An engine fault detection method and system

Abstract

The present application relates to the technical field of data processing, in particular to an engine fault detection method and system, the method comprising: determining a standardized energy deviation of a current working cycle by a working condition similarity of the current working cycle and a plurality of historical reference working cycles and vibration energy of vibration signals of each historical reference working cycle during engine operation; decomposing the current working cycle into a power stroke and a compression stroke, and determining a stroke asymmetry index according to an angular velocity fluctuation degree of the power stroke and an angular velocity fluctuation degree of the compression stroke; determining an instantaneous risk index of the current working cycle by integrating the standardized energy deviation and the stroke asymmetry index; and performing time series accumulation analysis on the instantaneous risk index to detect whether the engine has a fault. The method improves the accuracy and reliability of engine fault detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to an engine fault detection method and system. BACKGROUND

[0002] As the core component of various power equipment, the reliability and safety of engine operation is of great importance. Online fault diagnosis technology can monitor the health status of the engine in real time during operation, and provide early warning for early faults, so as to avoid catastrophic accidents and reduce maintenance costs. Among many diagnosis technologies, fault analysis based on vibration signals is widely used due to its non-invasive and rich information.

[0003] In the prior art, a common engine fault detection method is a fixed threshold method based on vibration signal energy. This method collects engine cylinder vibration signals through a vibration sensor, takes a single working cycle of the engine as an analysis unit, calculates the energy or other statistical characteristics of the vibration signal in each cycle, and compares it with a pre-set fixed fault threshold. When the calculated characteristic value exceeds the threshold for many times in succession, the system determines that the engine has failed.

[0004] However, there are obvious defects in actual application. The operating conditions of the engine are dynamically changing, especially in the scenarios of vehicle acceleration or deceleration or sudden load change of the generator set. Changes in operating conditions will cause a sharp fluctuation in the normal vibration level of the engine. For example, when the engine switches from low-speed idle condition to high-speed high-load condition, the healthy vibration energy of the engine itself will increase significantly.

[0005] At this time, the fixed threshold method is easy to misjudge the vibration enhancement caused by normal condition change as a fault, thereby generating a large number of false alarms. This high false alarm rate seriously reduces the credibility and practicality of the diagnosis system, making it impossible for the operator to effectively distinguish between real faults and normal condition fluctuations. The fundamental reason is that this method uses a static evaluation standard to measure a dynamically changing system, and fails to establish an internal correlation between the vibration characteristics and the real-time operating state of the engine. SUMMARY

[0006] To solve the technical problems that the detection method based on the fixed threshold in the prior art cannot adapt to the dynamic change of the engine operating conditions and is prone to false alarms, the present application provides an engine fault detection method and system.

[0007] In a first aspect, the present application provides an engine fault detection method, comprising:

[0008] Take any one working cycle in the engine running process as a current working cycle, and determine a plurality of historical reference working cycles; the current working cycle / historical reference working cycle includes the vibration signal and the crank angle of the engine;

[0009] According to the working condition similarity of each historical reference working cycle and the current working cycle, the vibration energy of the vibration signal of each historical reference working cycle is weighted and calculated to determine the dynamic reference energy of the vibration signal of the current working cycle; in response to the comparison result of the energy and the dynamic reference energy of the vibration signal of the current working cycle, the standardized energy deviation of the current working cycle is determined;

[0010] According to the crank angle of the current working cycle, the current working cycle is decomposed into a working stroke and a compression stroke, and the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke are determined respectively;

[0011] Fusion of the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke, determine a stroke asymmetry index for representing the dynamic balance of the engine internal working and force accumulation process;

[0012] The instantaneous risk index of the current working cycle is determined by comprehensively considering the standardized energy deviation and the stroke asymmetry index, and the time series cumulative analysis is carried out on the instantaneous risk index to judge whether the engine has a fault.

[0013] The technical scheme firstly constructs a dynamic reference energy which is adjusted in real time according to the working condition at the reference setting level, and completely discards the traditional fixed threshold. It makes the evaluation of the vibration energy no longer based on a static absolute value, but is transformed into a relative value which is accurately matched with the current engine state. This dynamic adaptive mechanism can effectively eliminate the normal vibration fluctuation caused by the change of the rotating speed and the load, and is the key step to solve the false alarm problem. Secondly, at the feature extraction level, the stroke asymmetry index is introduced, which has a clear physical meaning. It is no longer limited to analyzing the size of the vibration energy, but goes deep into the internal working mechanism of the engine. By comparing the angular velocity smoothness of the working stroke and the compression stroke, the change of the power output form caused by the abnormal combustion such as misfire and knock is captured. More importantly, the index is designed to use the standardized energy deviation for gain adjustment, so that it is not sensitive to the normal working condition change, but highly sensitive to the form deviation of the real fault, realizing the deepening analysis from one-dimensional energy to two-dimensional form. Finally, at the decision confirmation level, the time series cumulative analysis mechanism is adopted to conduct the continuity test on the instantaneous risk index. The mechanism can effectively accumulate the risk signals generated by the real and continuous fault in the time dimension, and make the isolated risk value caused by random electromagnetic interference or instantaneous data jump decay quickly, ensuring the accuracy and reliability of the fault detection result. In summary, through the organic combination of the dynamic adaptive reference, the deep physical features and the time series cumulative analysis, the high-precision, low-false-alarm and high-robustness diagnosis of the engine fault is realized.

[0014] Optionally, the plurality of historical reference working cycles are determined based on the following manner:

[0015] All historical working cycles before the current working cycle are obtained; the energy of the vibration signal of the first historical working cycle is calculated; starting from the second historical working cycle, when the energy of the vibration signal of the historical working cycle is less than the product of the energy of the vibration signal of the last historical working cycle and a preset coefficient, the historical working cycle is determined as a historical reference working cycle.

[0016] The technical scheme constructs an online adaptive screening mechanism of the historical reference working cycle set, adopts a recursive relative comparison criterion for dynamic screening, which can effectively intercept the energy surge signals caused by sudden or severe faults, prevent these data points representing significant abnormal states from entering the reference set, and thus guarantee the sensitivity and stability of the fault detection method in long-term operation.

[0017] Optionally, the working condition similarity of each historical reference working cycle and the current working cycle is determined by the following formula:

[0018]

[0019] In the formula, is the similarity of the working condition of the first historical reference working cycle and the current working cycle, is the average speed and average load of the current working cycle, and is the average speed and average load of the first historical reference working cycle, and is a preset adjustment factor, is a natural exponential function, is an absolute value symbol. This technical solution provides a specific and efficient working condition similarity quantification method, which ensures that the calculation of dynamic reference energy is mainly determined by the historical reference working cycle that best matches the current real-time working condition, thereby realizing the precise and smooth following of the dynamic reference energy to the working condition change, and providing an accurate basis for the establishment of the entire adaptive diagnosis framework. Optionally, the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke are determined based on the following manner: based on the working process of the engine, the crankshaft rotation angle intervals corresponding to the working stroke and the compression stroke are set, and the instantaneous angular velocity subsequence corresponding to the working stroke and the instantaneous angular velocity subsequence corresponding to the compression stroke are extracted from the instantaneous angular velocity sequence of the current working cycle, respectively; wherein the instantaneous angular velocity sequence is calculated based on the crankshaft rotation angle and the crankshaft rotation speed parameter; the variances of the instantaneous angular velocity subsequence of the working stroke and the instantaneous angular velocity subsequence of the compression stroke are taken as the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke.

