure that it covers common operating conditions; Under each load condition, according to the method of step S1, a vibration signal and a crankshaft speed signal are collected every 100 engine cycles, and the signals are used as a reference vibration signal and a reference crankshaft speed signal under the load condition, respectively; Determine a reference order spectrum according to the reference vibration signal and the corresponding reference crankshaft speed signal under the load condition according to the operation in step S2, that is, the method for determining each reference order spectrum under the load condition is consistent with the method for determining the real-time order spectrum; Considering that a single measurement may be affected by accidental factors such as instantaneous oil fluctuations, ambient temperature changes, random sensor noise, etc., resulting in deviations in the collected data and failing to fully represent the stable and healthy state under the working condition, a total of 10 measurements are performed to obtain 10 benchmark order spectra under the load condition, which can effectively smooth random errors and weaken the influence of random noise.

[0069] In summary, according to this method, 10 reference order spectra of a healthy engine under each load condition can be obtained.

[0070] S32: Compare the real-time order spectrum with the reference order spectrum to determine the energy deviation of each order.

[0071] Based on the real-time load condition of the engine (the output power at the current moment), all reference order spectra of a healthy engine under the same load condition are found. For any order in the real-time order spectrum, the mean and standard deviation of the vibration amplitude of that order in all reference order spectra are used as the reference amplitude of that order and the fluctuation value of the reference amplitude of that order, respectively. For example, set the current time to At this moment, the engine is The real-time load at this moment is , The real-time order spectrum at time ,for The Order, get the healthy state of the engine under load All the benchmark order spectra under The reference amplitude of the order Hedi The fluctuation value of the reference amplitude of the order .

[0072] It is particularly noted that if the real-time load condition is between two load conditions of a healthy engine (in step S31), the reference amplitude of each order and the fluctuation value of the reference amplitude of each order under the real-time load condition can be calculated by linear interpolation or other methods to achieve coverage of the entire operating range.

[0073] Calculate the absolute difference between the vibration amplitude of the order and the reference amplitude, as well as the cumulative value of the reference amplitude and the fluctuation value of the reference amplitude; determine the ratio of the absolute difference to the cumulative value as the normalized deviation rate of the vibration amplitude of the order; determine the energy gain factor of the order as the ratio of the vibration amplitude of the order to the reference amplitude; and determine the energy deviation of the order by fusing the normalized deviation rate of the vibration amplitude of the order and the energy gain factor of the order by multiplication.

[0074] For example, Real-time order spectrum at time The The energy deviation of each order is:

[0075] In this formula, for Real-time order spectrum at time The The energy deviation of the order, For a whole character, for The The vibration amplitude of the order, For the The reference amplitude of the order, For the The fluctuation value of the base value of the vibration amplitude of the order, Is the absolute value symbol. In extreme cases, or If it is 0, set it to a very small positive number, such as , to avoid the denominator being 0 and affecting subsequent calculations.

[0076] In this formula, the first half is The normalized deviation rate of the vibration amplitude of the order reflects the The deviation between the vibration amplitude of each order and the reference value accounts for the proportion of the deviation in the normal vibration amplitude range, which realizes the adaptive normalization of the deviation. For the The absolute difference between the vibration amplitude of the order and the reference amplitude, For the The accumulated value of the base amplitude of the order and the fluctuation value of the base amplitude.

[0077] In this formula, the second half is the energy gain factor of the order, which reflects the energy gain of the first order in the real-time order spectrum. The vibration amplitude of each order is how many times the reference value is, the instantaneous energy gain, which is a nonlinear amplification factor.

[0078] By multiplying the normalized deviation rate and the instantaneous energy gain, a synergistic amplification effect is achieved. When it is only slightly exceeded, Slightly larger than When , the instantaneous energy gain is approximately equal to 1, Mainly determined by the normalized deviation rate, which reflects the and The deviation size. Significantly higher than When , not only the normalized deviation rate will become larger, but the instantaneous energy gain will also act as a powerful gain coefficient to amplify the result twice, which makes It is extremely sensitive to drastic energy jumps.

