An emergency diesel generator fault monitoring method, device, equipment and product

Abstract

The application relates to the technical field of emergency diesel generator monitoring, and discloses an emergency diesel generator fault monitoring method, device, equipment and product, which comprises the following steps: obtaining operation data, health state parameters and operation data of an emergency diesel generator at different time points in a historical period; adopting a reinforced mean clustering algorithm, combining with working condition auxiliary data, performing clustering on feature data to obtain a plurality of clustering clusters, and determining the working conditions corresponding to the clustering clusters; respectively determining initial alarm threshold values of various features according to probability density curves; correcting the alarm threshold values according to the health state parameters; and determining a fault state according to current operation data and the alarm threshold values; in the application, the data obtained through detection points are used for monitoring mechanical faults, the alarm threshold values are determined in combination with operation data obtained through the detection points, the fault state at the current time is judged based on the alarm threshold values, the problem of insufficient mechanical fault monitoring effect in a traditional monitoring method is solved, and the monitoring result is more accurate.

Description

Technical Field

[0001] This invention relates to the field of generator monitoring technology, specifically to a method, device, equipment, and product for monitoring faults in emergency diesel generators. Background Technology

[0002] For fault monitoring of emergency diesel generator sets, a local monitoring system is configured on-site. This system mainly monitors conventional thermal parameters of the emergency diesel generator sets, such as generator speed, cylinder exhaust temperature, cooling water pressure, temperature, and fuel pressure. These thermal parameters have a certain monitoring effect on leakage, generator overheating, and performance degradation, but they are insufficient for monitoring mechanical faults such as generator wear and fracture. Summary of the Invention

[0003] This invention provides a method, device, equipment, medium, and product for monitoring faults in emergency diesel generator sets, in order to solve the problem that existing emergency diesel generator set fault monitoring methods are insufficient in monitoring mechanical faults.

[0004] In a first aspect, the present invention provides a method for monitoring faults in an emergency diesel generator, comprising: Acquire the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the characteristic data of various features of different detection points in the emergency diesel generator. The enhanced mean clustering algorithm is adopted, and the auxiliary data of the emergency diesel generator at different times in the historical period are combined to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and the operating conditions corresponding to each cluster are determined. For each detection point, the initial alarm threshold for each feature under different operating conditions is determined based on the probability density curves of each feature in different clusters. The initial alarm threshold is corrected based on the current health status parameters of the emergency diesel generator, and the current alarm thresholds of each detection point under different operating conditions are obtained. The current fault status of each detection point is determined based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

[0005] By adding additional detection points to key components of emergency diesel generators, the system achieves the goal of monitoring mechanical faults that cannot be considered in traditional fault prediction methods. The system inputs the operating data collected from each detection point into a clustering algorithm to automatically classify different operating conditions such as startup, steady state, and transition. Based on historical health data under each type of operating condition, it generates initial thresholds for each detection point under different operating conditions, forming a differentiated operating condition early warning strategy. On this basis, the thresholds are dynamically corrected by combining health status parameters, so that the alarm judgment can adaptively reflect the actual degree of equipment health degradation, avoiding false alarms or missed alarms caused by fixed thresholds, thereby improving the accuracy and reliability of fault identification.

[0006] In one optional implementation, the feature data of various characteristics at different detection points include actual rotational speed data, and the operating condition auxiliary data includes load rate data. An enhanced mean clustering algorithm is used, combined with the operating condition auxiliary data of the emergency diesel generator at different times in historical periods, to cluster the feature data of each detection point at different times, obtaining multiple clusters corresponding to each detection point. The operating conditions corresponding to each cluster are then determined, including: Based on the actual rotational speed data of each detection point at different times and the rated rotational speed data of each detection point, the relative rotational speed data of each detection point at different times are determined. Based on the relative speed data of each detection point at different times and the load rate data of the emergency diesel generator at the corresponding times, a clustering parameter vector of each detection point at different times is formed. For each detection point, the enhanced mean clustering algorithm is used to cluster the clustering parameter vectors of the detection points at different times to obtain multiple clusters; For each detection point, the operating conditions corresponding to each cluster are determined based on the clustering parameter vectors in each cluster.

[0007] By enhancing the mean clustering algorithm to divide the clustering parameter vector, multiple physically meaningful operating condition intervals can be obtained. Based on this, independent threshold calculations are performed on each operating condition interval, enabling differentiated judgments using different standards for different operating conditions. Furthermore, this division method comprehensively covers all operational stages of the equipment, from start-up and shutdown to steady state, and from typical to transitional states, effectively eliminating monitoring blind spots and thus improving the accuracy of fault identification.

[0008] In one optional implementation, for each detection point, the initial alarm threshold for each feature under different operating conditions is determined based on the probability density curves of each feature in different clusters, including: Probability density calculations are performed on standardized detection point data corresponding to different working conditions to determine the probability density distribution of various feature data corresponding to each detection point under different working conditions. Based on the probability density distribution of the feature data corresponding to each detection point and the alarm classification rules, the initial alarm threshold corresponding to each feature of each detection point is determined.

[0009] Because different emergency diesel generators have different operating states, and the installation locations and operating environments of the detection points on the same emergency diesel generator are different, the data characteristics collected by each detection point are also different. This solution independently constructs the probability density distribution curve of the operating data for each detection point, and combines it with the preset alarm classification rules to transform the fuzzy judgment criteria into clear numerical thresholds. In this way, the goal of using independent standards for differentiated judgment of each detection point is achieved, which significantly improves the accuracy of fault identification.

[0010] In one optional implementation, the health status parameters include vibration characteristic trend coefficient, speed stability coefficient, and temperature adaptability coefficient. The initial alarm threshold is corrected based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds for each characteristic at each detection point under different operating conditions, including: The health index is determined by weighted summation of the vibration characteristic trend coefficient, rotational speed stability coefficient, and temperature adaptability coefficient. Determine the threshold correction coefficient based on the health index; The initial alarm threshold is corrected based on the threshold correction coefficient to obtain the current alarm threshold for each feature of each detection point under different operating conditions.