[0020] Optionally, the stroke asymmetry index is determined based on the following manner:

[0021] The absolute difference and the cumulative value of the angular velocity fluctuation degree of the compression stroke and the angular velocity fluctuation degree of the working stroke are calculated, and the ratio of the absolute difference and the cumulative value is taken as the stroke asymmetry index of the current working cycle.

[0022] This technical solution realizes an adaptive normalization operation by dividing the absolute difference by the sum of the angular velocity fluctuation degrees, effectively eliminating the influence of the angular velocity fluctuation degree caused by the overall rotation speed and load change of the engine, and ensuring that the stroke asymmetry index can stably measure the relative smoothness of the physical signal corresponding to the internal work during engine combustion under any working condition.

[0023] Optionally, one determination manner of the instantaneous risk index of the current working cycle is:

[0024] The absolute difference and the cumulative value of the angular velocity fluctuation degree of the compression stroke and the angular velocity fluctuation degree of the working stroke are calculated, and the ratio of the absolute difference and the cumulative value is taken as the stroke asymmetry index of the current working cycle.

[0025] The absolute difference and the cumulative value of the angular velocity fluctuation degree of the compression stroke and the angular velocity fluctuation degree of the working stroke are calculated, and the ratio of the absolute difference and the cumulative value is taken as the stroke asymmetry index of the current working cycle.

[0026] The absolute difference and the cumulative value of the angular velocity fluctuation degree of the compression stroke and the angular velocity fluctuation degree of the working stroke are calculated, and the ratio of the absolute difference and the cumulative value is taken as the stroke asymmetry index of the current working cycle.Pre-acquire multiple working cycles of the engine under multiple working conditions; acquire a feature combination composed of a standardized energy deviation and a stroke asymmetry index of each working cycle; construct a two-dimensional feature space with the standardized energy deviation as the horizontal coordinate axis and the stroke asymmetry index as the vertical coordinate axis, to obtain a data point set composed of all feature combinations corresponding to the multiple working cycles in the two-dimensional feature space;

[0027] Perform clustering analysis on the data point set to determine a healthy core cluster, and take the centroid of the healthy core cluster as a health center point; determine an instantaneous risk index based on the distance between the data point corresponding to the feature combination of the current working cycle in the two-dimensional feature space and the health center point, wherein the instantaneous risk index and the distance are determined through a positive correlation relationship.

[0028] This technical solution introduces an adaptive risk assessment model based on unsupervised machine learning, which autonomously learns and refines a core cluster and its centroid representing a healthy state from a two-dimensional feature space composed of a standardized energy deviation and a stroke asymmetry index through a clustering analysis algorithm. Therefore, the instantaneous risk index is accurately quantified as the distance between the current state point and the healthy state, improving the accuracy, adaptability, and automation level of diagnosis.

[0029] Optionally, time series accumulation analysis is performed on the instantaneous risk index to determine whether the engine has a fault, including:

[0030] Calculate the attenuation memory fault confirmation value of the current working cycle; ; in the formula, is the attenuation memory fault confirmation value of the current working cycle, is the instantaneous risk index of the current working cycle, is the attenuation memory fault confirmation value of the previous working cycle of the current working cycle, is a preset forgetting factor; when the attenuation memory fault confirmation values of a preset number of consecutive working cycles all exceed a fault confirmation threshold, it is determined that the engine has a fault.

[0031] Optionally, another determination method of the instantaneous risk index of the current working cycle is:

[0032] ; in the formula, is the instantaneous risk index of the current working cycle, is the Euclidean distance between the data point corresponding to the feature combination of the current working cycle in the two-dimensional feature space and the health center point, is a normalized distance factor, is a modulation factor of the current working cycle; wherein,​​ is the centroid of all suspected fault clusters, the nearest suspected fault cluster; is a natural exponential function, is is the Euclidean distance between ; wherein all suspected fault clusters are all clusters other than the healthy core cluster in the result of the cluster analysis.

[0033] The technical solution introduces a modulation factor on the basis of the risk assessment, enhances the directionality and certainty of the risk assessment, enables the diagnostic system to not only identify abnormalities, but also give higher risk weight to abnormal states that present typical fault patterns, and thus significantly improves the recognition sensitivity of real faults.

[0034] Optionally, the current working cycle / historical reference working cycle further comprises a load of the engine, and the vibration signal is acquired by an acceleration sensor; the crank angle is acquired by a crank position sensor; and the load is acquired by reading an electronic control unit coupled with the engine.

[0035] In a second aspect, the present application further provides an engine fault detection system, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to realize the steps of any one of the engine fault detection methods.

[0036] The present application has the following effects:

[0037] The present application can identify the physical nature of the fault by constructing a dynamic reference energy that is self-adaptive to the working condition to realize accurate relative comparison, and determining the stroke asymmetry index by analyzing the physical signal form during combustion work. On this basis, a data-driven decision mechanism is introduced, that is, a healthy prototype is autonomously learned through unsupervised machine learning, and time series cumulative analysis is combined to distinguish between persistent faults and random interference, thereby improving the accuracy and reliability of the final fault detection result. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a flowchart of the present application;

[0039] Figure 2 is a comparison diagram of the speed signal of the current cycle and the speed signal of the historical cycle in the present application;

[0040] Figure 3 is a comparison diagram of the load signal of the current cycle and the load signal of the historical cycle in the present application;

[0041] Figure 4 ​A comparison diagram of the working condition similarity between the historical cycle and the current cycle in the application;

[0042] Figure 5 A comparison diagram of the instantaneous angular velocity signal in the working stroke and the compression stroke of the current cycle and the healthy cycle in the application;

[0043] Figure 6 A two-dimensional feature space constructed in the application and a corresponding clustering analysis result diagram;

[0044] Figure 7 A diagram of the process of dynamically obtaining the decay memory fault confirmation value according to the change trend of the instantaneous risk index in the application. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application.

[0046] With reference to Figure 1 The application provides a complete flow of an engine fault detection method. The method follows a logical order of a benchmark establishment, feature extraction, and decision confirmation. First, the system customizes a dynamic health energy benchmark for the current working condition through adaptive screening of historical operation data; then, deep features capable of representing faults are extracted from two dimensions of energy and working mode; finally, the authenticity and continuity of the faults are finally confirmed through a time sequence analysis model with a time memory function.

[0047] The steps will be described in detail as follows:

[0048] Step S101: Collecting multi-modal data of the engine and determining historical benchmark working cycles.