[0079] In short, quantify how much the real-time vibration energy deviates from its normal value under the current working conditions. This is a key step from qualitative comparison to quantitative evaluation. The larger the dynamic energy deviation, the more likely the engine's operating state is abnormal, and vice versa. This energy deviation calculation formula is precisely designed for variable load and non-stationary working conditions of the engine. Its core value lies in the sensitive capture of fault characteristics and adaptation to all working conditions through collaborative quantification in two dimensions. The normalized deviation rate in the first half measures the absolute deviation between the real-time amplitude and the benchmark amplitude within the framework of the benchmark value and the normal fluctuation range, and dynamically adjusts the denominator (the higher the load, the greater the normal fluctuation, The larger the value, the better) eliminates the interference of normal vibration amplitude differences under different loads, ensures the uniformity of deviation evaluation standards under different working conditions such as idling and full load, and avoids false alarms or missed alarms caused by fixed thresholds; the energy gain factor in the second half forms a nonlinear amplification mechanism through the ratio of the real-time amplitude to the reference value. When the engine has an early minor fault, the amplitude is slightly higher and the energy gain factor is close to 1. The value primarily reflects linear deviation, ensuring that even minor anomalies are not overlooked. As the fault intensifies, the amplitude increases significantly, and the gain factor is amplified simultaneously, creating a synergistic effect with the deviation rate, highlighting dramatic energy jumps. This design not only adapts to the dynamic progression of engine faults from minor to severe, but also ensures consistent assessment under different loads and orders, providing an accurate quantitative basis for subsequent fault risk assessment.

[0080] S4: Determine the morphological stability of all waveform segments corresponding to each order in the real-time order spectrum.

[0081] In engine condition monitoring, changes in order energy amplitude can be used to initially determine whether the engine's operating status is abnormal. This step further considers that occasional factors such as unstable combustion and external random impacts can also cause transient increases in vibration energy, which can easily be confused with vibration caused by actual, persistent mechanical defects such as bearing pitting and gear tooth breakage.

[0082] Therefore, by analyzing the vibration characteristics of real mechanical faults, we can see that these characteristics recur stably over continuous cycles, with their waveforms exhibiting periodic consistency. By quantifying the morphological stability of the waveform segments corresponding to each order in the real-time order spectrum, we can provide continuous verification of energy deviations, effectively distinguishing between sporadic interference and real mechanical vibrations, and enhancing the reliability of operational status monitoring.

[0083] The specific process of this step is as follows: for any order in the real-time order spectrum, first extract all waveform segments corresponding to this order in multiple consecutive engine cycles; then, remove the mean value of any two waveform segments to retain the morphological characteristics, and determine the similarity between the two by calculating the peak value of the normalized cross-correlation coefficient under different time delays; then, based on the similarity between all waveform segments, convert the similarity into a negatively correlated morphological difference, and calculate the average morphological difference by averaging the sum of the squares of all combined morphological differences to evaluate the morphological dispersion of the waveform segment of this order; finally, use the morphological dispersion as a negative indicator; the higher the dispersion, the lower the stability, to quantify the morphological stability of the waveform segment corresponding to this order.

[0084] Specifically, they include: S41: extracting all waveform segments corresponding to each order in the real-time order spectrum.

[0085] Since step S2 is to obtain multiple periodic vibration signals with a period of 360 degrees of crankshaft rotation by performing angle domain conversion on the vibration signal at the current moment, on this basis, the periodic vibration signal is subjected to frequency domain analysis through Fourier transform, and each order and its corresponding characteristic parameters (such as vibration amplitude) are extracted from the vibration characteristics. This is to determine the order based on the vibration and obtain a real-time order spectrum.

[0086] In this step, the process of extracting all waveform segments corresponding to each order in the real-time order spectrum follows the technical logic of reverse positioning, specifically: Based on each order in the real-time order spectrum, the waveform segment matching the order (corresponding vibration mode) is reversely separated from each periodic vibration signal divided in step S2 using methods such as bandpass filtering, thereby obtaining the waveform segment of the order within a single period. Finally, all waveform segments extracted from all periodic vibration signals of the order are aggregated to form all waveform segments corresponding to the order.

[0087] This operation of first analyzing the order through the vibration signal and then reversely locating the waveform segment based on the order is a conventional and reasonable technical means used in order analysis to achieve refined feature extraction. The logic is self-consistent and conforms to the basic laws of signal processing.

[0088] like Figure 7The figure below intuitively illustrates all waveform segments corresponding to the second-order vibration mode extracted from five consecutive cycles of vibration signals. The five curves correspond to the waveform changes of the vibration mode corresponding to the second-order vibration mode within the five cycles of the vibration signal. As can be seen, these five curves overlap to a large extent, with highly consistent overall trends, yet slight differences exist. This is consistent with the physical characteristics of real, persistent mechanical faults, which recur periodically but are susceptible to minor perturbations. By calculating the similarity between these waveform segments, their morphological stability can be quantified, providing a key basis for subsequently distinguishing faults from interference.

[0089] S42: Calculate the similarity between any two waveform segments corresponding to each order.