[0011] The correction coefficient is calculated using three dimensions: vibration characteristic trend coefficient, rotational speed stability coefficient, and temperature adaptability coefficient. This results in a health index that is more accurate and comprehensive than correction coefficients calculated using a single indicator. By dynamically adjusting the threshold correction coefficient based on the health index, the threshold remains lenient when the equipment is healthy, avoiding false alarms caused by normal fluctuations. When the equipment health value declines, the correction coefficient decreases, the threshold is lowered, and fault signals are more easily detected at an early stage.

[0012] In one optional implementation, the current fault status of each detection point is determined based on the current operating data of the emergency diesel generator and the current alarm thresholds of various characteristics of each detection point under different operating conditions, including: The current operating condition of the emergency diesel generator is determined based on the auxiliary operating data in the current operating data and the characteristic data of different detection points in the emergency diesel generator. The current fault status of the emergency diesel generator is obtained by comparing the feature data of each characteristic in the current operating data with the current alarm threshold corresponding to the current operating condition.

[0013] By combining the clustering parameter ranges and alarm threshold ranges for each detection point under various operating conditions obtained in the above process, this judgment method first judges the operating conditions of the data and then judges the fault status of the data, thus achieving the purpose of judging the corresponding status using different standards and avoiding misjudgments caused by different operating conditions or different equipment health statuses.

[0014] In an optional implementation, if multiple detection points are located on the same component, after determining the initial alarm thresholds for each feature under different operating conditions based on the probability density curves of each feature in different clusters, and before correcting the initial alarm thresholds based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds for each feature of each detection point under different operating conditions, the method further includes: Correlation calculations are performed on the characteristic data of each detection point in the same component under the same working condition to obtain the correlation coefficient; If the correlation coefficient is lower than the correlation threshold, the current initial alarm threshold division is deemed invalid. Check the data quality of the emergency diesel generator's operating data at different times in the historical period. If the operating data is incorrect, reacquire the operating data of the emergency diesel generator at different times in the historical period. Re-cluster the data based on the adjusted operating data to obtain multiple clusters corresponding to each detection point, and determine the operating conditions corresponding to each cluster. For each detection point, determine the initial alarm thresholds for each feature under different operating conditions based on the probability density curves of each feature in different clusters. If the running data is correct, adjust the clustering parameters of the enhanced mean clustering algorithm, re-cluster according to the adjusted enhanced mean clustering algorithm, obtain multiple clusters corresponding to each detection point, and determine the working conditions corresponding to each cluster. For each detection point, determine the initial alarm threshold of each feature under different working conditions according to the probability density curve of each feature in different clusters.

[0015] By performing correlation calculations on the characteristic data of each detection point in the same component under the same operating conditions, it is possible to verify whether the data changes between each detection point are consistent, thereby determining whether the currently generated initial alarm threshold is reasonable and effective. This ensures the quality of the data involved in fault judgment from the source and avoids threshold failure due to problems such as sensor abnormalities or unreasonable division of operating conditions. At the same time, when the correlation coefficient is lower than the set threshold, an error correction path is provided to ensure that the alarm threshold finally used for fault judgment has better reliability.

[0016] Secondly, the present invention provides an emergency diesel generator fault monitoring device, comprising: The data acquisition module is used to acquire the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the feature data of various characteristics of different detection points in the emergency diesel generator. The working condition classification module is used to use the enhanced mean clustering algorithm, combined with the working condition auxiliary data of the emergency diesel generator at different times in the historical period, to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and to determine the working condition corresponding to each cluster. The threshold confirmation module is used to determine the initial alarm threshold for each feature under different operating conditions for each detection point based on the probability density curve of each feature in different clusters. The threshold correction module is used to correct the initial alarm threshold based on the health status parameters of the emergency diesel generator at the current moment, so as to obtain the current alarm threshold of each feature of each detection point under different operating conditions. The fault diagnosis module is used to determine the current fault status of each detection point based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

[0017] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform an emergency diesel generator fault monitoring method according to the first aspect or any corresponding embodiment described above.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute an emergency diesel generator fault monitoring method according to the first aspect or any corresponding embodiment described above.

[0019] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute an emergency diesel generator fault monitoring method according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first step of an emergency diesel generator fault monitoring method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the layout of detection points for an emergency diesel generator set according to an embodiment of the present invention, which is a method for monitoring faults in an emergency diesel generator. Figure 4 This is a structural block diagram of an emergency diesel generator fault monitoring device according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0024] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] As an optional application scenario of this invention, such as Figure 1 As shown, the emergency diesel generator fault monitoring system may include at least one terminal device and at least one server. Figure 1 The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.

[0026] Specifically, the terminal device can be a smartphone, tablet, laptop, PDA, desktop computer, game console, smart TV, smart wearable device, in-vehicle terminal, VR (Virtual Reality) device, AR (Augmented Reality) device, etc. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranet, local area network, wide area network, mobile communication network, and combinations thereof.

[0027] This invention provides a method for monitoring faults in emergency diesel generators. By installing detection points at various parts of the emergency diesel generator and acquiring operational data, and by clustering the operational data acquired from each detection point, the alarm thresholds of each detection point under different operating conditions and at different states of the emergency diesel generator are determined. Based on this, the fault state at the current moment is judged, which solves the problem of insufficient mechanical fault monitoring effect in traditional monitoring methods and achieves more accurate monitoring results.

[0028] According to an embodiment of the present invention, an embodiment of an emergency diesel generator fault monitoring method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0029] This embodiment provides a method for monitoring faults in emergency diesel generators. Figure 2 This is a flowchart of an emergency diesel generator fault monitoring method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the feature data of various characteristics of different detection points in the emergency diesel generator.

[0030] Operational data refers to the data collected during the operation of the emergency diesel generator (EDG) from multiple newly added monitoring points on various components of the high-temperature gas-cooled reactor. Auxiliary operating data refers to supplementary parameters that do not directly reflect the mechanical vibration state of the equipment but describe its operating environment, workload, and physical condition. As an example, auxiliary operating data may include coolant temperature, engine oil temperature, load rate, and operating time.

