Oil monitoring fault diagnosis method and system based on Internet of Things

Abstract

The invention relates to the technical field of fault diagnosis, in particular to an oil monitoring fault diagnosis method and system based on the Internet of Things, and the method comprises the following steps: collecting a mechanical vibration signal through oil monitoring, setting a time window, executing short-time Fourier transform, converting a time domain signal into a frequency domain signal, and obtaining frequency spectrum recognition data. According to the method, batch data can be processed in real time through the Internet of Things technology, a classifier is trained through a machine learning method, the fault type recognition precision is further improved, the operation state and efficiency of the machine are allowed to be evaluated through introduction of entropy calculation, and through comparison with a set fault level standard, the fault type recognition accuracy is improved. The fault risk of the machine is found and sorted in time, and the combination of real-time monitoring and automatic identification greatly improves the capability of prevention and maintenance and the operation efficiency of the equipment, reduces the risk of accidental shutdown, and brings remarkable benefits in the aspects of cost control and productivity improvement.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to an oil monitoring fault diagnosis method and system based on the Internet of Things. Background Art

[0002] The field of fault diagnosis technology focuses on identifying potential problems by detecting anomalies in equipment or systems. It involves collecting and analyzing various types of data, including physical parameters such as temperature, pressure, vibration, etc., to determine whether the equipment is working properly or needs repair. Fault diagnosis methods can be based on traditional sensor technology or integrate modern information technology such as the Internet of Things, big data, and machine learning to improve the accuracy and efficiency of diagnosis. The application of technology can greatly reduce downtime, improve production efficiency and equipment life.

[0003] Among them, the IoT oil monitoring fault diagnosis method refers to the use of IoT technology to monitor and analyze the oil status in mechanical equipment, such as oil temperature, viscosity, and impurity content. By monitoring the changes in oil in real time, the failure and wear of mechanical equipment can be discovered in time. The main uses of this technology include extending the service life of mechanical equipment, preventing sudden failures, and optimizing maintenance plans to achieve cost-effectiveness and optimization of operating efficiency in industrial production. Summary of the invention

[0004] In order to solve the technical problems that the existing technology relies on traditional sensors and periodic detection, and has significant deficiencies in real-time and fault prevention. It cannot timely reflect the dynamic changes of mechanical equipment in actual operation, delays the discovery of faults, causes equipment damage or production line shutdown, and further affects the overall production efficiency and maintenance costs. In terms of oil monitoring, traditional technology cannot reflect slight changes in oil quality in real time. If the changes are not detected in time, mechanical parts will wear out or fail prematurely. In a production environment that relies heavily on efficiency and stability, this deficiency leads to significant economic losses and safety risks. An embodiment of the present invention provides an oil monitoring fault diagnosis method and system based on the Internet of Things. The technical solution is as follows:

[0005] On the one hand, a method for oil monitoring fault diagnosis based on the Internet of Things is provided, the method comprising:

[0006] S1: Collect mechanical vibration signals through oil monitoring, set the time window, perform short-time Fourier transform, convert time domain signals into frequency domain signals, and obtain spectrum recognition data;

[0007] S2: Based on the spectrum recognition data, analyze the energy spectrum density, accumulate energy of the data within the target frequency range, identify the frequency characteristics associated with the mechanical fault, and obtain the fault indication frequency;

[0008] S3: using the fault indication frequency, by calculating the degree of change in the differentiated time series, identifying an abnormal mode that deviates from a normal vibration mode, and obtaining an abnormal vibration mode;

[0009] S4: using the abnormal vibration pattern as input data, using the Internet of Things to train a machine learning classifier, identifying the fault type in the vibration data by setting target parameters, and obtaining a fault type identification result;

[0010] S5: using the fault type identification result, extracting fault samples of the oil monitoring data, executing multiple stages of entropy value calculation, calculating the data entropy at differentiated time points, recording entropy value data, evaluating the operating speed and efficiency of the machine, and obtaining entropy value data records;

[0011] S6: Based on the entropy data record, the data is compared with the pre-set fault level standard, the health status of the machine is identified through threshold judgment, and the processing priority of the mechanical failure risk is sorted to obtain the fault risk assessment result.

[0012] As a further solution of the present invention, the spectrum identification data includes the key frequency, corresponding amplitude and spectrum distribution of the vibration signal, the fault indication frequency includes the key fault frequency, associated energy peak and frequency width, the vibration abnormality mode includes the abnormal frequency mode, deviation degree and mode stability, the fault type identification result includes the identified fault type, fault severity and credibility score, the entropy value data record includes the calculated entropy value, entropy value change rate and time series entropy value comparison, and the fault risk assessment result includes the assessed fault level, risk ranking and expected maintenance time point.

[0013] As a further solution of the present invention, the steps of collecting mechanical vibration signals through oil monitoring, setting a time window, performing short-time Fourier transform, converting time domain signals into frequency domain signals, and obtaining spectrum recognition data are specifically as follows:

[0014] S101: Collect mechanical vibration signals through oil monitoring, set time window parameters, intercept the vibration signals in segments, check the signal quality by adjusting the signal interception, and perform normalization processing to adjust the amplitude of the intercepted signal to generate segmented time domain signals;

[0015] S102: Based on the segmented time domain signal, using short-time Fourier transform technology, setting the frequency resolution in the conversion process, and calculating the frequency domain characteristics of the signal in the differentiated time window by dynamically adjusting the conversion window length to obtain frequency domain signal data;

[0016] S103: Using the frequency domain signal data, perform spectrum recognition, filter frequency components through peak detection, combine signal strength and frequency position, identify key spectrum features, and obtain spectrum recognition data.