[0090] After obtaining all waveform segments corresponding to each order, calculating the similarity between any two waveform segments is a crucial prerequisite for quantifying morphological stability and is irreplaceable. This is because waveform segment similarity directly reflects the morphological consistency of the same-order vibration characteristics across different cycles. For persistent mechanical faults, the vibration waveform will recur stably over consecutive cycles, and the similarity between any two waveform segments should be significantly higher. For occasional disturbances (such as unstable combustion or random shocks), the waveform morphology will exhibit low similarity due to irregular fluctuations.

[0091] By quantifying the similarity, the subjective judgment of "whether the morphology is consistent" can be converted into objective data, providing a basic basis for the subsequent calculation of morphological differences and evaluation of discreteness, and ultimately achieving accurate distinction between stably recurring fault characteristics and random interference signals.

[0092] Specifically: For any two waveform segments among all waveform segments corresponding to any order, they are respectively recorded as the first waveform segment and the second waveform segment; the first waveform segment and the second waveform segment are subjected to de-averaging processing to eliminate DC offset and retain only the waveform morphological characteristics; the normalized cross-correlation coefficients of the de-averaged first waveform segment and the second waveform segment at different time delays are calculated; and the peak value of the normalized cross-correlation coefficient is used as the similarity between the first waveform segment and the second waveform segment.

[0093] For example, specifically, Among all waveform segments corresponding to the order, the normalized correlation coefficient between any two waveform segments is:

[0094] In this formula, For the The order corresponding to the Waveform segment and The normalized cross-correlation coefficient of the waveform segments, For the Waveform segment and The time sequence number of each waveform segment, The value range is , For the Waveform segment and The total number of moments of the waveform segment, due to the Waveform segment and The time lengths of the waveform segments are the same, then The moment will get The vibration amplitude of the first waveform segment at this moment and the The vibration amplitude of each waveform segment at that moment. For the The order corresponding to the The waveform segment is in The vibration amplitude at a moment, For the The order corresponding to the The average value of the vibration amplitude of the waveform segment at all times, For the The order corresponding to the The waveform segment is in The vibration amplitude at a moment, The time delay ranges from 0.001 to 0.01. For the The order corresponding to the The average value of the vibration amplitude of a waveform segment at all times.

[0095] The numerator is the quantitative carrier of trend consistency. It directly reflects the trend consistency of the vibration pattern by calculating the product sum of the two waveform segments after removing the mean. and Representing the 、 The waveform segment at time The fluctuation component that deviates from its own mean (and the corresponding moment after a delay of τ0). When the two waveform segments have similar shapes (simultaneous peaks and decays), the product of the two fluctuation components is positive, and the sum is a large positive number. When the shapes are opposite (one rises and the other falls), the product of the two fluctuation components is negative, and the sum is a small negative number. When the shapes are irregular (one has a clear cycle and the other is chaotic), the positive and negative products cancel each other out, and the result approaches 0. This design accurately captures the essential difference between the stable recurrence of persistent fault waveforms and the random fluctuations of occasional interference waveforms.

[0096] The denominator is a normalization guarantee to eliminate amplitude differences. As a normalization factor, the denominator converts the "absolute correlation" of the numerator into a "relative coefficient" by taking the square root of the product of the fluctuation energy of the two waveform segments. This process solves the problem that "amplitude differences do not represent morphological differences" in engine vibration. For example, the waveform amplitude of the same order at high load may be greater than that at low load. However, as long as the morphology is consistent (such as the peak position and fluctuation rhythm), a high similarity result can still be obtained through the scaling effect of the denominator. Ultimately, Strictly limited to Within the range, it is ensured that the similarity results under different orders and different loads are directly comparable.

[0097] because is the time delay, and its value ranges from 0.001 to 0.01. Each time delay can calculate a , all Peak As the first The order corresponding to the Waveform segment and The normalized cross-correlation coefficient of the waveform segments is Waveform segment and The time alignment method that makes the two waveform segments most similar.

[0098] The introduction of time delay is an engineering optimization for small fluctuations in engine speed. Since the duration of different cycles may vary slightly due to speed fluctuations, time delay adjustment can achieve precise alignment of waveform segments, avoid misjudging morphological differences due to time axis misalignment, and accurately reflect inherent similarities. For example, when two waveform segments have the same actual morphology but a time difference of 0.002 seconds, By delaying the waveform by 0.002 seconds, the interference of the duration difference can be eliminated and the similarity of the waveform segment itself can be focused.