[0031] For example, the locations of the multiple newly added detection points on various components of the high-temperature gas-cooled reactor emergency diesel generator are as follows: Figure 3 As shown, detection point 12 is a newly added key phase detection point, which can monitor the phase signals of the EDG rotor, crankshaft, and emergency diesel generator rotor, such as the time or interval period when the keyway passes through the sensor, the rotor instantaneous phase angle and the synchronous phase of speed fluctuations calculated in conjunction with the speed, and combined with the phase characteristics of the vibration signal, can locate the specific location of faults such as rotor imbalance, misalignment, and shaft bending, providing accurate phase reference for instantaneous speed detection points, eliminating the cumulative error of speed measurement, and ensuring the speed measurement accuracy at high speeds.

[0032] For example, detection point 11 is a newly added speed detection point, which can monitor the instantaneous speed value during EDG operation, such as dynamic fluctuations of 1500 r / min ± 5 r / min; it can monitor speed fluctuation rate, such as the maximum speed deviation at rated speed, the rate of speed increase and decrease; it can monitor speed step changes, such as the speed climb curve during the start-up phase, and the speed change amplitude during loading and unloading; by monitoring the speed climb rate during the EDG start-up phase, the coordinated working status of the starter motor, fuel injection system, and ignition system can be determined, and the start-up performance can be judged; it can capture speed fluctuations during loading or unloading, such as the emergency load input of a nuclear power plant, in real time, and the governor response speed and stability can be evaluated.

[0033] For example, detection points 5-10 are six newly added cylinder block vibration detection points, which can monitor the vibration acceleration, velocity and displacement of each cylinder block surface, such as covering characteristic frequencies of combustion impact, piston knocking and valve seating, with a frequency range of 10Hz~1kHz; they can monitor vibration peak values ​​and effective values, such as the peak value of impact vibration generated by combustion and the effective value of vibration during steady-state operation; they can monitor vibration spectrum characteristics, such as the combustion main frequency, piston second-order vibration and frequency components of valve mechanism vibration; and they can provide early warning of EDG combustion status, inter-cylinder imbalance and mechanical faults.

[0034] For example, detection points 1-4 are four newly added crankcase vibration detection points, which can monitor the vibration acceleration and velocity of the crankcase housing, such as low-frequency characteristics including crankshaft bearing vibration, connecting rod oscillation, and crankshaft torsional vibration, with a frequency range of 5Hz to 500Hz; they can monitor vibration harmonic components, such as crankshaft rotation frequency, double rotation frequency, and other shaft system characteristic frequencies; they can monitor vibration time-domain waveforms, such as impact pulse signals, reflecting transient events such as bearing impact and connecting rod impact; they can provide early warning of crankshaft faults, and, combined with lubricating oil temperature and oil pressure data, judge the lubrication effect by the crankcase vibration intensity.

[0035] For example, detection points 14 and 15 are two newly added gearbox vibration detection points, which can monitor the vibration acceleration and velocity of the gearbox housing, such as covering gear meshing frequency, bearing frequency, and shaft rotation frequency, with a frequency range of 10Hz~2kHz; they can monitor gear meshing frequency and harmonics, such as the characteristic frequency corresponding to the number of gear teeth × rotational speed, reflecting the gear meshing state; they can monitor vibration kurtosis and peak factor, such as capturing impact signals generated by gear tooth surface wear, pitting, broken teeth and other faults; their main function is to monitor the gear meshing state and provide early warning of gearbox bearing faults.

[0036] For example, detection point 13 is a newly added vibration detection point for the bearing of the emergency diesel generator. It can monitor the vibration acceleration, velocity, and displacement of the bearing housing of the emergency diesel generator, such as covering the bearing frequency, rotor speed, and electromagnetic vibration frequency, with a frequency range of 10Hz to 1kHz; it can monitor the electromagnetic components in the vibration spectrum, such as the 100Hz frequency component generated by the stator electromagnetic force, reflecting the electromagnetic operating condition of the emergency diesel generator; it can monitor the steady-state and transient characteristics of vibration, such as the vibration abrupt change when the emergency diesel generator is loaded and the vibration stability during steady-state operation; it can identify faults such as bearing wear, grease aging, and excessive bearing clearance, and monitor the bearing condition of the emergency diesel generator; combined with the electromagnetic components in the vibration, it can provide early warning of electromagnetic problems such as loose stator windings and uneven rotor air gap, and can perform electromagnetic fault diagnosis.

[0037] For example, noise reduction and normalization calculations can also be performed on the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time.

[0038] As an example, the denoising process can remove outliers from the data. A combination of 3σ and gradient detection is used: first, the 3σ criterion removes discrete points deviating from the mean by three times the standard deviation; then, gradient detection removes abrupt changes such as start-stop shocks. Abnormalities are defined as the difference between adjacent data points greater than five times the mean difference. For vibration-related detection points, wavelet packet filtering can be used, decomposing the data into eight layers to remove irrelevant high-frequency interference above 1kHz and the 50Hz power frequency. For speed or key phase detection points, Kalman filtering can be used to suppress random noise in speed fluctuations.

[0039] As an example, since temperature can interfere with the accuracy of vibration data, it is necessary to remove the influence of temperature to ensure that the vibration data only reflects the state of the equipment itself. Therefore, a fitting relationship between temperature and vibration can be established to correct the vibration data for temperature. The vibration RMS correction coefficient is determined based on the temperature change, and this coefficient is used to correct the measured vibration data. For example, for every 10°C increase in oil temperature, the vibration RMS correction coefficient can be 0.98.

[0040] As an example, the Z-score standardization formula can be used to convert detection point data with different dimensions into dimensionless data, facilitating cross-detection point fusion. The standardization formula can be:

[0041] in, To standardize data, The data is raw data that has not undergone standardized calculations. The data is the mean. The standard deviation data is used.

[0042] Step S202: The enhanced mean clustering algorithm is adopted. The auxiliary data of the emergency diesel generator at different times in the historical period are combined to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and the working conditions corresponding to each cluster are determined.