[0049] This step is the input end and the benchmark setting link of the entire diagnosis framework, and the purpose is to provide an accurate and reliable data basis for subsequent relative comparison. Unlike directly using all historical data, this step introduces an online adaptive screening mechanism, aiming to solve the potential problem that the data set used to calculate the benchmark may be contaminated by persistent fault signals.

[0050] The design logic flow of this step is as follows:

[0051] First, take any working cycle in the engine operation process as the current working cycle, and synchronously collect multi-source sensor data of the engine. Then, the continuous data stream is divided into data frames with working cycles by using the crank angle signal. Then, through a recursive relative comparison criterion, the qualified cycles are screened from the historical data stream to form a historical benchmark working cycle set.

[0052] Detailed implementation description:

[0053] In this embodiment, the vibration signal is acquired by an acceleration sensor mounted on the engine block , the crank angle signal is acquired by a high-resolution optical encoder or other crank position sensor , and the load parameter is acquired by reading the ECU (Electronic Control Unit) data of the engine through the CAN (Controller Area Network) bus.

[0054] Using the TDC (Top Dead Center) pulse in the crank angle signal, the continuous data stream is accurately divided into a complete working cycle of 720° crank angle. Each working cycle contains the vibration signal, crank angle and load within the cycle, and the derivative is obtained to obtain the crank angle speed .

[0055] Then, a plurality of historical reference working cycles are determined based on the following method:

[0056] All historical working cycles before the current working cycle are acquired, for example, 1000 (empirical value) working cycles. The selection of 1000 historical working cycles is an empirical choice to balance data coverage and real-time calculation efficiency. If the number is too small, the working condition coverage is insufficient, and if the number is too large, the system response is slowed down. Specifically, on the one hand, 1000 historical working cycles can cover common engine working conditions such as idle low speed low load or high speed high load, on the other hand, the data amount of 1000 historical working cycles will not occupy too much memory or prolong the calculation time, which meets the real-time requirement of online fault detection. For all historical working cycles before the current working cycle, the system performs a traversal screening. First, the vibration energy of the first historical working cycle is calculated, i.e. the energy of the vibration signal of the first historical working cycle (the sum of the squares of all amplitudes of the vibration signal). Starting from the second historical working cycle, the energy of the vibration signal is calculated and compared with the energy of the vibration signal of the immediately preceding historical working cycle. If the energy of the vibration signal of the current historical working cycle is less than the product of the energy of the vibration signal of the immediately preceding historical working cycle and a predetermined coefficient , it is determined that the historical working cycle belongs to a historical reference working cycle. The setting of the coefficient aims to relax the admission threshold and only exclude data with energy increasing by several times, which can be determined as fault, while allowing fluctuation data within the normal range to enter the sample pool, thereby ensuring the data quality and sample diversity of the healthy sample pool. This simple screening can effectively prevent energy increasing data caused by obvious faults from polluting the sample pool. Among them, The core purpose of this coefficient is to intercept historical operating cycles that cause obvious malfunctions, while retaining those that represent normal fluctuations. Engineering experience shows that vibration energy fluctuations in a healthy engine typically do not exceed three times that of the previous operating cycle, while malfunctions (such as knocking or misfires) can cause a dramatic increase in vibration energy, usually exceeding three times. Therefore, setting this coefficient... It can filter out work cycles that fluctuate under normal operating conditions and effectively capture work cycles corresponding to obvious faults.

[0057] Step S102: Calculate the dynamic reference energy based on the engine's real-time operating conditions.

[0058] This step is a direct means of addressing the core defects of the background technology. It combines the clean historical benchmark working cycle prepared in step S101 with the current real-time operating conditions of the engine to generate a dynamically changing, personalized dynamic benchmark energy. This transforms the diagnostic criteria from static to dynamic and is the core of achieving condition adaptation.

[0059] The design logic flow for this step is as follows:

[0060] First, obtain the average speed and average load of the current working cycle. Next, iterate through each historical benchmark working cycle in the historical benchmark working cycle set and calculate its operating condition similarity to the current working cycle. Then, using the operating condition similarity as a weight, perform a weighted average of the vibration energy of all historical benchmark working cycles to obtain the dynamic benchmark energy of the current working cycle.

[0061] Detailed implementation instructions:

[0062] For the current cycle to be tested, its average rotational speed is , It is a whole, with an average load of Iterate through all historical baseline work cycles and calculate the similarity of the working conditions between each historical baseline work cycle and the current work cycle.

[0063] In this embodiment, the similarity of the current work cycle and the historical baseline work cycle satisfies the following relationship:

[0064]

[0065] In the formula, For the first Similarity of operating conditions between a historical baseline work cycle and the current work cycle. and The average speed and average load of the current working cycle, and The first Average speed and average load of a historical work cycle and is a preset adjustment factor, is a natural exponential function, is an absolute value symbol. Consider two extreme cases: one is that the engine suddenly stalls in the current working cycle due to failure (such as misfire, oil interruption), resulting in is 0, the other is that the load is misjudged as 0 in the current working cycle due to sensor failure (such as sensor signal loss, circuit short circuit) or data transmission error, which will cause a mathematical error of 0 in the denominator in this formula, resulting in a calculation crash. Therefore, in order to avoid the extreme case or is 0, which leads to an unreasonable case of 0 in the denominator, it is uniformly set here. If the extreme case or is 0, in order to avoid calculation errors, set the two parameters to a very small positive number to avoid unreasonable calculation. Specifically, by setting a very small positive number when , set , when is 0, set The core of this replacement strategy is to ensure the robustness of the calculation, avoid mathematical calculation errors, and ensure that the formula can be normally operated. In the field of numerical calculation, a very small positive number is used to handle the denominator of 0, which is a classic robustness design method. The reason for choosing is that it is small enough and is a positive number close to 0, which avoids mathematical calculation errors after replacement, and completely conforms to the original design logic, ensuring the correctness of the calculation result.

[0066] The core of this formula is to quantify the similarity between the two through an exponential decay model that considers the influence of both speed and load. reflects the relative difference rate or deviation degree of the average speed of the historical reference working cycle relative to the corresponding parameter of the current working cycle. reflects the relative difference rate or deviation degree of the average load of the historical reference working cycle relative to the corresponding parameter of the current working cycle, and is normalized by dividing the average speed and average load of the current cycle, respectively. The benefit of normalization is to eliminate the influence of the different dimensions and numerical ranges of speed and load. For example, a 100 RPM difference in speed and a 10 unit difference in load may not be comparable in absolute value, but by calculating the relative difference rate, they can be compared in the same dimension.