[0017] As a further solution of the present invention, the formula of the short-time Fourier transform technology is as follows:

[0018]

[0019] Among them, X′(ω) is the frequency domain signal data, ω is the angular frequency, t is the time offset, n is the discrete time index, x(n) is the signal amplitude at time point n, w(nt) is the window function with a width of w centered at t, α is the time window weight attenuation coefficient, β is the frequency window weight attenuation coefficient, and e is a natural constant.

[0020] As a further solution of the present invention, based on the spectrum recognition data, the energy spectrum density is analyzed, the energy of the data within the target frequency range is accumulated, the frequency characteristics associated with the mechanical fault are identified, and the steps of obtaining the fault indication frequency are specifically as follows:

[0021] S201: Based on the spectrum identification data, perform energy spectrum density analysis, sort out the energy of each frequency component, evaluate the energy contribution of each frequency point by signal strength, combine the frequency point energies, and obtain a cumulative energy spectrum;

[0022] S202: defining a target frequency range through the accumulated energy spectrum, iteratively aggregating and analyzing the accumulated energy data within the target range, identifying frequency regions with energy anomalies by setting thresholds, and generating target frequency analysis results;

[0023] S203: Using the target frequency analysis result, the identified abnormal frequency characteristics are compared with the known mechanical fault frequency characteristics, and the frequency characteristics associated with the mechanical fault are identified through matching analysis to generate a fault indication frequency.

[0024] As a further solution of the present invention, the fault indication frequency is used to identify an abnormal mode that deviates from a normal vibration mode by calculating the degree of change in a differentiated time series, and the steps of obtaining the abnormal vibration mode are specifically as follows:

[0025] S301: Based on the fault indication frequency, identifying the frequency difference between the current vibration data of the mechanical equipment and the standard vibration mode, analyzing the degree of change of each fault indication frequency through a differential time series, and obtaining a frequency difference time series;

[0026] S302: using the frequency difference time series, performing time window analysis on the data in the series, distinguishing between normal and abnormal vibration modes by calculating the degree of deviation of the vibration mode in each time window, and generating a time window analysis result;

[0027] S303: using the time window analysis result, comparing the identified abnormal vibration pattern with the normal vibration pattern, verifying the vibration behavior deviating from the normal pattern through pattern recognition technology, and generating a vibration abnormality pattern.

[0028] As a further solution of the present invention, the abnormal vibration pattern is used as input data, the Internet of Things is used to train the machine learning classifier, and the fault type in the vibration data is identified by setting the target parameters. The steps of obtaining the fault type identification result are specifically as follows:

[0029] S401: using the abnormal vibration pattern as input data, using the IoT device to collect and transmit data, initializing and setting parameters for the machine learning classifier, and training the classifier to obtain a training data set;

[0030] S402: using the training data set, executing a classifier training process, adjusting target parameters in classification, including a learning rate and a number of iterations, optimizing a response sensitivity to abnormal vibration modes, and generating parameter training results;

[0031] S403: applying the classifier to the current vibration data through the parameter training result, identifying the fault type by analyzing the data, and matching the prediction result with the known fault type to generate a fault type identification result.

[0032] As a further solution of the present invention, the fault type identification result is used to extract fault samples of oil monitoring data, and multiple stages of entropy value calculation are performed. By calculating the data entropy at differentiated time points, the entropy value data is recorded, and the operating speed and efficiency of the machine are evaluated. The steps of obtaining the entropy value data record are specifically as follows:

[0033] S501: based on the fault type identification result, extracting a target fault sample from the oil monitoring data corresponding to the fault type, extracting key data points from the target fault sample, and obtaining selected fault sample data;

[0034] S502: using the selected fault sample data, performing entropy calculation, evaluating the randomness and volatility of the data points, calculating the data entropy at the differentiated time points, and generating a volatility evaluation result;

[0035] S503: Analyze the overall data entropy trend according to the volatility assessment result, and evaluate the speed and efficiency changes of the mechanical operation by recording and analyzing the entropy value changes to obtain the entropy value data record.

[0036] As a further solution of the present invention, based on the entropy data record, the data is compared with the preset fault level standard, the health status of the machine is identified by threshold determination, and the processing priority of the mechanical failure risk is sorted. The steps of obtaining the fault risk assessment result are specifically as follows:

[0037] S601: Based on the entropy data record, a decision tree algorithm is used to compare and analyze multiple operating parameters and entropy values ​​of the machine, perform an initial determination of the fault level, and obtain a fault level determination result;

[0038] The formula of the decision tree algorithm is as follows:

[0039]

[0040] Among them, R is the fault level judgment result, x a is the actual entropy value of the ath operating parameter of the machine, t a is the preset fault entropy value threshold of the ath operating parameter, w a is the weight coefficient of the ath operating parameter, d a is the parameter deviation adjustment coefficient, λ is the adjustment parameter, N is the number of operating parameters, and e is a natural constant.

[0041] S602: using the fault level judgment result, performing threshold judgment, performing risk assessment on differentiated fault levels, sorting the severity of mechanical faults by analyzing risk coefficients and fault occurrence probabilities, and generating fault risk sorting results;

[0042] S603: Integrate and compare the fault risk data of different levels through the fault risk ranking results, evaluate the health status of the machinery through oil monitoring, identify and record the processing priority of the mechanical fault risk, and obtain the fault risk assessment result.