[0099] when When Waveform segment and The more similar the shapes of the two waveform segments are, the more consistent the fluctuation trends and peak positions are. The more similar the shapes of the two waveform segments are, the more consistent the fluctuation trends and peak positions are. The vibration mode corresponding to the order shows stable and repetitive characteristics in these two waveform segments, which is consistent with the "periodic recurrence" characteristics of persistent mechanical failures; when Or when it is a negative value, it means The order corresponding to the Waveform segment and The shapes of the waveform segments are very different, indicating that The vibration mode corresponding to each order is unstable in these two waveform segments, with irregular shapes or opposite trends, and is more likely to be caused by occasional interference such as unstable combustion and random impact. This quantitative result provides an objective basis for the subsequent calculation of morphological differences and evaluation of overall stability.

[0100] S43: Determine the degree of morphological discreteness of all waveform segments corresponding to each order.

[0101] Calculating the degree of morphological dispersion of all waveform segments corresponding to each order is crucial for distinguishing mechanical failures from occasional interference and is therefore crucial. This is because the similarity of individual waveform segments can reflect the morphological correlation between them, while the degree of morphological dispersion, through statistical analysis of the similarity of all waveform segments, can be used to comprehensively assess the stable reproducibility of the vibration mode of that order.

[0102] For persistent mechanical faults, the vibration waveform should be highly consistent over successive cycles, with low overall dispersion. For sporadic disturbances, the waveform is disorganized and exhibits significantly higher dispersion. By quantifying the degree of dispersion, we can transform the local similarities of multiple waveform segments into a global stability indicator, providing a decisive basis for determining the nature of the vibration anomaly and effectively improving the accuracy of fault diagnosis.

[0103] Therefore, the process of determining the degree of discreteness in this step based on the statistical characteristics of the morphological similarity of all waveform segments is as follows: first, the similarity between the two waveform segments is converted into a negatively correlated morphological difference, and then the average value of the sum of the squares of the morphological differences of all combinations is calculated to obtain the average morphological difference, which is used as the morphological discreteness of all waveform segments of this order.

[0104] Specifically, for the first waveform segment and the second waveform segment in all waveform segments corresponding to any order (step S42), the morphological difference between the first waveform segment and the second waveform segment is determined based on the similarity between the first waveform segment and the second waveform segment, and the morphological difference is negatively correlated with the similarity; the first waveform segment and the second waveform segment are taken as a combination, and the sum of the squares of the morphological differences corresponding to the combination is divided by the total number of combinations to obtain an average morphological difference; the average morphological difference is used as the morphological discreteness of all waveform segments corresponding to the order.

[0105] For example, the real-time order spectrum The degree of morphological discreteness of all waveform segments corresponding to the order is calculated based on the following formula:

[0106] In this formula, is the first order in the real-time order spectrum The discrete degree of the order in M ​​continuous vibration signals quantifies the order from a global perspective. The degree of morphological discreteness of all waveform segments corresponding to the order. The smaller the value, the higher the order. The higher the reproducibility of the vibration mode corresponding to the order, the more stable it is. For the The order corresponding to the Waveform segment and The morphological similarity of the waveform segments.

[0107] In this formula, the numerator is the cumulative quantification of the total discrete degree, first through Morphological similarity is converted into morphological difference. Each pair of waveform segments is grouped together, and the square root of the square root of the square root of the square root of the square root of the square root is taken. This squaring operation amplifies the impact of significant differences (such as large morphological deviations caused by occasional interference) while suppressing the interference of minor fluctuations (such as slight vibration deviations during normal operation). This ensures that the calculation of the total dispersion level is more consistent with the requirement for determining "stable and reproducible fault characteristics." The square root operation aligns the resulting dimension with the morphological difference, facilitating an intuitive understanding of the overall dispersion level.

[0108] The denominator is a normalization calibration, which is the total number of combinations of all waveform segments. Its function is to convert the total dispersion to an average level. This process eliminates the influence of sample size differences. Whether the number of waveform segments is 5 or 500, the final result reflects the average dispersion of any two waveform segments. This makes the dispersion of different orders and different sample sizes directly comparable. For example, for the same total dispersion, a larger sample size will reduce the average dispersion, which better reflects the characteristics of stable reproduction.

[0109] In short, The smaller it is, the smaller the average morphological difference of all waveform segments of this order is, and the more reproducible the vibration mode is in the continuous cycle, which is consistent with the characteristics of persistent faults such as bearing pitting and gear tooth breakage. On the contrary, The larger the value, the more chaotic and irregular the waveform is, and is more likely to be caused by occasional interference such as unstable combustion and external random impact. This provides a core indicator for subsequent evaluation of morphological stability, transforming stable reproduction from a qualitative description into directly comparable quantitative data.

[0110] S44: Determine the morphological stability of all waveform segments corresponding to each order.