[0067] In this formula, and is a preset weight coefficient or adjustment factor, respectively used to adjust the importance of the speed difference and the load difference in the overall similarity calculation, and is determined in the following way:

[0068] The coefficient of variation of the speed and the coefficient of variation of the load of all historical reference working cycles are calculated, the coefficient of variation is equal to the standard deviation divided by the average value, and the coefficient of variation is usually represented by , so the coefficient of variation of the speed is denoted as CV1, CV1 reflects the fluctuation degree of the speed, the greater the fluctuation degree, the greater the impact on the working condition, and vice versa, and similarly, the coefficient of variation of the load is denoted as CV2, CV2 reflects the fluctuation degree of the load, the greater the fluctuation degree, the greater the impact on the working condition, and vice versa, so is set, which reflects the proportion of the fluctuation degree of the speed in the overall fluctuation degree, the greater the proportion, the greater the impact of the fluctuation degree of the speed on the working condition in the overall fluctuation degree of the speed and the load, so a greater weight is set, and vice versa, and is set, which reflects the proportion of the fluctuation degree of the load in the overall fluctuation degree, the greater the proportion, the greater the impact of the fluctuation degree of the load on the working condition in the overall fluctuation degree of the speed and the load, so a greater weight is set, and vice versa.

[0069] In this formula, is a comprehensive working condition difference index obtained by weighted sum of the relative difference rate of the speed and the relative difference rate of the load. This comprehensive index comprehensively reflects the overall difference degree of the two working cycles in the two key dimensions of speed and load. The greater the difference, the greater the value of this item. The constructs an exponential decay function, which constructs a negative correlation between the overall difference degree and the working condition similarity. The similarity weight is very sensitive to the change of the working condition difference. When the historical working cycle and the current working cycle are completely consistent in working condition, i.e. the difference of speed and load is 0, the value in the bracket is 0, equal to 1, and the working condition similarity reaches the maximum value 1. With the increase of the working condition difference, i.e. the deviation of speed or load, the absolute value of the exponential part of the exponential decay function increases, and the working condition similarity will quickly decrease and tend to 0. This rapid decay feature accurately reflects the physical meaning of being closer and more relevant, and can effectively amplify the weight of samples with similar working conditions, while quickly suppressing the influence of samples with large working condition difference.

[0070] Overall, the formula's overall effect is to build a robust and sensitive operating condition similarity assessment model. It not only simply measures the absolute gap of speed and load, but also scientifically integrates the influence of these two key parameters through normalization processing and adjustable weight factors.

[0071] In summary, by utilizing the characteristics of the exponential decay function, this multi-dimensional operating condition difference is mapped to the operating condition similarity between This operating condition similarity intuitively and effectively reflects that the closer the average speed and average load of the historical reference working cycle to the current state, the more it represents the health level under the current state, and therefore is given a higher weight. This weight distribution is crucial for subsequent health state assessment, fault diagnosis or performance prediction tasks, because it ensures that the model will focus on the most relevant historical data, thereby improving the accuracy and reliability of the analysis.

[0072] Please refer to Figure 2 , Figure 3 and Figure 4 , which collectively and step-by-step demonstrate the basis and results of operating condition similarity calculation, providing intuitive evidence for the effectiveness of this embodiment. Figure 2 The comparison results of the speed signals of the current cycle and the historical cycles are shown. The red line represents the speed signal of the current cycle, and the other color lines represent five different historical cycles. From a visual perspective, it can be clearly judged that the overall trend of the speed signals of historical cycle 1 and historical cycle 2 is closest to the current cycle, indicating that their average speed difference is the smallest. Figure 3 The comparison of the load signals of the historical cycles and the load signal of the current cycle is shown. The load signals of historical cycle 1 and historical cycle 2 are closest to the current cycle. Figure 4 The final results obtained by substituting these signal differences into the operating condition similarity formula are shown. It can be seen that although the operating conditions of historical cycles 1 and 2 are visually very close to the current cycle, after precise calculation, the operating condition similarity score of historical cycle 1 is the highest, and the operating condition similarity values of other historical cycles with larger operating condition differences are significantly reduced.

[0073] Subsequently, the dynamic reference energy of the current cycle is calculated by weighting:

[0074]

[0075] In this formula, is the dynamic reference energy of the current cycle, is the serial number of the historical reference working cycle, is the total number of historical reference working cycles, is the The working condition similarity between the historical benchmark working cycle and the current working cycle, whose value is between 0 and 1, the greater the value, the more similar the working condition (speed and load) of the two working cycles. This quantifies the importance or relevance, which makes the historical benchmark working cycle that best matches the working condition of the current working cycle dominant in the calculation, while the historical benchmark working cycle with large working condition difference is marginalized. The energy of the vibration signal of the first historical benchmark working cycle (the sum of the squares of all amplitudes of the vibration signal) represents the energy level that the vibration signal should have when the engine is running under the specific working condition of the historical benchmark working cycle.

[0076] The numerator part of the formula will weight and sum the energy of the vibration signal of each historical benchmark working cycle according to the working condition similarity between the historical benchmark working cycle and the current working cycle, and collect the energy information of all historical benchmark working cycles. The contribution of the energy of the vibration signal of the historical benchmark working cycle with high working condition similarity to the total will be greater, and the value close to its original value will be counted into the total; while the contribution of the energy of the vibration signal of the historical benchmark cycle with low working condition similarity to the total will be smaller.

[0077] The denominator part of the formula is to add all the working condition similarities for normalization. By dividing by the sum of the working condition similarities (the sum of the weights), the weighted energy sum can be converted to an average energy value with actual physical meaning.

[0078] In summary, the essence of the formula is to use weighted average based on working condition similarity to construct a dynamic benchmark energy. This dynamic benchmark energy can accurately follow the normal energy fluctuations of the vibration signal caused by changes in speed and load, effectively distinguishing between normal working condition changes and real equipment abnormalities, and reducing false positives caused by traditional fixed threshold methods.

[0079] In this way, when a real fault occurs, the energy of the abnormal vibration signal will be greatly different from the dynamic benchmark energy, making the fault features clearly prominent, improving the sensitivity and reliability of fault detection.

[0080] Step S103: Extract the standardized energy deviation and stroke asymmetry index.

[0081] This step further quantifies the standardized energy deviation of the current working cycle, reflecting whether the energy of the vibration signal of the current working cycle is abnormal. At the same time, combined with the internal working mechanism of the engine, the fault features are extracted from these two dimensions, providing more abundant and more targeted information for subsequent accurate decision-making, which is the key to realizing fault detection to fault classification.

[0082] The design logic flow of this step is: ​

[0083] First, the comparison result of the energy of the vibration signal of the current working cycle and the dynamic reference energy determines the normalized energy deviation of the current working cycle; then, the crank angle of the current working cycle decomposes the current working cycle into a power stroke and a compression stroke; next, the angular velocity fluctuation degree of the power stroke and the angular velocity fluctuation degree of the compression stroke are determined; finally, the stroke asymmetry index is determined by fusing the angular velocity fluctuation degrees of the power stroke and the compression stroke.

[0084] Detailed implementation description:

[0085] In the embodiment, the normalized energy deviation of the current working cycle satisfies the following relationship:

[0086] Calculate the actual vibration energy of the current cycle , and obtain the normalized energy deviation:

[0087]

[0088] In the formula, is the normalized energy deviation of the current cycle, is the energy of the vibration signal of the current cycle, is the dynamic reference energy of the current cycle.

[0089] This formula defines a dimensionless normalized energy deviation, and its core function is to normalize the actual vibration energy of the current working cycle against its health reference under the current working condition. It directly reflects how many times the actual energy is the dynamic reference energy, and reflects the degree to which the energy of the vibration signal of the current working cycle deviates from its dynamic reference energy.