[0043] On the other hand, an oil monitoring fault diagnosis system based on the Internet of Things is provided, which is applied to an oil monitoring fault diagnosis method based on the Internet of Things, and the system includes:

[0044] The signal acquisition module collects mechanical vibration signals through oil monitoring, sets the acquisition time window, performs short-time Fourier transform on the acquired time domain signals, converts them into frequency domain signals, and obtains spectrum data;

[0045] The energy analysis module uses the spectrum data to analyze the energy spectrum density, performs energy accumulation on the data within the selected target frequency range, and identifies the fault indication frequency;

[0046] The abnormal pattern recognition module obtains the vibration abnormal pattern by analyzing the change of the differential time series based on the fault indication frequency;

[0047] The machine learning training module inputs the abnormal vibration pattern into the machine learning classifier, trains the classifier to identify the fault type in the differentiated vibration data, and generates a fault type identification result;

[0048] The entropy analysis module uses the fault type identification results to extract fault samples from the oil monitoring data, analyzes the data entropy at differentiated time points, records the entropy data, compares it with the set fault level standard, identifies the health status of the machinery, and obtains the fault risk assessment results.

[0049] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0050] The accuracy and real-time performance of fault diagnosis can be significantly improved by collecting and analyzing mechanical vibration signals. By performing a short-time Fourier transform, the time domain signal is effectively converted into a frequency domain signal, and the energy spectrum density analysis of the frequency domain signal is performed, so that the frequency characteristics associated with mechanical faults can be accurately identified. By using the Internet of Things technology, batch data can be processed in real time and classifiers can be trained through machine learning methods to further improve the accuracy of fault type identification. The introduction of entropy calculation allows the evaluation of the operating status and efficiency of the machinery, and the timely detection and sorting of the failure risks of the machinery by comparing with the set fault level standards. This combination of real-time monitoring and automatic identification greatly improves the ability of preventive maintenance and the operating efficiency of equipment, reduces the risk of unexpected downtime, and brings significant benefits in cost control and productivity improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0052] Figure 2 This is a detailed flow chart of S1 of the present invention;

[0053] Figure 3 This is a detailed flow chart of S2 of the present invention;

[0054] Figure 4 This is a detailed flow chart of S3 of the present invention;

[0055] Figure 5 This is a detailed flow chart of S4 of the present invention;

[0056] Figure 6 This is a detailed flow chart of S5 of the present invention;

[0057] Figure 7 This is a detailed flow chart of S6 of the present invention;

[0058] Figure 8 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0059] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0060] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0061] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0062] See also Figure 1 The embodiment of the present invention provides an oil monitoring fault diagnosis method based on the Internet of Things. The processing flow of the method may include the following steps:

[0063] S1: Collect mechanical vibration signals through oil monitoring, set the time window, perform short-time Fourier transform, convert time domain signals into frequency domain signals, and initialize and organize the spectrum data to obtain spectrum recognition data;

[0064] S2: Based on the spectrum recognition data, analyze the energy spectrum density, accumulate the energy of the data within the target frequency range, identify the frequency characteristics associated with the mechanical fault, and obtain the fault indication frequency;

[0065] S3: using the fault indication frequency, by calculating the degree of change in the differentiated time series, identifying the abnormal mode that deviates from the normal vibration mode, and obtaining the vibration abnormal mode;

[0066] S4: Taking the vibration abnormal pattern as input data, using the Internet of Things to train the machine learning classifier, identifying the fault type in the vibration data by setting the target parameters, and obtaining the fault type identification result;

[0067] S5: Using the fault type identification results, extract the fault samples of the oil monitoring data, perform multiple stages of entropy value calculation, calculate the data entropy at the differentiated time points, record the entropy value data, evaluate the running speed and efficiency of the machine, and obtain the entropy value data record;

[0068] S6: Based on the entropy data records, the data is compared with the pre-set fault level standards, the health status of the machinery is identified through threshold judgment, and the processing priority of the mechanical failure risk is sorted to obtain the fault risk assessment result.

[0069] The spectrum identification data includes the key frequencies, corresponding amplitudes and spectrum distribution of the vibration signal. The fault indication frequencies include the key fault frequencies, associated energy peaks and frequency widths. The abnormal vibration patterns include abnormal frequency patterns, degree of deviation and pattern stability. The fault type identification results include the identified fault type, fault severity and credibility score. The entropy data records include the calculated entropy values, entropy change rates and time series entropy value comparisons. The fault risk assessment results include the assessed fault level, risk ranking and expected maintenance time point.

[0070] See also Figure 2 , collect mechanical vibration signals through oil monitoring, set the time window, perform short-time Fourier transform, convert the time domain signal into the frequency domain signal, and obtain the spectrum recognition data in the following steps:

[0071] S101: Collect mechanical vibration signals through oil monitoring, set time window parameters, intercept the vibration signals in segments, check the signal quality by adjusting the signal interception, and adjust the amplitude of the intercepted signal by normalization processing. The execution process of generating segmented time domain signals is as follows;

[0072] In the monitoring of mechanical vibration signals, in order to ensure the signal quality, the signal is segmented by setting the time window parameters. The core of this process is to determine the optimal length of the time window, which needs to be adjusted according to the mechanical operation characteristics and vibration characteristics. For example, for mechanical equipment with rapid periodic changes, the time window should be shorter to capture detailed vibration information. On the contrary, for equipment with stable changes, a longer time window can be set to reduce the amount of data processing. Signal normalization processing includes adjusting the signal amplitude to a certain range to avoid data distortion caused by excessive or small amplitude in subsequent analysis. For example, the amplitude can be normalized to between -1 and 1 to ensure that data collected from different devices or at different times are comparable. The signal after amplitude adjustment is convenient for subsequent processing and analysis, ensuring data quality and accuracy of analysis, and generating segmented time domain signals.