[0043] Enhanced mean clustering is an unsupervised machine learning method whose core objective is to automatically divide a dataset into a predetermined number of categories based on the similarity between data points. The algorithm uses the mean of the data points in each category as the centroid, representing the overall characteristics of that category. Through iterative calculations, the algorithm continuously optimizes the position of the centroids, ensuring that data points within the same category are as close to their centroids as possible, while centroids in different categories are as far apart as possible.

[0044] By enhancing the mean clustering algorithm, the feature data of each detection point at different times can be divided into multiple classes based on the similarity of the feature data itself and the similarity of the auxiliary data of the working conditions. The set of data in each class is called a cluster. The number of clusters is a preset value, and each cluster represents a specific working condition.

[0045] Step S203: For each detection point, determine the initial alarm threshold for each feature under different operating conditions based on the probability density curves of each feature in different clusters.

[0046] A probability density curve is a graph used to describe the frequency or probability of a class of data across all possible values. In this embodiment, it is used to depict the normal fluctuation range and the probability of extreme values ​​occurring at each detection point of the equipment under specific operating conditions. The initial alarm threshold is a preliminary boundary between abnormal data and normal operating data calculated statistically based on this distribution curve, used to identify potential abnormal states.

[0047] Step S204: Correct the initial alarm threshold based on the health status parameters of the emergency diesel generator at the current moment, and obtain the current alarm threshold of each feature of each detection point under different operating conditions.

[0048] Health status parameters are comprehensive indicators used to quantify the current health level of equipment. Since equipment operating data changes with its health status, and initial alarm thresholds are set based on operating data when the equipment is healthy, it is necessary to dynamically adjust the current alarm thresholds according to the equipment's current health status when determining equipment anomalies.

[0049] For example, alarm thresholds can be divided into warning thresholds, alarm thresholds, and emergency alarm thresholds.

[0050] Step S205: Determine the current fault status of each detection point based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

[0051] This process serves as the final execution step in real-time fault monitoring. It compares the data monitored in real-time at each detection point with various alarm thresholds and determines the current fault status based on the threshold range into which the data monitored in real-time at each detection point falls.

[0052] This embodiment provides a fault monitoring method for emergency diesel generators. By adding additional detection points to key parts of the emergency diesel generator, it achieves the goal of monitoring mechanical faults that cannot be considered in traditional fault prediction methods. The system inputs the operating data collected from each detection point into a clustering algorithm to automatically classify different operating conditions such as startup, steady state, and transition. Based on historical health data under each type of operating condition, it generates initial thresholds for each detection point under different operating conditions, forming a differentiated operating condition early warning strategy. On this basis, the thresholds are dynamically corrected by combining health status parameters, so that the alarm judgment can adaptively reflect the current actual health degradation of the equipment, avoiding false alarms or missed alarms caused by fixed thresholds, thereby improving the accuracy and reliability of fault identification.

[0053] In an optional embodiment, the feature data of various characteristics at different detection points include actual rotational speed data, and the operating condition auxiliary data includes load rate data. Step S202 involves using an enhanced mean clustering algorithm, combining the operating condition auxiliary data of the emergency diesel generator at different times in historical periods, to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and determining the operating conditions corresponding to each cluster. Specifically, this includes: Step a1: Based on the actual rotational speed data of each detection point at different times and the rated rotational speed data of each detection point, determine the relative rotational speed data of each detection point at different times.

[0054] Actual speed data refers to the real rotational speed of the rotating components of an emergency diesel generator at every instant during operation. Rated speed data is a fixed design parameter, referring to the target speed value specified in the design of the emergency diesel generator set for long-term stable operation. Relative speed data refers to the dimensionless ratio obtained by dividing the actual speed of the emergency diesel generator by the rated speed, used to describe the degree to which the current equipment's actual speed reaches the rated target. As an example, relative speed data can be calculated by dividing the actual speed by the rated speed.

[0055] Step a2: Based on the relative speed data of each detection point at different times and the load rate data of the emergency diesel generator at the corresponding times, a clustering parameter vector of each detection point at different times is formed.

[0056] Load factor data refers to the ratio between the actual power output of an emergency diesel generator at the current moment and its maximum rated power, used to represent the current output level of the emergency diesel generator. Clustering parameter vectors are a set of vectors formed by combining multiple key feature parameters describing the equipment's operating status, creating multidimensional data points that serve as input values ​​for the clustering algorithm.

[0057] For example, the clustering parameter vector may include, in addition to the relative speed data and load rate data of each detection point at different times, other auxiliary data to assist the clustering algorithm in clustering. If only the relative speed data of each detection point at different times and the load rate data of the emergency diesel generator at the corresponding times are used for clustering, although the external load state of the equipment can be distinguished, the internal thermodynamic state of the equipment cannot be distinguished. In EDG, the vibration, clearance, and friction characteristics are completely different under the same speed and load, depending on the thermal state. Therefore, by introducing a temperature parameter, the clustering algorithm can distinguish between cold steady state and hot steady state, thereby avoiding deviations in the generation of thresholds and associated features in subsequent processes.

[0058] Engine oil temperature characterizes lubrication status and mechanical clearances. Engine oil viscosity changes significantly with temperature. At low temperatures, engine oil is viscous, forming a thick film but with poor flowability and high damping; at high temperatures, engine oil is thin, forming a thin film but with good flowability. Changes in oil film damping alter the vibration energy transmission path, affecting vibration transmission. The clearance between the piston and cylinder liner also changes with temperature; a large clearance in a cold state can easily cause knocking, while a small clearance in a hot state results in smoother operation and affects impact energy. In clustering algorithms, engine oil temperature can be used to distinguish between the initial startup state (oil temperature around 30℃) and the hot standby state (oil temperature around 60℃). Both states should have the rated speed S≈1 and the load 0. Due to the different oil temperatures, their vibration baselines should also differ.

[0059] Cooling water temperature can characterize heat load and engine block thermal deformation, reflecting the thermal load status of components such as cylinder liners and cylinder heads. As temperature rises, large castings such as the engine block and crankshaft undergo thermal expansion, causing micron-level thermal deformation in critical fits such as bearing clearances and alignment. Under conditions such as emergency starts, the entry of cold water into the engine can trigger drastic temperature changes, generating thermal shock, which in turn creates thermal stress and may even induce cracks. In clustering algorithms, cooling water temperature helps identify the current heat load of the unit.