[0111] The morphological stability of all waveform segments corresponding to each order is a key step in converting discreteness into an intuitive diagnostic indicator. While morphological discreteness reflects the magnitude of waveform differences, morphological stability directly links these differences to "fault reproducibility" through a clear functional relationship. For operational status monitors, the stability indicator provides a more direct way to determine whether the vibration pattern meets the characteristics of a persistent fault. Higher stability indicates greater reproducibility of the vibration pattern in continuous cycles, and a higher confidence level for the fault. Conversely, lower stability indicates a more likely case of sporadic interference. This quantification process transforms abstract morphological consistency into a directly applicable diagnostic basis.

[0112] The process of determining the degree of morphological stability is as follows: using the degree of morphological dispersion as a negative indicator, a function that can reflect the negative correlation between the two (such as an exponential function) is used to quantify the morphological stability of all waveform segments corresponding to each order, thereby obtaining a more intuitive indicator in which the larger the value, the more stable it is.

[0113] For example, The morphological stability of all waveform segments corresponding to the order satisfies the following relationship:

[0114] In this formula, For the The morphological stability of all waveform segments corresponding to the order, is a natural constant, is the first order in the real-time order spectrum The degree of morphological discreteness of all waveform segments corresponding to the order, It is an adjustment coefficient greater than 0, with an empirical value of 0.5, which is used to adapt to the differences in vibration characteristics of different orders. For example, for low orders such as 1st order and 2nd order related to core components, the dispersion of their normal vibration should be extremely low, so it can be appropriately increased. (such as 0.6) to increase the sensitivity to small discreteness; for high orders, because the vibration mode is more complex, the normal discreteness is slightly higher, which can be reduced The default value of 0.5 is set to ensure the horizontal comparability of the stability of different orders, so that diagnosticians can evaluate the stability differences of different orders under the same standards.

[0115] In short, the use of exponential function to construct the negative correlation between the degree of dispersion and the degree of stability is based on the characteristic of "nonlinear change of fault characteristics from stable to unstable". When it approaches 0 (waveform height is consistent), It approaches 1, which directly reflects the characteristics of high stability, which is consistent with the vibration law of continuous faults such as bearing pitting and gear tooth breakage. Increases, that is, the waveform difference becomes larger, when It decays exponentially, rapidly approaching zero, accurately capturing the "stability drop" phenomenon caused by occasional interference. This nonlinear transformation better reflects the characteristics of actual fault development. Small changes in morphological dispersion have little impact on stability, while significant increases in morphological dispersion lead to a sharp drop in stability. This facilitates the distinction between acceptable normal fluctuations and alarming abnormal morphological dispersion.

[0116] at last, The value range strictly falls within Within the range, the larger the value, the higher the stability of the morphology: When , it indicates that the vibration waveform of this order is highly consistent in continuous cycles, which is a typical feature of periodic recurrence of mechanical failure and requires special attention; when When the waveforms differ significantly, it's more likely due to occasional interference, such as unstable combustion or external shocks, and the priority of attention can be reduced. This intuitive quantitative result provides a clear decision-making basis for fault diagnosis, transforming the key characteristic of stable reproducibility from an abstract description into a quantifiable and comparable indicator.

[0117] S5: Determine the comprehensive failure risk index of each order by combining energy deviation and morphological stability.

[0118] Calculating the comprehensive fault risk index of each order is a key step in achieving the transition from feature quantification to fault decision-making. Energy deviation can only reflect the degree of abnormality of vibration energy, but cannot distinguish whether its source is a real fault or occasional interference; morphological stability can only reflect the reproducibility of the vibration mode, but cannot explain the severity of the fault. Both of them have limitations when used alone. If based solely on energy deviation, occasional interference such as unstable combustion may be misjudged as a fault; if relying solely on morphological stability, it is difficult to quantify the actual impact of the fault. By integrating the two, a coordinated assessment of severity and credibility can be achieved, ensuring that only when the energy is significantly abnormal and the morphology is stably reproducible will it be judged as high risk, effectively improving the accuracy and reliability of fault diagnosis, and providing an accurate decision-making basis for engine status monitoring.

[0119] Therefore, for any order, this step takes the morphological stability of all waveform segments corresponding to the order as input, and generates a gating weight through a nonlinear activation function transformation; the energy deviation of the order and the gating weight are weightedly fused to obtain the comprehensive fault risk index of the order.

[0120] Specifically, using The function is used as a nonlinear activation function, and the fusion is realized based on the following formula to obtain the comprehensive fault risk index of each order:

[0121] In this formula, It is The comprehensive failure risk index of the order, It is The energy deviation of the order represents the severity of the fault. is a natural constant, It is The morphological stability of all waveform segments corresponding to the order, is the activation threshold of the morphological stability, which is set to 0.8. This value is determined based on the statistics of the engine vibration characteristics. When , it shows that the waveform is highly consistent in the continuous cycle, which is consistent with the characteristics of periodic recurrence of mechanical failure. At this time, the gate weight begins to increase significantly; when When , the morphological stability is insufficient, the gating weight is low, and the interference of unstable features is effectively filtered out.