[0073] S102: Based on the segmented time domain signal, the short-time Fourier transform technology is used to set the frequency resolution in the conversion process, and the frequency domain characteristics of the signal in the differentiated time window are calculated by dynamically adjusting the conversion window length to obtain the frequency domain signal data. The execution process is as follows;

[0074] The formula for the short-time Fourier transform technique is as follows:

[0075]

[0076] Among them, X′(ω) is the frequency domain signal data, ω is the angular frequency, t is the time offset, n is the discrete time index, x(n) is the signal amplitude at time point n, w(nt) is the window function with a width of w centered at t, α is the time window weight attenuation coefficient, β is the frequency window weight attenuation coefficient, and e is a natural constant.

[0077] The definition and calculation process of the parameters involved in the formula are as follows:

[0078] ω (angular frequency): This is the frequency used to represent the components in the signal in frequency domain analysis. It is determined by the sampling rate of the signal. If the sampling rate of the signal is set to 8000Hz, the angular frequency ω can range from 0 to 4000Hz.

[0079] t (time offset): represents the center point of the window function in the signal. For example, if the sampling frequency is 8000 Hz, a specific window function is selected with the center at the 1000th sample point of the signal.

[0080] n (discrete time index): represents the length range of the entire signal, from the beginning to the end of the signal. For example, for a signal with a length of 1 second, at a sampling rate of 8000 Hz, n ranges from 0 to 7999.

[0081] x(n) (signal amplitude): the actual amplitude of the signal at time point n, which is obtained by reading the actual collected signal data;

[0082] w(nt) (window function): Hanning window or Hamming window is used, and the window width w is set to 256 sample points, which means that each window covers a signal of 256 sample points;

[0083] α (time window weight decay coefficient): controls the decay speed of the edge of the window function, determined by experiments or optimization algorithms. For the Hanning window, the empirical value α is set to 0.46;

[0084] β (frequency window weight attenuation coefficient): adjusts the accuracy of frequency resolution. For audio signals, this value is obtained by minimizing the spectrum leakage in the overlapping area. The commonly used value is 0.001;

[0085] Calculation derivation example: Set t = 1000, ω = 2π × 1000 (corresponding to 1000 Hz), α = 0.46, β = 0.001, and use the Hanning window;

[0086] Calculate w(100-n) and x(n):

[0087]

[0088] calculate and

[0089]

[0090] Comprehensively calculate the partial values ​​of X′(ω):

[0091]

[0092] Because -372600 ≈0, the whole sum tends to 0, indicating that under extremely high attenuation conditions, the energy of the signal almost disappears. This result shows that the high attenuation coefficient causes the energy of the signal at this frequency to be completely suppressed, reflecting that when designing the window function and attenuation coefficient, it is necessary to reasonably select parameters to ensure the effectiveness and accuracy of signal analysis.

[0093] S103: Using the frequency domain signal data, performing spectrum recognition, filtering the frequency components through peak detection, combining the signal strength and frequency position, identifying the key spectrum features, and obtaining the spectrum recognition data. The execution process is as follows;

[0094] In the processing of frequency domain signal data, spectrum recognition is a key step. Peak detection is used to screen frequency components. This process is completed by identifying significant peaks in the spectrum. Each peak represents a major frequency component. For example, by setting a threshold, only frequency components with amplitudes exceeding the threshold are considered. Combined with signal strength and frequency position, key spectrum features can be identified, including the exact location of the frequency and the corresponding amplitude value. The data is crucial for subsequent mechanical condition monitoring, helping to identify potential faults or abnormalities and obtain spectrum recognition data.

[0095] See also Figure 3 Based on the spectrum recognition data, the energy spectrum density is analyzed, the energy of the data in the target frequency range is accumulated, the frequency characteristics associated with the mechanical fault are identified, and the steps of obtaining the fault indication frequency are as follows:

[0096] S201: Based on the spectrum recognition data, perform energy spectrum density analysis, sort out the energy of each frequency component, evaluate the energy contribution of each frequency point by signal strength, combine the frequency point energy, and obtain the cumulative energy spectrum. The execution process is as follows;

[0097] In the processing of spectrum data, it is necessary to evaluate the energy contribution of each frequency point. This requires obtaining the energy of each frequency component based on energy spectrum density analysis, calculating the energy value of each frequency point through mathematical conversion, and accumulating the energy of all frequency points. This processing involves the organization of spectrum data and energy calculation, which ensures the accuracy and feasibility of the analysis and provides basic data for subsequent steps. This analysis not only helps to understand the energy distribution of the signal at different frequencies, but also provides key information for subsequent fault analysis to obtain the accumulated energy spectrum.

[0098] S202: By accumulating the energy spectrum, defining the target frequency range, iteratively gathering and analyzing the accumulated energy data within the target range, identifying the frequency region with energy anomalies by setting a threshold, and generating the target frequency analysis result. The execution process is as follows;

[0099] By setting the threshold to identify the frequency area of ​​energy anomaly, according to formula E total =∑E z Calculate the accumulated energy in the entire target frequency range. Where, E total Represents the total accumulated energy within the target frequency range, E z Represents the energy of a single frequency point z. Determine the energy E of each frequency point z z , can be directly obtained through frequency analysis equipment or calculated through spectrum data processing algorithms. For example, if five frequency points are set, and the energy values ​​are 10, 20, 30, 40, and 50 respectively, the total cumulative energy E total =10+20+30+40+50=150. The key is to perform a simple arithmetic accumulation of the energy of all target frequency points to obtain the total energy. The total value is the basis for evaluating whether the energy in the frequency area is abnormal or not.