[0060] As an intermittent emergency device, the body temperature of an EDG (Emergency Controller) often lags behind changes in speed and load. Incorporating temperature parameters can effectively distinguish between different thermodynamic stages such as cold start, hot start, and steady-state thermal equilibrium, avoiding the mixing of vibration data caused by different thermal states under the same speed and load.

[0061] Vibration signals are sensitive to temperature. Temperature clustering ensures that the physical characteristics of the data, such as mechanical clearance and oil film damping, remain consistent within the same operating condition cluster, thus providing a more solid physical basis for subsequent feature correlation verification and dynamic threshold generation.

[0062] When equipment malfunctions, by observing the changes in temperature and load in the parameter vector, it is possible to make a preliminary judgment as to whether the fault originates from mechanical components or the thermal system.

[0063] As an example, the remaining auxiliary data used to help the clustering algorithm perform clustering could be coolant temperature and engine oil temperature. Therefore, the clustering parameter vector can be represented as... ,in, This is relative rotational speed data. This represents the actual load factor. For cooling water temperature, This refers to the engine oil temperature. As an example, the relative speed data can be calculated based on the ratio of the current speed to the rated speed, and the actual load rate can be calculated based on the ratio of the current power to the rated power, with a value range of 0 to 1.

[0064] Step a3: For each detection point, the enhanced mean clustering algorithm is used to cluster the clustering parameter vectors of the detection points at different times to obtain multiple clusters.

[0065] An enhanced mean clustering algorithm is used to divide the clustering parameter vectors of the detection points at different times into multiple clusters. The number of clusters is the same as the preset number of working condition classifications.

[0066] For example, in the reinforced mean clustering algorithm, each clustering parameter vector is treated as a data point. The algorithm first randomly selects a point from all data points as the first initial cluster center. For the remaining points, it calculates the shortest distance from each point to an existing center. The greater the distance, the higher the probability of the point being selected as the next center. This process is repeated until a preset number of centers are selected. For the iterative process: the feature vector of each data point is traversed, and the Euclidean distance from that point to each center is calculated. The point is then assigned to the cluster containing the nearest center. For each cluster, the average value of the feature vectors of all data points within that cluster is calculated, and the average value replaces the original center as the new center. This iterative process is repeated until the change in the center's position is less than a set threshold or no longer changes, resulting in a classification result for multiple clustering parameter vectors. Each cluster contains multiple clustering parameter vectors.

[0067] Step a4: For each detection point, determine the working condition corresponding to each cluster based on the clustering parameter vector in each cluster.

[0068] Based on the fact that each cluster contains multiple clustering parameter vectors, the operating condition type to which each cluster belongs is determined.

[0069] For example, operating conditions can be categorized according to the following table, where, This is relative rotational speed data. The actual load rate is given by Tc, where Tc is the coolant temperature, To is the oil temperature, and ds / dt is the rate of temperature change.

[0070] As an example, the range of clustering parameter vectors corresponding to the partitioned operating conditions can be: Start-up operating condition: Steady-state operating conditions: Sudden load change conditions: Shutdown conditions: .in, This is relative rotational speed data. This represents the actual load rate. Furthermore, the process from startup to steady-state operation can be further subdivided into three transitional conditions: the ignition burst period, the acceleration period, and the pre-grid connection stabilization period. The ignition burst period represents the process of speed increasing from 0 with a large impact load; the acceleration period represents the process of linear speed increase and gradual oil film establishment; and the pre-grid connection stabilization period represents the process of speed approaching the rated speed. Load mutation conditions can be further subdivided into the instantaneous mutation, the adjustment period, and the recovery to steady-state period. The instantaneous mutation represents a millisecond-level impact response process; the adjustment period represents the governor response stage; and the recovery to steady-state period represents entering a new steady-state stage.

[0071] This embodiment provides a fault monitoring method for emergency diesel generators. By using an enhanced mean clustering algorithm to divide the clustering parameter vector, multiple physically meaningful operating condition intervals can be obtained. Based on this, independent threshold calculations are performed on each operating condition interval, enabling differentiated judgments using different standards for different operating conditions. Furthermore, this division method comprehensively covers all operating stages of the equipment, from start-up and shutdown to steady state, and from typical to transitional states, effectively eliminating monitoring blind spots and thus improving the accuracy of fault identification.

[0072] In an optional embodiment, step S203, for each detection point, determines the initial alarm threshold for each feature under different operating conditions based on the probability density curves of each feature in different clusters, specifically including: Step b1 involves calculating the probability density of standardized detection point data corresponding to different working conditions to determine the probability density distribution of various feature data corresponding to each detection point under different working conditions.

[0073] Probability density calculation is a method used to describe the likelihood of a continuous random variable occurring around a specific value. The probability density distribution refers to the result obtained by calculating the probability density of standardized test point data. A probability density curve is a smooth curve describing the probability of continuous random data occurring around various values; it is a visual representation of the probability density distribution.

[0074] For example, probability density calculation can be performed using kernel density estimation (KDE) to obtain the probability density curve.

[0075] Step b2: Determine the initial alarm threshold corresponding to each feature of each detection point based on the probability density distribution of each feature data and the alarm classification rules.

[0076] The alarm classification rule is a preset threshold classification standard used to determine the position of the initial alarm threshold in the probability density distribution. The initial alarm threshold is a specific value extracted directly from the probability density distribution according to the alarm classification rule. It is the original alarm threshold limit without subsequent dynamic correction.

[0077] For example, the alarm classification rule can be: take the 95th percentile of the probability density distribution as the basic early warning threshold for the operating condition range, the 99th percentile as the basic alarm threshold, and the 99.9th percentile as the basic emergency alarm threshold.

[0078] This embodiment provides a fault monitoring method for emergency diesel generators. Due to the differences in the operating states of different emergency diesel generators, and the different installation locations and operating environments of the detection points on the same emergency diesel generator, the data characteristics collected by each detection point are also different. This solution independently constructs the probability density distribution curve of the operating data for each detection point, and combines it with preset alarm classification rules to transform the fuzzy judgment criteria into clear numerical thresholds. Thus, it achieves the purpose of using independent standards for differentiated judgment at each detection point, significantly improving the accuracy of fault identification.