[0090] In the formula, when : this means , that is, the actual energy is almost completely consistent with the dynamic reference energy, which represents the most ideal and healthy running state of the engine under the working condition. When : this means , indicating that the actual energy exceeds the health reference under the current working condition. The size of the value quantifies the degree of excess, for example, , clearly indicating that the actual energy is 1.3 times the health reference, that is, 30% higher, which strongly suggests that the engine may have faults such as knock, pre-ignition, etc. that cause abnormal intense energy release. When : this means , indicating that the actual energy fails to reach the health level it should have under the current working condition. For example, This indicates that the actual energy is only 60% of the health benchmark, i.e. 40% lower, which strongly suggests that the engine may have a fault such as misfire, partial fire or insufficient combustion, etc. leading to insufficient energy output.

[0091] At the same time, according to the crank angle of the current working cycle, the current working cycle is decomposed into a power stroke and a compression stroke. For a standard four-stroke engine, the piston needs to complete four strokes of intake, compression, power and exhaust to constitute a complete working cycle, which corresponds to two revolutions of the crankshaft, i.e. 720 degrees of crank angle. In these 720 degrees, different strokes have different functions, among which the compression stroke is the main energy consumption and force accumulation stage, and the power stroke is the only energy generation and force stage. The conventional method is difficult to distinguish between normal working condition changes and real combustion faults, and the fundamental reason is that it fails to effectively utilize the inherent and stable balance relationship between the two strokes. This step is to construct a feature that is not sensitive to working condition changes but highly sensitive to combustion abnormalities by accurately extracting and comparing the two stroke signals that constitute a unity of opposites in a physical sense.

[0092] Decomposing the current working cycle into a power stroke and a compression stroke is based on the basic working principle of the Otto cycle or Diesel cycle of a four-stroke engine. This principle clearly states that in a 720-degree cycle with the intake stroke top dead center as the 0-degree reference, the crank angle in the 180-degree to 360-degree interval must correspond to the compression stroke, and in the 360-degree to 540-degree interval must correspond to the power stroke. This angle correspondence is determined by the mechanical structure of the engine and is constant and universal.

[0093] Therefore, for the current working cycle, the TDC (top dead center) signal is used as the reference of the angle coordinate system, the crank angle in the 180-degree to 360-degree interval is defined as the compression stroke, and the crank angle in the 360-degree to 540-degree interval is defined as the power stroke, achieving the decomposition of the current working cycle.

[0094] According to the crank angle of the engine at each time in the current working cycle, the crank angle sequence of the current working cycle is obtained; according to the crank angle sequence of the current working cycle and the crank speed parameter, the instantaneous angular velocity sequence of the current working cycle is obtained; the instantaneous angular velocity is the rate of change of the angle of the crank angle with time, that is, the instantaneous angular velocity is the ratio of the angle of the crankshaft turned to a small angle and the time spent. The final result of this process is to obtain the crank angle sequence and the instantaneous angular velocity sequence of the current working cycle, and the two sequences have the same length and are one-to-one corresponding based on time.

[0095] Traverse the instantaneous angular velocity sequence of the current working cycle, extract all the crank angles falling into the above compression stroke interval, and according to the one-to-one correspondence between the crank angle sequence and the instantaneous angular velocity sequence, obtain the instantaneous angular velocity values corresponding to these crank angles, and form the instantaneous angular velocity sub-sequence corresponding to the compression stroke;

[0096] Similarly, extract all the crank angles falling into the working stroke interval, and according to the one-to-one correspondence between the crank angle sequence and the instantaneous angular velocity sequence, obtain the instantaneous angular velocity values corresponding to these crank angles, and form the instantaneous angular velocity sub-sequence corresponding to the working stroke.

[0097] Then, the variance of the instantaneous angular velocity sub-sequence corresponding to the working stroke and the variance of the instantaneous angular velocity sub-sequence corresponding to the compression stroke are calculated, which represent the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke, respectively.

[0098] In this way, targeted analysis can be performed within a fixed window with clear physical meaning. Instead of observing the overall performance of the entire cycle, the smoothness of the engine in the two key actions of charging and working is directly examined. This self-referencing comparison method can naturally offset the influence of overall angular velocity rise or fall caused by changes in speed and load, but can extremely sensitively capture signs of the balance between charging and working being broken due to misfire, knock and other faults.

[0099] Finally, based on the angular velocity fluctuation degree of the working stroke and the angular velocity fluctuation degree of the compression stroke, a stroke asymmetry index is determined to represent the deviation of the combustion process from the standard symmetrical state.

[0100] In the present embodiment, the stroke asymmetry index is calculated by the following formula:

[0101]

[0102] In the formula, is the stroke asymmetry index of the current working cycle, is the angular velocity fluctuation degree of the working stroke of the current working cycle, which quantifies the running smoothness of the engine in the working stage. The larger the variance, the more intense the angular velocity fluctuation, the more unsmooth the running, and vice versa. is the angular velocity fluctuation degree of the compression stroke of the current working cycle, which quantifies the running smoothness of the engine in the charging stage. The larger the variance, the more intense the angular velocity fluctuation, the more unsmooth the running, and vice versa, is the absolute value symbol.

[0103] A healthy engine, its energy consumption and production process should follow a kind of harmony, symmetry rhythm. Each working cycle of the engine, its core is the energy exchange of two key strokes, compression stroke is the input or force stage of the engine, consume energy to compress the mixture in the cylinder, provide potential energy for work, the running smoothness of this process (quantified by ) reflects the stability of the input stage. The working stroke is the output or force stage of the engine, the mixture burns, the chemical energy is converted into mechanical work, and the piston does work. The running smoothness of this process (quantified by ) reflects the stability of the output stage. If the engine is in an ideal healthy state, the running smoothness of the two strokes should be in a dynamic symmetrical balance, that is and The size is close, any abnormal combustion (such as knock or misfire) will fundamentally break this symmetry.

[0104] The formula will destroy this physical symmetry into a measurable stroke asymmetry index, the numerator reflects the size of the asymmetry, which is the most direct measure of the imbalance. It calculates the absolute difference between the running smoothness of the output stage and the running smoothness of the input stage. In normal cases, the better the symmetry of the two strokes, the smaller the absolute difference, the closer to 0, the smaller the stroke asymmetry index, when the engine fails, whether it is caused by dramatic increase, or caused by sudden decrease, will increase the absolute difference, the worse the symmetry of the two strokes, resulting in a larger stroke asymmetry index. The denominator is a normalized scale that adapts to the working condition. The overall angular velocity fluctuation of the engine will naturally be larger at high speed and high load than at idle speed. If only the absolute difference of the numerator is considered, it may be misjudged at different working conditions. By dividing by the sum of the two angular velocity fluctuation levels, an adaptive normalization is achieved. It considers the amplitude of asymmetry in the context of the current total fluctuation level, thus completely decoupling the overall influence of the working condition (speed, load).