[0100] S203: Using the target frequency analysis result, the identified abnormal frequency characteristics are compared with the known mechanical fault frequency characteristics, and the frequency characteristics associated with the mechanical fault are identified through matching analysis, and the execution process of generating the fault indication frequency is as follows;

[0101] When identifying frequency characteristics associated with mechanical failures, the target frequency analysis results are compared with the known mechanical failure frequency characteristics. The comparison process includes a detailed analysis of each abnormal frequency characteristic to ensure that the frequency characteristics are accurately matched. This comparison is processed by algorithms, such as using least squares or statistical matching techniques to identify frequency characteristics associated with mechanical failures. This process ensures the accuracy of the analysis and the reliability of the operation. Comparative analysis is not only a key step in identifying faults, but also a necessary measure to ensure the safety of equipment operation and generate fault indication frequencies.

[0102] See also Figure 4 , using the fault indication frequency, by calculating the degree of change in the differentiated time series, identifying the abnormal mode that deviates from the normal vibration mode, the steps for obtaining the abnormal vibration mode are as follows:

[0103] S301: Based on the fault indication frequency, identify the frequency difference between the current vibration data of the mechanical equipment and the standard vibration mode, analyze the degree of change of each fault indication frequency through the differential time series, and obtain the execution process of the frequency difference time series as follows;

[0104] Analyze the frequency difference between the current vibration data of the mechanical equipment and the standard vibration mode. The process involves differential time series analysis. By evaluating the degree of change of each fault indication frequency, this analysis process ensures accurate monitoring and analysis of vibration data, helps to understand the vibration characteristics of mechanical equipment under different working conditions, and provides data support for subsequent fault prevention and maintenance. By recording and accumulating each measured difference between the fault frequency and the standard frequency, it is used for further analysis and fault diagnosis to generate a frequency difference time series.

[0105] S302: Using the frequency difference time series, perform time window analysis on the data in the sequence, and distinguish between normal and abnormal vibration modes by calculating the degree of deviation of the vibration mode in each time window. The execution process of generating the time window analysis result is as follows;

[0106] By calculating the degree of vibration mode deviation in each time window, according to the formula Calculate the deviation. In the formula, ΔP represents the degree of deviation of the vibration mode, P obs represents the vibration mode observed in each time window, P norm Represents the normal vibration mode. For each vibration data P observed in the time window obs And standard vibration data P norm Perform difference calculation. For example, if the vibration data observed in a certain time window is 15, 20, 22, and the normal data is 10, 15, 20, then the sum of squares of the differences is calculated as (15-10) 2 +(20-15) 2 +(22-20) 2 =25+25+4=54, then take the square root to get the degree of deviation The key to the calculation step is to quantify the degree of vibration deviation through mathematical methods to provide a numerical basis for identifying abnormal vibration.

[0107] S303: using the time window analysis result, comparing the identified abnormal vibration pattern with the normal vibration pattern, verifying the vibration behavior deviating from the normal pattern through pattern recognition technology, and generating the abnormal vibration pattern. The execution process is as follows;

[0108] The identified abnormal vibration pattern is compared with the normal vibration pattern. This process is achieved through pattern recognition technology. During the comparative analysis, by verifying the vibration behavior that deviates from the normal pattern, this analysis process not only helps to quickly identify potential equipment failures, but also effectively guides maintenance personnel to perform targeted maintenance. The identification of abnormal vibration patterns is achieved by calculating and comparing the deviations between the vibration data in each time window and the standard vibration pattern, ensuring the accuracy and reliability of vibration analysis, providing important technical support for the health status monitoring of mechanical equipment, and generating abnormal vibration patterns.

[0109] See also Figure 5 , taking the vibration abnormal pattern as input data, using the Internet of Things, training the machine learning classifier, and identifying the fault type in the vibration data by setting the target parameters. The specific steps to obtain the fault type identification result are as follows:

[0110] S401: Based on the abnormal vibration pattern as input data, the IoT device is used to collect and transmit data, the machine learning classifier is initialized and parameterized, and the classifier is trained to obtain the training data set. The execution process is as follows;

[0111] Initialize and set parameters for the machine learning classifier according to the formula Calculate the loss function. In the formula, L(θ) represents the loss function, θ represents the classifier parameters, and N i represents the total number of samples, y i represents the true label of the i-th sample, Represents the predicted label. Initialize the classifier parameter θ. For example, if the initial value of the parameter is 0, the sample data is {0, 1, 0, 1}, and the initial value of the predicted label is 0.5, substitute it into the formula, and the loss function is calculated as The key to this calculation step is to adjust the parameters by optimizing the loss function to improve the accuracy and robustness of the classifier.

[0113] S402: Using the training data set, executing the classifier training process, adjusting the target parameters in the classification, including the learning rate and the number of iterations, optimizing the response sensitivity to the abnormal vibration mode, and generating the parameter training results. The execution process is as follows;

[0114] In the process of executing classifier training, the key is to adjust the target parameters in the classification, including the learning rate and the number of iterations. The optimization of parameters can significantly improve the response sensitivity to abnormal vibration modes. Through mathematical optimization techniques, such as gradient descent, the parameters are adjusted to minimize the classification error. Through performance evaluation in the iterative process, it is ensured that each step of parameter adjustment is based on the learning results of the previous step. The optimization process includes multiple iterations, and each iteration evaluates the classifier performance and adjusts the parameters. The training process is accurately controlled to obtain the best classifier settings and generate parameter training results.