[0079] In an optional embodiment, the health status parameters include vibration characteristic trend coefficient, speed stability coefficient, and temperature adaptability coefficient. Step S204 involves correcting the initial alarm threshold based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds for each feature at each detection point under different operating conditions. Specifically, this includes: Step c1 involves weighted summation of the vibration characteristic trend coefficient, rotational speed stability coefficient, and temperature adaptability coefficient to determine the health index.

[0080] The vibration characteristic trend coefficient is a quantitative indicator describing the trend of equipment vibration level changes over a recent period of time. The speed stability coefficient is a quantitative indicator describing the degree of fluctuation of the actual speed of the emergency diesel generator relative to the rated speed. The temperature adaptability coefficient is a quantitative indicator describing the degree of deviation of the current operating temperature of the equipment from its optimal operating temperature. The health index is an indicator describing the current health status of the equipment.

[0081] Vibration characteristic trend coefficient, speed stability coefficient, and temperature adaptability coefficient are core indicators for judging the health status of equipment. They are calculated from three dimensions: mechanical state, control accuracy, and thermodynamic state. The health status of the equipment is calculated by combining the importance weights of the three dimensions affecting the health status, thereby obtaining a health index to describe the health status of the equipment.

[0082] For example, the vibration characteristic trend coefficient can be calculated using the following formula:

[0083] in, This is the vibration characteristic trend coefficient. The slope of the linear regression. The reference slope is used for normalization. Wherein, It is obtained by performing linear regression calculations on time and vibration data, with a reference slope. The average degradation slope measured before historical failures can be used, or it can be set to the maximum allowable degradation rate of the equipment. express , express As an example, the maximum allowable degradation rate of the device can be an increase of 0.01 mm / s in vibration data per day, and the vibration data can be taken from nearly 90 days of vibration data.

[0084] As an example, if a reference slope cannot be obtained... The vibration characteristic trend coefficient can then be calculated using the following formula:

[0085] in, This is the vibration characteristic trend coefficient. The slope of the linear regression. This represents the maximum degradation slope in historical statistics.

[0086] For example, the rotational speed stability coefficient can be calculated using the following formula:

[0087] in, This is the rotational speed stability coefficient. The standard deviation of the rotational speed represents the magnitude of the absolute fluctuation. The average rotational speed within the window represents the magnitude of the absolute fluctuation. For the maximum allowable relative fluctuation, This is the coefficient of variation, used to eliminate dimensions. As an example, it represents the maximum permissible relative fluctuation. The value can be set to a speed fluctuation rate of no more than 0.5%, i.e., 0.005.

[0088] For example, the temperature adaptability coefficient can be calculated according to the following formula:

[0089] in, Temperature adaptability coefficient, For optimal operating temperature, The maximum allowable deviation range, This is the current temperature data. As an example, The value can be ,in These are the upper and lower boundaries of the optimal operating temperature range. The value can be As an example, if the optimal temperature for engine oil is 80°C, a fluctuation of ±10°C is permissible. The current measured temperature is 85℃, then If the current temperature is 95℃, .

[0090] For example, the health index can be calculated using the following formula:

[0091] in, For health index, The trend weight of vibration characteristics, This is the vibration characteristic trend coefficient. As the weight for rotational speed stability, This is the rotational speed stability coefficient. For temperature adaptability weight, This is the temperature adaptability coefficient. As an example, The value can be 0.4. The value can be 0.3. The value can be 0.3.

[0092] Step c2: Determine the threshold correction coefficient based on the health index.

[0093] The threshold correction factor is a dynamic adjustment factor calculated based on the current health index of the device. It is used to adjust the initial alarm threshold to a current alarm threshold that is more consistent with the current health status of the device.

[0094] For example, it can be done by Calculate the threshold correction coefficient, where, As a health index.

[0095] Step c3: Correct the initial alarm threshold according to the threshold correction coefficient to obtain the current alarm threshold of each feature of each detection point under different working conditions.

[0096] The current alarm threshold refers to the final alarm threshold obtained after correcting the initial alarm threshold with a threshold correction coefficient, and is used to determine anomalies.

[0097] For example, it can be done by The final alarm threshold is calculated. As an example, after adjusting to obtain the final alarm threshold, the final alarm threshold also needs to meet the constraint that the corrected threshold must not exceed ±30% of the initial manual threshold. The initial manual threshold is a safety threshold manually set based on a safety baseline when the device has no historical data, used to ensure safety redundancy.

[0098] This embodiment provides a fault monitoring method for emergency diesel generators. It calculates the correction coefficient using three dimensions: vibration characteristic trend coefficient, speed stability coefficient, and temperature adaptability coefficient. This results in a health index that is more accurate and comprehensive than correction coefficients calculated using a single indicator. By dynamically adjusting the threshold correction coefficient based on the health index, the threshold remains lenient when the equipment is healthy, avoiding false alarms caused by normal fluctuations. When the equipment health value declines, the correction coefficient decreases, the threshold is lowered, and fault signals are more easily detected at an early stage.

[0099] In an optional embodiment, step S205, determining the current fault status of each detection point based on the current operating data of the emergency diesel generator and the current alarm thresholds of various characteristics of each detection point under different operating conditions, specifically includes: Step d1: Determine the current operating condition of the emergency diesel generator based on the auxiliary operating data in the current operating data and the characteristic data of different detection points in the emergency diesel generator.

[0100] The clustering parameter vector, which is composed of the auxiliary operating data and the feature data of different detection points in the emergency diesel generator collected at the current moment, is compared with the clustering parameter intervals corresponding to various operating conditions obtained by classification, so as to determine the operating condition corresponding to the current data features.

[0101] Step d2: Compare the feature data of each feature in the current operating data with the current alarm threshold corresponding to the current operating condition to obtain the current fault status of the emergency diesel generator.