[0105] Referring to Figure 5 , by comparing the instantaneous angular velocity of a healthy cycle and the current cycle, the physical connotation of the stroke asymmetry index is intuitively revealed. Figure 5 In Figure 5The right half shows the waveform pattern of the instantaneous angular velocity of a healthy cycle. It can be seen that the waveform is regular and stable in amplitude, reflecting the smooth and balanced energy consumption (compression stroke) and generation (power stroke) processes of the engine in a healthy state. At this time, the angular velocity fluctuation degree of the power stroke is close to that of the compression stroke, and the calculated stroke asymmetry index tends to be 0. While Figure 5 The left half shows the waveform pattern of the instantaneous angular velocity of the current cycle. During the power stroke phase near 400° crank angle, the peak value of the angular velocity shows a significant and abnormal bulge, indicating that the engine has experienced a violent and unstable energy release during the power stroke of the cycle. Such a dramatic change in the waveform will result in a large difference between the angular velocity fluctuation degree of the power stroke and that of the compression stroke, thereby significantly increasing the stroke asymmetry index. The stroke asymmetry index can effectively capture the imbalance of the engine's internal power form, which is an effective feature that distinguishes from traditional energy analysis and deeply analyzes the physical nature of the fault, providing a key morphological basis for diagnosis.

[0106] In summary, the stroke asymmetry index of the current working cycle is a highly refined physical index for measuring the internal symmetry of the engine's core working cycle. It strictly tends to zero in a healthy state; any imbalance in the smoothness of the input and output stages caused by abnormal combustion will significantly increase its value. Therefore, the size of the asymmetry index can be directly, objectively and stably used as a quantitative basis for the health of the engine's combustion process.

[0107] Step S104: Construct a two-dimensional feature space and calculate the instantaneous risk index.

[0108] This part is the pre-stage of the decision-making layer of the diagnosis system, which uses the indicators extracted in step S103 to construct a two-dimensional feature space and evaluate the instantaneous risk index based on the distribution of the current working cycle in the two-dimensional feature space.

[0109] Design logic flow:

[0110] First, a two-dimensional feature space is constructed with the standardized energy deviation as the horizontal coordinate axis and the stroke asymmetry index as the vertical coordinate axis. Then, a healthy core cluster is determined through clustering analysis. Finally, based on the distance between the data point corresponding to the feature combination of the current working cycle in the two-dimensional feature space and the healthy center point, the instantaneous risk index is determined.

[0111] Specifically:

[0112] The multiple working cycles of the engine under multiple working conditions, including healthy working conditions and various classic types of faults, are pre-acquired, and the standardized energy deviation and the stroke asymmetry index of each working cycle are obtained through the foregoing steps S101-S103, to obtain a feature combination composed of the standardized energy deviation and the stroke asymmetry index of each working cycle;

[0113] A two-dimensional feature space is constructed with the standardized energy deviation as the horizontal coordinate axis and the stroke asymmetry index as the vertical coordinate axis, and each data point of the two-dimensional feature space corresponds to a feature combination , and all feature combinations corresponding to the multiple working cycles form a data point set in the two-dimensional feature space.

[0114] The data points of different health states will exhibit obvious clustering characteristics, and the data point set is subjected to clustering analysis, for example, the DBSCAN (density-based spatial clustering with noise) algorithm can be used, which does not need to specify the number of clusters in advance, and can identify clusters of any shape, and can also identify outliers that do not belong to any cluster as noise, which is perfectly matched with the scene. The data points corresponding to the health state form a high-density core cluster, various fault data points can form other clusters with slightly lower density, and some random interference will be naturally identified as noise.

[0115] After the DBSCAN algorithm is executed, a plurality of clusters are output, and according to prior knowledge, the cluster containing the most data points and the centroid closest to the health center point is recorded as a health core cluster, and the centroid of the cluster is taken as a health core point CH, so that the centroid of the cluster is solidified as a core parameter, and all clusters other than the cluster closest to the health center point are taken as suspected fault clusters. Among them, the horizontal coordinate of the health center point represents that the actual vibration energy of the engine completely coincides with the dynamic health baseline under the current working condition, representing the ideal health state in the energy dimension. The vertical coordinate represents that the angular velocity fluctuation degrees of the power stroke and the compression stroke of the engine are completely equal, representing the perfect symmetry and balance of the internal power form of the engine.

[0116] It can be seen that this coordinate point is determined by the physical meanings of the two core feature indexes constructed by the application, and is the benchmark point that best represents the ideal health state of the engine. The health data points in actual operation will closely surround the theoretical center point and form a Gaussian distribution, and any fault will cause the data points to systematically deviate from the center in the feature space.

[0117] Then, based on the distance between the data point corresponding to the feature combination of the current work cycle and the health center point in the two-dimensional feature space, an instantaneous risk index is determined, wherein the instantaneous risk index and the distance are determined by a positive correlation formula.

[0118] Reference Figure 6 As shown in the figure, this figure is a visualization of the two-dimensional feature space constructed by the present invention and the analysis results. Figure 6 The x-axis represents the standardized energy deviation (NED), and the y-axis represents the stroke asymmetry index (SAI). This shows that the data points for different health states exhibit obvious clustering characteristics, with the core health clusters tightly clustered around the theoretical health points. The surrounding area validated the physical expectation of small energy deviation and good symmetry under healthy conditions. Various suspected fault clusters and noise clusters were distributed far from the healthy core and were clearly separated from each other. The data points corresponding to the current cycle in two-dimensional space were significantly deviated from the healthy center point. The red dashed line connecting the two intuitively represents the Euclidean distance between them, which is the basis for calculating the instantaneous risk index R. This two-dimensional feature space effectively separates different engine health states, and the risk assessment method based on the distance from the data point to the healthy center is intuitive, quantitative, and reliable, providing a solid basis for subsequent fault decisions.

[0119] In this embodiment, the calculation process for the instantaneous risk index for the current working cycle during actual engine operation is as follows:

[0120] Find the data point in the two-dimensional feature space corresponding to the characteristic combination of the standardized energy deviation and the stroke asymmetry exponent of the current work cycle, denoted as . Instantaneous risk index of the current work cycle Defined as Go to the health center The normalized distance, i.e.:

[0121]

[0122] In this formula, This represents the instantaneous risk index for the current work cycle. This represents the data points in the two-dimensional feature space corresponding to the characteristic combination of the standardized energy deviation and stroke asymmetry index of the current working cycle. As a health center, for and The Euclidean distance between them The larger the value, the further the engine state is from a healthy state in the current work cycle, and the greater the instantaneous risk index of the current work cycle, and vice versa. is a normalization distance factor, which is a constant used to scale the range of the instantaneous risk index, and it can be set as the average radius of the healthy core cluster.

[0123] In summary, the formula converts the fault risk into a standardized, interpretable, and adaptive numerical indicator in a data-driven manner, which is a key bridge connecting advanced feature extraction and intelligent fault decision-making, and improves the accuracy and adaptability of detection.