[0115] S403: Based on the parameter training results, the classifier is applied to the current vibration data, the fault type is identified by analyzing the data, and the prediction result is matched with the known fault type. The execution process of generating the fault type identification result is as follows;

[0116] The classifier is applied to the current vibration data. By analyzing the data, the specific fault type can be identified. In this process, the application of the classifier is based on the previously trained parameters and models, and the prediction results are matched with the known fault types. This matching process is achieved through algorithm comparison. This process not only confirms the effectiveness of the classifier, but also provides decision-making information for fault handling. The classifier responds and analyzes new vibration data quickly to ensure the accuracy and efficiency of fault identification and generate fault type identification results.

[0117] See also Figure 6 , using the fault type identification results, extract the fault samples of the oil monitoring data, perform multiple stages of entropy calculation, calculate the data entropy at differentiated time points, record the entropy data, evaluate the running speed and efficiency of the machine, and obtain the steps of entropy data recording as follows:

[0118] S501: Based on the fault type identification result, extract the target fault sample in the oil monitoring data corresponding to the fault type, extract the key data points in the target fault sample, and obtain the selected fault sample data. The execution process is as follows;

[0119] Extract target fault samples from the oil monitoring data corresponding to the fault type. The key to this process is to accurately identify and extract specific sample data related to the identified fault type from the batch monitoring data. Extracting key data points from the target fault samples is to deeply analyze the specific manifestations and causes of the fault, which involves data screening, feature calibration and recording of key data points. The data points represent the key monitoring indicators when the fault occurs, which provides a basis for subsequent data analysis and fault diagnosis, ensures the pertinence and accuracy of the analysis, and obtains the selected fault sample data.

[0120] S502: Using the selected fault sample data, executing entropy value calculation, evaluating the randomness and volatility of the data points, calculating the data entropy at the differentiated time points, and generating the volatility evaluation result. The execution process is as follows;

[0121] By evaluating the randomness and volatility of the data points, the data entropy is calculated according to the formula H = -∑p(x)logp(x). In the formula, H represents the data entropy and p(x) represents the probability of the data point x appearing. Determine the probability p(x) of each data point appearing, set the data point distribution to {10, 10, 20, 20, 20, 30}, then the probability of each point is Substituting into the formula, the data entropy is calculated as The key to this calculation step is to express the uncertainty and complexity of the data through entropy value, providing a numerical basis for volatility assessment.

[0122] S503: Analyze the overall data entropy trend according to the volatility assessment result, and evaluate the speed and efficiency changes of the machine operation by recording and analyzing the entropy value changes. The execution process of obtaining the entropy value data record is as follows;

[0123] Analyze the overall data entropy trend. This process is achieved through continuous recording and analysis of entropy changes in the data. The steps to evaluate the speed and efficiency changes of mechanical operation include entropy calculation, trend drawing and change interpretation. It not only reflects the stability and efficiency of the mechanical operation status, but also provides a basis for predicting the future operation trend of the machinery. The analysis of entropy helps the maintenance team identify potential operating problems, ensures the timeliness and effectiveness of mechanical maintenance, and generates entropy data records.

[0124] See also Figure 7 Based on the entropy data records, the data is compared with the pre-set fault level standards, the health status of the machine is identified through threshold judgment, and the priority of mechanical failure risk treatment is sorted. The specific steps to obtain the fault risk assessment results are as follows:

[0125] S601: Based on the entropy data record, a decision tree algorithm is used to compare and analyze multiple operating parameters and entropy values ​​of the machine, and an initial judgment of the fault level is performed. The execution process of obtaining the fault level judgment result is as follows;

[0126] The formula for the decision tree algorithm is as follows:

[0127]

[0128] Among them, R is the fault level judgment result, x a is the actual entropy value of the ath operating parameter of the machine, t a is the preset fault entropy value threshold of the ath operating parameter, w ais the weight coefficient of the ath operating parameter, d a is the parameter deviation adjustment coefficient, λ is the adjustment parameter, N is the number of operating parameters, and e is a natural constant;

[0129] x a : Actual entropy value, obtained in real time through the mechanical operation monitoring system. The machine is set with three parameters, and the entropy values ​​obtained by monitoring are x 1 =0.2,x 2 =0.3,x 3 =0.1;

[0130] t a :Preset fault entropy value threshold, set according to mechanical historical data and fault records, the set value is as follows: 1 =0.5, t 2 =0.3, t 3 =0.2;

[0131] w a : Weight coefficient, set according to the degree of influence of the parameter, here set w 1 =0.5, w 2 =0.3, w 3 =0.2;

[0132] d a : Parameter deviation adjustment coefficient, obtained through mechanical failure analysis data, set to d 1 =-1.0, d 2 =0,d 3 =1.0;

[0133] N: the number of parameters, 3 in this example;

[0134] λ: adjustment parameter, used to adjust the denominator when the number of parameters is large, set to 0.5;

[0135] Calculate the adjusted value for each parameter

[0136] For a=1,

[0137] For a=2,

[0138] For a=3,

[0139] Multiply the adjustment value by the weight factor and sum:

[0140]

[0141] Calculation results:

[0142] The result shows that the calculated fault level judgment result value is 0.029. When judging the fault level, a lower result indicates that the current mechanical state is close to the normal operating state, reflecting that the calculation is increased in detail and accuracy by adding the adjustment coefficient and weight adjustment formula, making the fault level judgment more in line with the actual operating conditions.