[0102] Based on the current operating conditions, the feature data collected by each detection point at the current moment is compared with the current alarm threshold corresponding to the current operating conditions in a hierarchical manner to determine the alarm threshold range into which the current measured feature data falls, thereby obtaining the final judgment result of the current fault status of the emergency diesel generator.

[0103] This embodiment provides an emergency diesel generator fault monitoring method. By combining the clustering parameter ranges and alarm threshold ranges of each detection point under various operating conditions obtained in the above process, the method first judges the operating conditions of the data and then judges the fault status of the data. This achieves the purpose of judging the corresponding status using different standards, avoiding misjudgments caused by different operating conditions or different equipment health statuses.

[0104] In an optional embodiment, if multiple detection points are located on the same component, after determining the initial alarm thresholds for each feature under different operating conditions based on the probability density curves of each feature in different clusters, and before correcting the initial alarm thresholds based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds for each feature of each detection point under different operating conditions, the method further includes: Step e1: Calculate the correlation of various feature data of each detection point in the same component under the same working condition to obtain the correlation coefficient.

[0105] This process, which involves verification and feedback after generating the initial alarm threshold and before dynamic correction, is used to ensure data quality and enhance the effectiveness of the mean clustering algorithm.

[0106] For the same component, although the installation locations of the detection points may differ, because they are located on the same component, ideally, after working condition clustering, the corresponding data collected by each detection point under the same working condition should meet certain correlation requirements. Based on the above physical foundation, after generating the initial alarm threshold, the correlation of various characteristic data of each detection point of the same component under each working condition can be calculated to verify the rationality of the threshold. The correlation coefficient is a statistical indicator that measures the degree of linear correlation between two variables and is the result obtained from the correlation calculation.

[0107] Step e2: If the correlation coefficient is lower than the correlation threshold, the current initial alarm threshold division is determined to be invalid.

[0108] If the correlation coefficient is lower than the correlation threshold, it indicates that the original data may contain errors, or it may indicate that the working condition division results obtained by the clustering algorithm are unreasonable, causing the operating data that should belong to other working conditions to be incorrectly classified into the current working condition. Therefore, it is judged that the current initial alarm threshold division is incorrect.

[0109] For example, the correlation coefficient can be the Pearson correlation coefficient.

[0110] For example, the correlation threshold can be 0.7. That is, the criterion for the correlation coefficient can be expressed as r ≥ 0.7, where r is the correlation coefficient.

[0111] Step e3: Check the data quality of the emergency diesel generator's operating data at different times in the historical period. If the operating data is incorrect, reacquire the operating data of the emergency diesel generator at different times in the historical period. Re-cluster the data based on the adjusted operating data to obtain multiple clusters corresponding to each detection point, and determine the operating conditions corresponding to each cluster. For each detection point, determine the initial alarm threshold of each feature under different operating conditions based on the probability density curve of each feature in different clusters.

[0112] Since the possibility of erroneous operational data is higher than the possibility of errors in the enhanced mean clustering algorithm, the data quality of operational data at different times can be checked first. If the data quality is incorrect, the operational data of the emergency diesel generator at different times in the historical period should be reacquired, and then the initial alarm threshold should be redefined. The method for determining the initial alarm threshold based on the operational data at different times is the same as steps S202 and S203, and will not be repeated here.

[0113] Step e4: If the running data is correct, adjust the clustering parameters of the enhanced mean clustering algorithm, re-cluster according to the adjusted enhanced mean clustering algorithm, obtain multiple clusters corresponding to each detection point, and determine the working conditions corresponding to each cluster. For each detection point, determine the initial alarm threshold of each feature under different working conditions according to the probability density curve of each feature in different clusters.

[0114] If the running data is correct, it indicates that the enhanced mean clustering algorithm has made a mistake in its partitioning. In this case, the clustering parameters need to be adjusted and the running data needs to be re-clustered. Based on the re-clustered operating conditions, the initial alarm thresholds for different operating conditions should be determined. The method for determining the initial alarm thresholds based on the running data at different times is the same as steps S202 and S203, and will not be repeated here.

[0115] This embodiment provides an emergency diesel generator fault monitoring method. By performing correlation calculations on various characteristic data of each detection point in the same component under the same operating conditions, it can verify whether the data changes between each detection point are consistent, thereby determining whether the currently generated initial alarm threshold is reasonable and effective. This ensures the quality of data involved in fault judgment from the source and avoids threshold failure due to problems such as sensor abnormalities or unreasonable division of operating conditions. At the same time, when the correlation coefficient is lower than the set threshold, an error correction path is provided to ensure that the alarm threshold finally used for fault judgment has better reliability.

[0116] This embodiment also provides an emergency diesel generator fault monitoring device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0117] This embodiment provides an emergency diesel generator fault monitoring device, such as... Figure 4 As shown, it includes: The data acquisition module 401 is used to acquire the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the feature data of various characteristics of different detection points in the emergency diesel generator. The working condition classification module 402 is used to use the enhanced mean clustering algorithm, combined with the working condition auxiliary data of the emergency diesel generator at different times in the historical period, to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and to determine the working condition corresponding to each cluster. The threshold confirmation module 403 is used to determine the initial alarm threshold of each feature under different working conditions for each detection point based on the probability density curve of each feature in different clusters. The threshold correction module 404 is used to correct the initial alarm threshold based on the health status parameters of the emergency diesel generator at the current moment, so as to obtain the current alarm threshold of each feature of each detection point under different operating conditions. The fault diagnosis module 405 is used to determine the current fault status of each detection point based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

[0118] The emergency diesel generator fault monitoring device provided in this embodiment of the invention can execute the emergency diesel generator fault monitoring method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0119] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0120] The following is a detailed reference. Figure 5 The diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 501, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 502 or a program loaded from memory 508 into random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the electronic device. The processor 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0121] Typically, the following devices can be connected to I / O interface 505: input devices 506 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 507 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 508 including, for example, magnetic tapes, hard disks, etc.; and communication devices 509. Communication device 509 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0122] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 509, or installed from a memory 508, or installed from a ROM 502. When the computer program is executed by the processor *01, it performs the functions defined in an emergency diesel generator fault monitoring method according to an embodiment of the present invention.