[0122] The core of the formula is to generate gating weights through the Sigmoid function. The second half of the formula is a typical Sigmoid function. The risk indicated by the energy deviation will only be recognized when the morphological stability is high enough. This design strictly follows the diagnostic logic that real faults must meet the requirements of energy anomaly and morphological stability at the same time. Mechanical faults (such as bearing pitting) will also manifest as large energy deviations. High and high degree of morphological stability High, the gate weight approaches 1. Approximately equal to , accurately reflects high risk; while occasional interference (such as unstable combustion) may cause large energy deviation, but the morphological stability is low, Low, the gating weight approaches 0, It is significantly suppressed to avoid misjudgment.

[0123] at last, The value of directly reflects the fault risk level corresponding to the order: When it is significantly higher (e.g., much higher than the historical normal threshold), it indicates that the order has significant energy anomalies and highly stable morphology, which is likely to be a mechanical structure failure and requires immediate attention. When it is low, it may be due to small energy deviation (normal state) or insufficient morphological stability (occasional interference), and the risk is low.

[0124] In short, this quantitative result simplifies complex multi-feature analysis into an intuitive risk indicator. It not only retains the characterization of fault severity by energy deviation, but also filters out unreliable interference signals through morphological stability, allowing the engine condition monitoring system to focus more accurately on real mechanical faults, significantly improving the efficiency and reliability of diagnostic decisions.

[0125] S6: Monitor the engine's operating status based on the comprehensive fault risk index of each order.

[0126] This step is the final output link of the entire monitoring process, which realizes real-time monitoring of the engine operating status through the comprehensive fault risk index of all orders in the real-time order spectrum.

[0127] Specifically, they include: S61: By analyzing the relationship between the order and the kinetic characteristics of the engine.

[0128] Each order corresponds to a specific vibration mode. The essence of the order is the "ratio of vibration frequency to crankshaft speed", which directly corresponds to the periodic vibration mode of specific components of the engine: low orders (such as 1st, 2nd and 3rd orders) are mostly related to the core motion system. For example: 1st order corresponds to the vibration mode generated by rotating parts such as crankshaft / flywheel every time it rotates one circle; 2nd order corresponds to the vibration mode generated by reciprocating parts such as piston / connecting rod every 2 circles; 3rd order corresponds to the vibration mode generated by gear meshing every 3 times per rotation, and the vibration mode of superposition of multi-cylinder combustion pulses.

[0129] Therefore, orders 1-3 serve as core indicators for engine vibration analysis. Their corresponding vibration excitation sources are directly related to fundamental moving components such as rotating systems, reciprocating systems, and core accessories. Early failures of these components (such as crankshaft imbalance, piston sticking, and water pump wear) can be clearly revealed through low-order vibration anomalies. These anomalies not only minimize signal interference and provide clear characteristics, facilitating rapid locating of the source of the fault, but also cover the vast majority of basic fault types. While high-order vibrations can reflect faults, they are often associated with the superposition of complex vibrations, such as high-frequency coupling of multi-cylinder combustion pulses and high-order resonances of components. Interpretation requires the integration of low-order anomalies, serving primarily as a supplemental analysis. Therefore, focusing on orders 1-3 can both efficiently capture early fault signals and achieve a balance between practicality and complexity.

[0130] Therefore, the relationship between each order (corresponding vibration mode) and the specific components of the engine is established as follows: 1st order: corresponds to rotating parts, including crankshafts and flywheels. When the 1st order vibration mode is abnormal, it usually means that the engine's rotating system is unbalanced (such as crankshaft bending and flywheel imbalance), indicating that the core rotating parts have eccentric vibrations during operation.

[0131] 2nd order: corresponds to reciprocating components, including pistons, fuel injectors, etc. When the 2nd order vibration mode is abnormal, it usually means that the reciprocating system is stuck or the combustion is uneven (such as the gap between the piston and the cylinder liner is too large, or the fuel injector is abnormal). This indicates that the reciprocating components have impact vibration during operation or the combustion stability is reduced. 3rd order: corresponds to other accessories, including water pumps, gears, etc. When the 3rd order vibration mode is abnormal, it usually means that the accessories are unbalanced or poorly meshed (such as water pump bearing wear, gear tooth surface damage), indicating that there is high-frequency abnormal friction in the accessory system during operation.

[0132] S62: Determine the normal range of the comprehensive failure risk index of each order.