[0124] To enhance the sensitivity to specific faults, a modulation factor related to the nearest fault cluster can be introduced, and another way to determine the instantaneous risk index of the current working cycle is:

[0125]

[0126] In the formula, is the instantaneous risk index of the current working cycle, is the Euclidean distance between and . is a normalization distance factor, which is a constant used to scale the range of the instantaneous risk index, and it can be set as the average radius of the healthy core cluster. is the fault modulation factor of the current working cycle, which is calculated as: , is the nearest suspected fault cluster among all suspected fault clusters, is the natural exponential function, is the Euclidean distance between and . The meaning of is that when not only is far from the healthy center, but also is very close to the center of a certain fault mode, it is more likely to be a real fault, and therefore is given a higher risk weight. The form of ensures that when is closer to the center of a certain fault mode, the value of

[0127] is closer to 1; otherwise, it quickly decays to 0.

[0128] This clustering analysis-based method makes the calculation of the instantaneous risk index completely data-driven. It not only measures the degree to which the working state corresponding to the current working cycle deviates from health, but also measures the degree to which the working state corresponding to the current working cycle is close to a typical fault through the fault modulation factor, which makes the evaluation of the instantaneous risk index more intelligent, comprehensive, and accurate.

[0129] This step is the final decision-making layer of the entire diagnostic framework. Its core purpose is to introduce a time-memory mechanism based on the instantaneous risk index to effectively distinguish between real, time-continuous faults and occasional, random interferences (such as electromagnetic noise), thereby ensuring the robustness of the final diagnostic results.

[0130] The design process for this step involves maintaining and continuously updating a variable called the decayed memory fault confirmation value. This variable functions similarly to a low-pass filter or leaky bucket integrator for risk signals. The decayed memory fault confirmation value is updated using an instantaneous risk index. Finally, a fault determination is made based on whether the decayed memory fault confirmation value consistently exceeds a threshold.

[0131] Specifically, it includes:

[0132] Initialization: For the first The work cycle, obtain the first The four consecutive (experience-based) work cycles preceding the first work cycle will... Decayed memory fault confirmation value per working cycle Set to 0 to achieve initialization. Four working cycles are the minimum effective window for initializing the decay memory fault confirmation value. Since the decay memory fault confirmation value needs to be iteratively updated, four working cycles are sufficient to allow the decay memory fault confirmation value to iterate to a stable level through the instantaneous risk index under normal operating conditions, avoiding the initial value deviation from interfering with subsequent fault judgment, and at the same time, it will not delay the system's start-up detection time due to an excessively large window.

[0133] Iterative update: For the first The work cycle, utilizing the first Decayed memory fault confirmation value per working cycle and the Instantaneous risk index per work cycle Determine the first Decayed memory fault confirmation value per working cycle ;

[0134] Specifically, no. The decay memory fault confirmation value for each working cycle is:

[0135]

[0136] In this formula, For the first The decayed memory fault confirmation value for each working cycle. For the first The instantaneous risk index of each work cycle, as a current risk input, represents the new risk information brought about by the current cycle. For the first The decaying memory fault confirmation value for each work cycle represents the memory characteristics of historical risks. The preset forgetting factor, a constant close to 1, controls the rate at which historical memory decays. The closer it is to 1, the stronger the memory; this is the setting. A value of 0.9 is an empirical value, representing a strong historical memory capability. The core purpose is to accurately match the essential characteristics of engine fault signals with the core objectives of fault detection. This is because engine faults, such as minor valve leaks, progressive cylinder misfires, and minor fuel line blockages, are typically continuous, gradually changing signals. The risk value of these faults needs to be gradually accumulated through historical data from multiple operating cycles to stand out from the background noise of normal operation and reach the alarm threshold. If the historical memory capability is too low, i.e., If the setting is too low, the weight of historical data will decay rapidly, causing weak fault signals to be excessively forgotten and difficult to accumulate effectively, ultimately leading to missed fault detection and defeating the fundamental purpose of detection. In short, setting... With a value of 0.9, it retains strong memory capabilities. Essentially, it selectively preserves the cumulative effect of continuous fault signals, avoiding the loss of fault information due to weak memory and improving the accuracy of fault detection.

[0137] In this formula, Part of this is a memory decay term, representing the inertia or memory of historical risks. When an occasional disturbance pulse causes the instantaneous risk index of the current work cycle to be very high, the decayed memory fault confirmation value of the current work cycle jumps accordingly. For example, It was very low, then suddenly a very high one appeared. The formula becomes The decaying memory fault confirmation value jumped instantly from 0.1 to 5.09. However, in the following work cycle, as the instantaneous risk index of the current work cycle returns to normal (close to 0), the decaying memory fault confirmation value of the current work cycle will decrease under the influence of the forgetting factor. In the next cycle... The interference disappeared. , Next cycle , The decay memory fault confirmation value is from... decay to Then decay to .

[0138] When a real fault occurs, the instantaneous risk index of the current work cycle remains positive for multiple consecutive work cycles. For example, real faults (such as fires or explosions) are physical and will continue to occur, so the instantaneous risk index will consistently output positive values. This allows the decaying memory fault confirmation value of the current work cycle to accumulate by adding the instantaneous risk index of the current work cycle, and its value will steadily increase until it breaks through and stabilizes at a high level. For example, When the value remains at 2, , , This accumulation and steady growth is driven by current risk inputs.

[0139] Reference Figure 7 As shown, the dynamic change process of the decay memory fault acknowledgment value (DMFV) is demonstrated, verifying its ability to suppress interference and accumulate fault signals. Figure 7 The upper part shows the instantaneous risk index as input. At the 5th working cycle, an isolated, high-amplitude interference pulse appears. (Corresponding to...) Figure 7 The lower half of the DMFV value (output) shows that the DMFV jumps to 5.55, but since the subsequent R regresses normally (close to 0), the forgetting factor... Under the influence of this mechanism, the DMFV value rapidly decayed to 5.10 in the next cycle, failing to continue increasing. This intuitively demonstrates the mechanism's ability to suppress intermittent interference. However, starting from the 11th working cycle, when R remained positive (representing a real fault), the DMFV value in the figure below began to accumulate steadily, increasing from 5.42 to 6.88, and then to 8.19, clearly demonstrating the cumulative effect on persistent fault signals. This verifies that the DMFV mechanism can effectively distinguish between isolated interference and persistent faults. By successfully filtering out transient noise through memory decay and accurately identifying the real fault through risk accumulation, the robustness and reliability of the final diagnostic results are improved.

[0140] Let the current work cycle number in step S101 be... The sequence number of the previous work cycle is . The decay memory fault confirmation value for the current work cycle is:

[0141]

[0142] In the formula, This is the decay memory fault confirmation value for the current work cycle. This represents the instantaneous risk index for the current work cycle. This is the decayed memory fault confirmation value from the previous work cycle of the current work cycle. The specific logic and the first... (The text abruptly ends here, likely due to an incomplete sentence or a formatting error.) The decay memory fault confirmation value is the same for each working cycle.