[0123] Figure 5 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0124] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, it implements the emergency diesel generator fault monitoring method shown in the above embodiments.

[0125] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0126] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for monitoring faults in an emergency diesel generator, characterized in that, The method includes: The system acquires the operating data of the emergency diesel generator at different times in a historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the feature data of various characteristics of different detection points in the emergency diesel generator. The enhanced mean clustering algorithm is adopted, and the auxiliary data of the emergency diesel generator at different times in the historical period are combined to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and the operating conditions corresponding to each cluster are determined. For each detection point, the initial alarm threshold for each feature under different operating conditions is determined based on the probability density curves of each feature in different clusters. The initial alarm threshold is corrected based on the current health status parameters of the emergency diesel generator to obtain the current alarm thresholds of each feature of each detection point under different operating conditions. The current fault status of each detection point is determined based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

2. The method according to claim 1, characterized in that, The feature data of each characteristic at different detection points includes actual rotational speed data, and the operating condition auxiliary data includes load rate data. An enhanced mean clustering algorithm is used, combined with the operating condition auxiliary data of the emergency diesel generator at different times in historical periods, to cluster the feature data of each detection point at different times, obtaining multiple clusters corresponding to each detection point, and determining the operating conditions corresponding to each cluster, including: Based on the actual rotational speed data of each detection point at different times and the rated rotational speed data of each detection point, the relative rotational speed data of each detection point at different times are determined. Based on the relative speed data of each detection point at different times and the load rate data of the emergency diesel generator at the corresponding times, a clustering parameter vector of each detection point at different times is formed. For each detection point, the enhanced mean clustering algorithm is used to cluster the clustering parameter vectors of the detection points at different times to obtain multiple clusters; For each detection point, the operating conditions corresponding to each cluster are determined based on the clustering parameter vectors in each cluster.

3. The method according to claim 1, characterized in that, For each detection point, the initial alarm threshold for each feature under different operating conditions is determined based on the probability density curves of each feature in different clusters, including: The probability density of the standardized detection point data corresponding to the different working conditions is calculated to determine the probability density distribution of each feature data corresponding to each detection point under different working conditions. Based on the probability density distribution of the feature data corresponding to each detection point and the alarm classification rules, the initial alarm threshold corresponding to each feature of each detection point is determined.

4. The method according to claim 1, characterized in that, The health status parameters include vibration characteristic trend coefficient, speed stability coefficient, and temperature adaptability coefficient. The initial alarm threshold is corrected based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds for each characteristic at each detection point under different operating conditions, including: The health index is determined by weighted summation of the vibration characteristic trend coefficient, rotational speed stability coefficient, and temperature adaptability coefficient. Determine the threshold correction coefficient based on the health index; The initial alarm threshold is corrected according to the threshold correction coefficient to obtain the current alarm threshold of each feature of each detection point under different operating conditions.

5. The method according to claim 1, characterized in that, The process of determining the current fault status of each detection point based on the current operating data of the emergency diesel generator and the characteristics of each detection point under different operating conditions and current alarm thresholds includes: The current operating condition of the emergency diesel generator is determined based on the operating condition auxiliary data in the current operating data and the characteristic data of different detection points in the emergency diesel generator. The current fault status of the emergency diesel generator is obtained by comparing the feature data of each characteristic in the current operating data with the current alarm threshold corresponding to the current operating condition.

6. The method according to claim 1, characterized in that, If multiple detection points are located on the same component, after the step of determining the initial alarm thresholds of each feature under different operating conditions based on the probability density curves of each feature in different clusters, and before the step of correcting the initial alarm thresholds based on the health status parameters of the emergency diesel generator at the current moment to obtain the current alarm thresholds of each feature at each detection point under different operating conditions, the method further includes: Correlation calculations are performed on the characteristic data of various detection points in the same component under the same working conditions to obtain the correlation coefficient; If the correlation coefficient is lower than the correlation threshold, the current initial alarm threshold division is deemed invalid. The data quality of the emergency diesel generator's operating data at different times in the historical period is checked. If the operating data is incorrect, the operating data of the emergency diesel generator at different times in the historical period is reacquired. The adjusted operating data is then re-clustered to obtain multiple clusters corresponding to each detection point, and the operating conditions corresponding to each cluster are determined. For each detection point, the initial alarm thresholds for each feature under different operating conditions are determined based on the probability density curves of each feature in different clusters. If the running data is correct, the clustering parameters of the enhanced mean clustering algorithm are adjusted, and the clusters are re-clustered according to the adjusted enhanced mean clustering algorithm to obtain multiple clusters corresponding to each detection point, and the working conditions corresponding to each cluster are determined; for each detection point, the initial alarm thresholds of each feature under different working conditions are determined according to the probability density curves of each feature in different clusters.

7. An emergency diesel generator fault monitoring device, characterized in that, The device includes: The data acquisition module is used to acquire the operating data of the emergency diesel generator at different times in the historical period, the health status parameters at the current time, and the operating data at the current time. The operating data includes the auxiliary operating data of the emergency diesel generator and the feature data of various features of different detection points in the emergency diesel generator. The working condition classification module is used to use the enhanced mean clustering algorithm, combined with the working condition auxiliary data of the emergency diesel generator at different times in the historical period, to cluster the feature data of each detection point at different times to obtain multiple clusters corresponding to each detection point, and to determine the working condition corresponding to each cluster. The threshold confirmation module is used to determine the initial alarm threshold for each feature under different operating conditions for each detection point based on the probability density curve of each feature in different clusters. The threshold correction module is used to correct the initial alarm threshold based on the health status parameters of the emergency diesel generator at the current moment, so as to obtain the current alarm threshold of each feature of each detection point under different operating conditions. The fault diagnosis module is used to determine the current fault status of each detection point based on the current operating data of the emergency diesel generator and the current alarm thresholds of each detection point under different operating conditions.

8. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform an emergency diesel generator fault monitoring method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute any one of claims 1 to 6 of the emergency diesel generator fault monitoring method.

10. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute any one of claims 1 to 6, a method for monitoring faults in an emergency diesel generator.

Images