[0133] In advance, the vibration signal and crankshaft speed signal of at least 1000 engine cycles (10 times the number in step S1) are collected within the full operating range of normal engine operation (no faults) (including idle, part load, and full load). The vibration signal and crankshaft speed signal are divided into 10 segments, with 100 engine cycles as one segment, to obtain 10 segments of vibration signal and 10 segments of crankshaft speed signal.

[0134] For each vibration signal segment, the real-time order spectrum of the vibration signal segment and the comprehensive fault risk index of each order in the real-time order spectrum of the vibration signal segment are calculated according to the method of steps S1 to S6. The orders in the real-time order spectra of these 10 vibration signals are the same.

[0135] For each order, calculate the mean and standard deviation of the comprehensive fault risk index of the order in 10 segments of vibration signals. According to the principle of three times the standard deviation in statistics, the normal range of the comprehensive fault risk index of the order is determined by using the mean ± 3 times the standard deviation. For example: order The average value of the comprehensive fault risk index in these 10 vibration signals is , the standard deviation is , then its normal range is According to this method, the normal range of the comprehensive fault risk index of each order in the real-time order spectrum of the vibration signal of the engine in a healthy state can be obtained.

[0136] S63: Monitor the operating status of the engine based on the comparison results of the comprehensive fault risk index of each order with the normal range.

[0137] The comprehensive fault risk index of each order in the real-time order spectrum obtained at the current moment is compared with its normal range. If the comprehensive fault risk index of a certain order exceeds the normal range of the comprehensive fault risk index of that order, it is determined that the operating state of the specific engine component associated with the vibration mode corresponding to that order is abnormal. If the comprehensive fault risk index of a certain order does not exceed the normal range of the comprehensive fault risk index of that order, it is determined that the operating state of the specific engine component associated with the vibration mode corresponding to that order is not abnormal. For example, if the comprehensive fault risk index of order 1 exceeds the normal range, it is determined that the operating state of the engine's rotating component is abnormal.

[0138] like Figure 8As shown, the final output of the entire process of the present invention shows the comprehensive fault risk index obtained after fusing energy deviation and morphological stability in step S5. Because the second-order vibration has both high energy deviation and high morphological stability, its final calculated comprehensive fault risk index significantly exceeds the preset upper limit of the normal range, while the comprehensive fault risk indices of other orders are within the normal range. This result clearly points to the abnormal operating state of the reciprocating components related to the second order, verifying that the method of the present invention can accurately and reliably monitor the operating state of the engine.

[0139] The present invention also provides an engine operating status monitoring system, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement operations such as steps S1 to S6, thereby accurately monitoring the operating status of the engine.

[0140] In summary, the present invention overcomes the defects of the prior art in processing variable speed engine signals through a complete and logically rigorous technical solution, and provides a new paradigm for condition monitoring that is more accurate, intelligent and reliable.

[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring engine operating status, characterized in that: include: Acquire the engine's vibration signal and engine crankshaft speed signal in real time; Based on the crankshaft speed signal, the vibration signal is converted from the time domain to the angle domain to obtain the vibration signal in the angle domain; Perform order spectrum analysis on the vibration signal in the angle domain to generate a real-time order spectrum with each order as the horizontal axis and the vibration amplitude of each order as the vertical axis; Comparing the real-time order spectrum with a pre-acquired reference order spectrum to determine the energy deviation of each order in the real-time order spectrum; For any order in the real-time order spectrum, all waveform segments corresponding to the order in multiple consecutive engine cycles are extracted as all waveform segments corresponding to the order, the similarity between any two waveform segments is calculated, and the degree of morphological dispersion of all waveform segments corresponding to the order is evaluated based on the statistical characteristics of the similarity between all arbitrary two waveform segments. The degree of morphological dispersion is used as a negative indicator to quantify the degree of morphological stability of all waveform segments corresponding to the order; The energy deviation of each order in the real-time order spectrum and the morphological stability of all waveform segments corresponding to each order are integrated to determine the comprehensive fault risk index of each order. The operating status of the engine is monitored based on the comprehensive fault risk index of each order.

2. The engine operating status monitoring method according to claim 1, characterized in that: The vibration signal in the angle domain is determined based on the following method: Perform time integration on the crankshaft speed signal to calculate the rotation angle of the crankshaft at each moment. Based on the time synchronization principle of the vibration signal and the crankshaft speed signal, determine the rotation angle and vibration amplitude corresponding to the vibration signal at each moment. With a rotation angle of 360 degrees as one period, divide the vibration signal into multiple periodic vibration signals. The rotation angle from 0 to 360 degrees is used as the horizontal coordinate, and the average of the vibration amplitudes corresponding to each rotation angle in all periodic vibration signals is used as the vertical coordinate of the rotation angle to obtain the vibration signal in the angle domain.