[0143] S602: Using the fault level judgment result, threshold judgment is performed to conduct risk assessment on differentiated fault levels, and the severity of mechanical faults is ranked by analyzing the risk coefficient and the probability of fault occurrence. The execution process of generating the fault risk ranking result is as follows;

[0144] The threshold determination process is performed. In this process, the risk coefficient and the probability of failure are analyzed by risk assessment of differentiated fault levels. The key is to sort the severity of mechanical failures. This analysis relies on mathematical models and statistical methods, such as calculating the cumulative distribution function or probability density function of each fault level, determining the risk coefficient of the fault, and sorting according to the coefficient to ensure that the assessment of fault risk is based on the accurate calculation of actual data and algorithms, providing scientific decision-making information for maintenance and prevention, and generating fault risk sorting results.

[0145] S603: Through the fault risk ranking results, the differentiated level fault risk data are integrated and compared, the health status of the machinery is evaluated through oil monitoring, and the processing priority of the mechanical fault risk is identified and recorded. The execution process of obtaining the fault risk assessment result is as follows;

[0146] Through oil monitoring, the health status of the machinery is evaluated, and the priority of handling mechanical failure risks is identified and recorded. This process not only involves data integration and comparison, but also includes monitoring and analysis of specific indicators in the mechanical oil. Through this comprehensive evaluation, the types of failures that need to be handled as a priority can be effectively identified to ensure the efficiency of maintenance work and the stable operation of mechanical equipment, and obtain failure risk assessment results.

[0147] See also Figure 8 On the other hand, an oil monitoring fault diagnosis system based on the Internet of Things is provided, and the system is applied to an oil monitoring fault diagnosis method based on the Internet of Things, and the system includes:

[0148] The signal acquisition module collects mechanical vibration signals through oil monitoring, sets the acquisition time window, performs short-time Fourier transform on the acquired time domain signals, converts them into frequency domain signals, and obtains spectrum data;

[0149] The energy analysis module uses spectrum data to analyze energy spectrum density, performs energy accumulation on data within the selected target frequency range, and identifies fault indication frequencies;

[0150] The abnormal pattern recognition module obtains the vibration abnormal pattern by analyzing the changes of the differential time series based on the fault indication frequency;

[0151] The machine learning training module inputs the vibration abnormality pattern into the machine learning classifier, trains the classifier to identify the fault type in the differentiated vibration data, and generates a fault type identification result;

[0152] The entropy analysis module uses the fault type identification results to extract fault samples from the oil monitoring data. By analyzing the data entropy at differentiated time points, the entropy data is recorded and compared with the set fault level standards to identify the health status of the machinery and obtain the fault risk assessment results.

[0153] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An oil monitoring fault diagnosis method based on the Internet of Things, characterized in that: The method comprises: Collect mechanical vibration signals through oil monitoring, set the time window, perform short-time Fourier transform, convert time domain signals into frequency domain signals, and obtain spectrum recognition data; Based on the spectrum identification data, analyzing the energy spectrum density, accumulating the energy of the data within the target frequency range, identifying the frequency characteristics associated with the mechanical fault, and obtaining the fault indication frequency; Using the fault indication frequency, by calculating the degree of change in the differentiated time series, identifying an abnormal mode that deviates from a normal vibration mode, and obtaining an abnormal vibration mode; The abnormal vibration pattern is used as input data, and the machine learning classifier is trained by using the Internet of Things. The fault type in the vibration data is identified by setting target parameters to obtain a fault type identification result; Utilizing the fault type identification result, extracting fault samples of the oil monitoring data, executing multiple stages of entropy value calculation, calculating data entropy at differentiated time points, recording entropy value data, evaluating the operating speed and efficiency of the machine, and obtaining entropy value data records; Based on the entropy data record, it is compared with the pre-set fault level standard, the health status of the machine is identified through threshold judgment, and the processing priority of the mechanical failure risk is sorted to obtain the fault risk assessment result.

2. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: The spectrum identification data includes the key frequency, corresponding amplitude and spectrum distribution of the vibration signal, the fault indication frequency includes the key fault frequency, associated energy peak and frequency width, the vibration abnormal mode includes the abnormal frequency mode, deviation degree and mode stability, the fault type identification result includes the identified fault type, fault severity and credibility score, the entropy value data record includes the calculated entropy value, entropy value change rate and time series entropy value comparison, and the fault risk assessment result includes the assessed fault level, risk ranking and expected maintenance time point.

3. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: The steps of collecting mechanical vibration signals through oil monitoring, setting the time window, performing short-time Fourier transform, converting the time domain signal into the frequency domain signal, and obtaining spectrum recognition data are as follows: Collect mechanical vibration signals through oil monitoring, set time window parameters, intercept vibration signals in segments, check signal quality by adjusting signal interception, and adjust the amplitude of intercepted signals through normalization processing to generate segmented time domain signals; Based on the segmented time domain signal, the short-time Fourier transform technology is used to set the frequency resolution in the conversion process, and the frequency domain characteristics of the signal in the differentiated time window are calculated by dynamically adjusting the conversion window length to obtain the frequency domain signal data; The frequency domain signal data is used to perform spectrum recognition, filter frequency components through peak detection, and identify key spectrum features in combination with signal strength and frequency position to obtain spectrum recognition data.

4. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 3 is characterized in that: The formula for the short-time Fourier transform technique is as follows: Among them, X′(ω) is the frequency domain signal data, ω is the angular frequency, t is the time offset, n is the discrete time index, x(n) is the signal amplitude at time point n, w(nt) is the window function with a width of w centered at t, α is the time window weight attenuation coefficient, β is the frequency window weight attenuation coefficient, and e is a natural constant.

5. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: Based on the spectrum identification data, the energy spectrum density is analyzed, the energy of the data within the target frequency range is accumulated, the frequency characteristics associated with the mechanical fault are identified, and the steps of obtaining the fault indication frequency are specifically as follows: Based on the spectrum identification data, perform energy spectrum density analysis, sort out the energy of each frequency component, evaluate the energy contribution of each frequency point by signal strength, combine the frequency point energies, and obtain a cumulative energy spectrum; By using the accumulated energy spectrum, a target frequency range is defined, the accumulated energy data within the target range is iteratively aggregated and analyzed, a frequency region with energy anomalies is identified by setting a threshold, and a target frequency analysis result is generated; The target frequency analysis result is used to compare the identified abnormal frequency characteristics with the known mechanical fault frequency characteristics, and the frequency characteristics associated with the mechanical fault are identified through matching analysis to generate a fault indication frequency.

6. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: The steps of using the fault indication frequency to identify the abnormal mode that deviates from the normal vibration mode by calculating the degree of change in the differentiated time series, and obtaining the abnormal vibration mode are specifically as follows: Based on the fault indication frequency, identifying the frequency difference between the current vibration data of the mechanical equipment and the standard vibration mode, analyzing the degree of change of each fault indication frequency through a differential time series, and obtaining a frequency difference time series; Using the frequency difference time series, performing time window analysis on the data in the series, distinguishing normal and abnormal vibration modes by calculating the degree of deviation of the vibration mode in each time window, and generating a time window analysis result; The time window analysis result is used to compare the identified abnormal vibration pattern with the normal vibration pattern, and the vibration behavior deviating from the normal pattern is verified by pattern recognition technology to generate an abnormal vibration pattern.

7. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: The abnormal vibration pattern is used as input data, and the machine learning classifier is trained using the Internet of Things. The fault type in the vibration data is identified by setting target parameters. The specific steps of obtaining the fault type identification result are as follows: Using the abnormal vibration pattern as input data, using the IoT device to collect and transmit data, initializing and setting parameters for the machine learning classifier, and training the classifier to obtain a training data set; Using the training data set, executing a classifier training process, adjusting target parameters in the classification, including a learning rate and a number of iterations, optimizing the response sensitivity to abnormal vibration modes, and generating parameter training results; The classifier is applied to the current vibration data through the parameter training results, the fault type is identified by analyzing the data, and the prediction result is matched with the known fault type to generate a fault type identification result.

8. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: By using the fault type identification result, extracting fault samples of oil monitoring data, executing multiple stages of entropy value calculation, calculating data entropy at differentiated time points, recording entropy value data, evaluating the running speed and efficiency of the machine, and obtaining the steps of entropy value data recording are specifically as follows: Based on the fault type identification result, extracting a target fault sample from the oil monitoring data corresponding to the fault type, extracting key data points from the target fault sample, and obtaining selected fault sample data; The selected fault sample data is used to perform entropy value calculation, and the data entropy at the differentiated time points is calculated by evaluating the randomness and volatility of the data points to generate a volatility evaluation result; According to the volatility assessment results, the overall data entropy trend is analyzed, and by recording and analyzing the entropy value changes, the speed and efficiency changes of the mechanical operation are evaluated to obtain the entropy value data record.

9. The oil monitoring fault diagnosis method based on the Internet of Things according to claim 1 is characterized in that: Based on the entropy data record, the data is compared with the preset fault level standard, the health status of the machine is identified through threshold judgment, and the processing priority of the mechanical fault risk is sorted. The steps of obtaining the fault risk assessment result are specifically as follows: Based on the entropy value data record, a decision tree algorithm is used to compare and analyze multiple operating parameters and entropy values ​​of the machine, perform an initial determination of the fault level, and obtain a fault level determination result; The formula of the decision tree algorithm is as follows: Among them, R is the fault level judgment result, x a is the actual entropy value of the ath operating parameter of the machine, t a is the preset fault entropy value threshold of the ath operating parameter, w a is the weight coefficient of the ath operating parameter, d a is the parameter deviation adjustment coefficient, λ is the adjustment parameter, N is the number of operating parameters, and e is a natural constant. Using the fault level judgment result, threshold judgment is performed to conduct risk assessment on differentiated fault levels, and the severity of mechanical faults is ranked by analyzing risk coefficients and fault occurrence probabilities to generate fault risk ranking results; By using the fault risk ranking results, the differentiated level fault risk data are integrated and compared, the health status of the machinery is evaluated through oil monitoring, the handling priority of the mechanical failure risk is identified and recorded, and the fault risk assessment result is obtained.

10. An oil monitoring fault diagnosis system based on the Internet of Things, characterized in that: According to the oil monitoring fault diagnosis method based on the Internet of Things according to any one of claims 1 to 9, the system comprises: The signal acquisition module collects mechanical vibration signals through oil monitoring, sets the acquisition time window, performs short-time Fourier transform on the acquired time domain signals, converts them into frequency domain signals, and obtains spectrum data; The energy analysis module uses the spectrum data to analyze the energy spectrum density, performs energy accumulation on the data within the selected target frequency range, and identifies the fault indication frequency; The abnormal pattern recognition module obtains the vibration abnormal pattern by analyzing the change of the differential time series based on the fault indication frequency; The machine learning training module inputs the abnormal vibration pattern into the machine learning classifier, trains the classifier to identify the fault type in the differentiated vibration data, and generates a fault type identification result; The entropy analysis module uses the fault type identification results to extract fault samples from the oil monitoring data, analyzes the data entropy at differentiated time points, records the entropy data, compares it with the set fault level standard, identifies the health status of the machinery, and obtains the fault risk assessment results.

Images