[0143] Final fault determination: An engine fault is officially determined only when the decay memory fault confirmation value exceeds a preset fault confirmation threshold for a consecutive preset number of working cycles. This criterion of continuously exceeding the threshold ensures that only stable and reproducible anomalies confirmed by time series are ultimately identified as real faults.

[0144] Specifically, a fault confirmation threshold of 0.8 (an empirical value) is set. Starting from the current work cycle, the decayed fault confirmation values ​​for the previous four work cycles (a total of five consecutive work cycles) are obtained. Only if the decayed fault confirmation values ​​for all five consecutive work cycles exceed 0.8 is an engine fault determined and a fault alarm issued. If at least one decayed fault confirmation value among the five consecutive work cycles does not exceed 0.8, it is determined that the engine has not experienced a fault in the current work cycle.

[0145] The threshold of 0.8 is determined through sample calibration using empirical values. Statistical analysis of numerous attenuated memory fault confirmation values ​​across multiple operating cycles in both healthy and faulty states revealed that the attenuated memory fault confirmation values ​​in multiple operating cycles under faulty conditions exceed 0.8. Therefore, 0.8 was set as the threshold. Furthermore, considering the persistent nature of engine faults (such as continuous misfires and chronic knocking), which typically involve many continuous operating cycles, while transient disturbances (such as electromagnetic noise) usually involve fewer, five consecutive operating cycles were set as the judgment condition. This eliminates transient disturbances and confirms the authenticity of the fault. When the attenuated memory fault confirmation values ​​for all five consecutive operating cycles exceed 0.8, it indicates that the anomaly is a persistent physical fault, not a random disturbance, conforming to the engineering diagnostic principle that faults must be reproduced for confirmation. This mechanism ensures that only stably reproducible anomalies are confirmed as faults and trigger fault alarms.

[0146] An engine fault detection system of the present invention includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the operations of steps S101-S105 in order to accurately detect engine faults.

[0147] In summary, by organically combining the above steps, this invention constructs a complete, scientific, and reliable engine fault diagnosis solution.

[0148] 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 within the protection scope of the present invention.

Claims

1. A method for detecting engine faults, characterized in that, include: The current working cycle is determined by taking any working cycle during engine operation and identifying multiple historical baseline working cycles. Both the current working cycle and the historical baseline working cycle include engine vibration signals and crankshaft angle; The similarity of operating conditions between each historical baseline work cycle and the current work cycle is determined by the following formula: ; In the formula, For the first Similarity of operating conditions between a historical baseline work cycle and the current work cycle. and The average speed and average load of the current working cycle, and The first Average speed and average load of a historical baseline working cycle and The preset adjustment factor, It is a natural exponential function. It is the absolute value symbol; Based on the similarity of operating conditions between each historical benchmark working cycle and the current working cycle, the vibration energy of the vibration signal of each historical benchmark working cycle is weighted and calculated to determine the dynamic benchmark energy of the vibration signal of the current working cycle. The standardized energy deviation of the current working cycle is determined by comparing the energy of the vibration signal in response to the current working cycle with the dynamic reference energy. Based on the crankshaft angle of the current working cycle, the current working cycle is decomposed into a power stroke and a compression stroke, and the angular velocity fluctuations of the power stroke and the compression stroke are determined respectively: Based on the engine's power stroke flow, crankshaft angle intervals corresponding to the power stroke and compression stroke are set, and instantaneous angular velocity subsequences corresponding to the power stroke and compression stroke are extracted from the instantaneous angular velocity sequence of the current working cycle, respectively; wherein, the instantaneous angular velocity sequence is calculated based on crankshaft angle and crankshaft speed parameters; the instantaneous angular velocity of the power stroke is... The variances of the angular velocity subsequence and the instantaneous angular velocity subsequence of the compression stroke are used as the degree of angular velocity fluctuation in the power stroke and the degree of angular velocity fluctuation in the compression stroke. By fusing the degree of angular velocity fluctuation in the power stroke and the degree of angular velocity fluctuation in the compression stroke, a stroke asymmetry index is determined to characterize the dynamic balance between the power and energy storage processes inside the engine. This includes calculating the absolute difference and the cumulative value of the degree of angular velocity fluctuation in the compression stroke and the degree of angular velocity fluctuation in the power stroke, and using the ratio of the absolute difference to the cumulative value as the stroke asymmetry index of the current working cycle. The method for determining the instantaneous risk index of the current working cycle by combining standardized energy deviation and stroke asymmetry index includes: pre-acquiring multiple working cycles of the engine under various operating conditions; acquiring the feature combination composed of standardized energy deviation and stroke asymmetry index for each working cycle; constructing a two-dimensional feature space with standardized energy deviation as the horizontal axis and stroke asymmetry index as the vertical axis, obtaining the set of data points formed by all feature combinations corresponding to multiple working cycles in the two-dimensional feature space; performing cluster analysis on the set of data points to determine a healthy core cluster, and taking the centroid of the healthy core cluster as the healthy center point; determining an instantaneous risk index based on the distance between the data point corresponding to the feature combination of the current working cycle in the two-dimensional feature space and the healthy center point, wherein the instantaneous risk index and the distance are determined by a positive correlation formula; and performing time series cumulative analysis on the instantaneous risk index to determine whether the engine has a fault, including: calculating the decay memory fault confirmation value of the current working cycle. In the formula, This is the decay memory fault confirmation value for the current work cycle. This represents the instantaneous risk index for the current work cycle. This is the decayed memory fault confirmation value from the previous work cycle of the current work cycle. The preset forgetting factor is used; when the decay memory fault confirmation value of a preset number of consecutive working cycles exceeds a fault confirmation threshold, the engine is determined to have a fault.

2. The engine fault detection method according to claim 1, characterized in that, Multiple historical baseline work cycles were determined based on the following method: Retrieve all historical work cycles preceding the current work cycle; Calculate the energy of the vibration signal from the first historical working cycle; Starting from the second historical working cycle, when the energy of the vibration signal in the historical working cycle is less than the product of the energy of the vibration signal in the previous historical working cycle and a preset coefficient, the historical working cycle is determined as a historical reference working cycle.

3. The engine fault detection method according to claim 1, characterized in that, Another way to determine the instantaneous risk index of the current work cycle is: ;In the formula, This represents the instantaneous risk index for the current work cycle. The data points corresponding to the feature combination of the current work cycle in this two-dimensional feature space. With Health Center The Euclidean distance between them For normalized distance factor, The modulation factor for the current work cycle; where, , For all suspected fault clusters, the centroid and Recent suspected faulty clusters; It is a natural exponential function. for and The Euclidean distance between them; where all suspected faulty clusters are all other clusters besides the healthy core cluster in the results of cluster analysis.

4. The engine fault detection method according to claim 1, characterized in that, The current duty cycle / historical baseline duty cycle also includes engine load, and vibration signals are obtained through an accelerometer; crankshaft angle is obtained through a crankshaft position sensor; The load is obtained by reading the electronic control unit coupled to the engine.

5. An engine fault detection system, characterized in that, The engine fault detection system includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the engine fault detection method as described in any one of claims 1-4.

Images