3. The engine operating status monitoring method according to claim 1, characterized in that: The real-time order spectrum is determined based on the following methods, including: Performing Fourier transform on the vibration signal in the angle domain to decompose it into a series of vibration modes with specific periodic characteristics, where each vibration mode corresponds to the vibration characteristics of a specific engine component; The number of repetitions of each vibration mode within each 360-degree rotation of the crankshaft is taken as the order of the vibration mode, and each vibration mode is characterized by each order, and the vibration amplitude of the vibration mode represented by each order is taken as the vibration amplitude of the order; The real-time order spectrum is generated with each order as the horizontal axis and the vibration amplitude of each order as the vertical axis.

4. The engine operating status monitoring method according to claim 1, characterized in that: The energy deviation of each order in the real-time order spectrum is determined based on the following method: According to the real-time load condition of the engine, all reference order spectra under the load condition are obtained; for any order in the real-time order spectrum, the mean and standard deviation of the vibration amplitudes of the order in all reference order spectra are used as the reference amplitude of the order and the fluctuation value of the reference amplitude of the order, respectively; Calculate the absolute difference between the vibration amplitude of the order and the reference amplitude, as well as the cumulative value of the reference amplitude and the fluctuation value of the reference amplitude; The ratio of the absolute difference and the accumulated value is determined as the normalized deviation rate of the vibration amplitude of the order; the ratio of the vibration amplitude of the order and the reference amplitude is determined as the energy gain factor of the order; the energy deviation of the order is determined by fusing the normalized deviation rate of the vibration amplitude of the order and the energy gain factor of the order by multiplication.

5. The engine operating status monitoring method according to claim 1, characterized in that: The base order spectrum is determined as follows: Run a healthy engine under multiple load conditions; Under each load condition, a plurality of vibration signals and a crankshaft speed signal are collected as a reference vibration signal and a reference crankshaft speed signal respectively; Determine a reference order spectrum according to each reference vibration signal and the corresponding reference crankshaft speed signal, and obtain multiple reference order spectra under the load condition; The method for determining each reference order spectrum is consistent with the method for determining the real-time order spectrum.

6. The engine operating status monitoring method according to claim 1, characterized in that: The similarity between any two waveform segments is determined based on the following: For any two waveform segments among all waveform segments corresponding to any order, they are respectively recorded as the first waveform segment and the second waveform segment; performing a de-averaging process on the first waveform segment and the second waveform segment to eliminate a DC offset and retain only waveform morphological features; Calculating normalized cross-correlation coefficients of the first waveform segment and the second waveform segment after the mean value removal process at different time delays; and taking a peak value of the normalized cross-correlation coefficient as the similarity between the first waveform segment and the second waveform segment.

7. The engine operating status monitoring method according to claim 6, characterized in that: The degree of morphological dispersion is based on the following method: determining a morphological difference between the first waveform segment and the second waveform segment based on the similarity between the first waveform segment and the second waveform segment, wherein the morphological difference is negatively correlated with the similarity; Taking the first waveform segment and the second waveform segment as a combination, dividing the sum of the squares of the morphological differences corresponding to the combination by the total number of combinations to obtain an average morphological difference; The average morphological difference is taken as the morphological discreteness of all waveform segments corresponding to the order.

8. The engine operating status monitoring method according to claim 1, characterized in that: The comprehensive failure risk index of each order is determined based on the following method: For any order, the morphological stability of all waveform segments corresponding to the order is used as input, and a gating weight is generated through a nonlinear activation function transformation; the energy deviation of the order and the gating weight are weightedly fused to obtain the comprehensive fault risk index of the order.

9. The engine operating status monitoring method according to claim 1, characterized in that: Monitor the engine's operating status based on comprehensive fault risk indices of various orders, including: Pre-obtain the correlation between the vibration mode corresponding to each order and the specific components of the engine, as well as the normal range of the comprehensive failure risk index of each order; If the comprehensive fault risk index of a certain order exceeds the normal range of the comprehensive fault risk index of the order, it is determined that the operating state of a specific component of the engine associated with the vibration mode corresponding to the order is abnormal; If the comprehensive fault risk index of a certain order does not exceed the normal range of the comprehensive fault risk index of the order, it is determined that the operating state of the specific component of the engine associated with the vibration mode corresponding to the order is not abnormal.

10. An engine operating status monitoring system, characterized in that: The engine operating state monitoring system includes a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the engine operating state monitoring method according to any one of claims 1 to 9.

Images