Industrial equipment health diagnosis method and system oriented to industrial internet

Abstract

The invention is suitable for the field of industrial equipment health diagnosis, and provides an industrial internet-oriented industrial equipment health diagnosis method and system, and the method comprises the steps: collecting industrial equipment operation parameters monitored by a plurality of types of sensors; acquiring temperature and humidity associated data in the operation parameters of the industrial equipment, and calculating and generating a temperature and humidity associated feedback value; acquiring power voltage associated data in the operation parameters of the industrial equipment, and calculating and generating a power voltage associated feedback value; lubricating oil associated data in the operation parameters of the industrial equipment are obtained, and a lubricating oil associated feedback value is calculated and generated; generating a health state feature vector; obtaining industrial equipment operation parameter historical data and health state feature vectors, and generating a health diagnosis model training set; constructing a health diagnosis prediction model; according to the method and the system, the risk of unplanned shutdown is reduced, the limitation of traditional single-parameter monitoring is broken through, and full-life-cycle management of the equipment is realized.

Description

Technical Field

[0001] The present invention belongs to the field of industrial equipment health diagnosis, and particularly relates to an industrial equipment health diagnosis method and system for industrial Internet. Background Art

[0002] At present, with the booming development of the industrial Internet, the efficient and stable operation of industrial equipment is crucial for the normal operation of the entire industrial production system. The health status of industrial equipment directly affects production efficiency, product quality, and the economic benefits of enterprises. Therefore, accurately and timely diagnosing the health status of industrial equipment has become a research hotspot in the industrial field.

[0003] Traditional industrial equipment health diagnosis methods mainly rely on monitoring and analyzing the operating parameters of the equipment. However, these methods often only focus on single parameter indicators, ignoring the mutual correlation and coupling effects between parameters, and it is difficult to comprehensively and accurately reflect the true health status of the equipment. In view of the above problems, it is urgent to develop a more mature industrial equipment health diagnosis method and system for industrial Internet. Summary of the Invention

[0004] The purpose of the present invention is to provide an industrial equipment health diagnosis method and system for industrial Internet, aiming to solve the problems raised in the above background art.

[0005] The present invention is implemented as follows. On the one hand, an industrial equipment health diagnosis method for industrial Internet, the method includes:

[0006] Collect the operating parameters of industrial equipment monitored by several types of sensors; the several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors;

[0007] Obtain the temperature and humidity correlation data in the operating parameters of industrial equipment, calculate and generate a temperature and humidity correlation feedback value;

[0008] Obtain the power voltage correlation data in the operating parameters of industrial equipment, calculate and generate a power voltage correlation feedback value;

[0009] Obtain the lubricating oil correlation data in the operating parameters of industrial equipment, calculate and generate a lubricating oil correlation feedback value;

[0010] Generate a health status feature vector based on the humidity correlation feedback value, power voltage correlation feedback value, and lubricating oil correlation feedback value;

[0011] Obtain the historical data of the operating parameters of industrial equipment, the health status feature vector, set the equipment health status grading, and generate a health diagnosis model training set based on the historical data of the operating parameters of industrial equipment, the health status feature vector, and the equipment health status grading, and construct a health diagnosis prediction model;

[0012] Collect and calculate the real-time health status feature vector, input the real-time health status feature vector into the health diagnosis and prediction model, and output the predicted health status grading of the device.

[0013] As a further solution of the present invention, the obtaining of the temperature and humidity correlation data in the operating parameters of the industrial device and the calculation and generation of the temperature and humidity correlation feedback value specifically include:

[0014] Extract and generate the temperature normalization value in the operating parameters of the industrial device and the humidity normalization value ;

[0015] Calculate the temperature normalization value and the humidity normalization value of the temperature and humidity correlation coefficient ;

[0016] The calculation process of the temperature and humidity correlation coefficient is:

[0017] ;

[0018] In the formula, is the time value, is the number of data samples, is the average value of the temperature normalization value, is the average value of the humidity normalization value;

[0019] Based on the temperature and humidity correlation coefficient , calculate and generate the temperature and humidity coupling coefficient ;

[0020] The calculation process of the temperature and humidity coupling coefficient is:

[0021] ;

[0022] In the formula, and are weight coefficients, and , is the time weight factor, is the time interval.

[0023] As a further solution of the present invention, the obtaining of the power voltage correlation data in the operating parameters of the industrial device and the calculation and generation of the power voltage correlation feedback value specifically include:

[0024] Extract the current data and the voltage data ;

[0025] The current data and voltage data Based on Fourier transform, generate the current components of each harmonic and voltage components ;

[0026] Calculate the total harmonic distortion rate of current and the total harmonic distortion rate of current ;

[0027] The calculation process of the total harmonic distortion rate of current is as follows:

[0028] ;

[0029] The calculation process of the total harmonic distortion rate of current is as follows:

[0030] ;

[0031] In the formula, is the harmonic order, is the fundamental current, is the fundamental voltage;

[0032] Based on the total harmonic distortion rate of current and the total harmonic distortion rate of current , respectively match the corresponding weight coefficients and ;

[0033] Calculate the harmonic pollution degree index ;

[0034] The calculation process of the harmonic pollution degree index is as follows:

[0035] ;

[0036] In the formula, and are weight coefficients, and .

[0037] As a further solution of the present invention, the obtaining of the lubricating oil related data in the operating parameters of the industrial equipment and the calculation and generation of the lubricating oil related feedback value specifically include:

[0038] Extract the dielectric constant of the lubricating oil in the operating parameters of the industrial equipment;

[0039] Calculate and generate the mean value and standard deviation of the dielectric constant of the lubricating oil;

[0040] Calculate and generate the deviation degree of the dielectric constant of the lubricating oil ;

[0041] The calculation process of the deviation degree of the dielectric constant of the lubricating oil is as follows:

[0042] ;

[0043] Based on a preset deviation threshold calculate and generate a quantization index of the dielectric constant of the lubricating oil ;

[0044] The calculation process of the quantization index of the dielectric constant of the lubricating oil is as follows:

[0045] .

[0046] As a further solution of the present invention, the generation of the health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value specifically includes:

[0047] Calculate the change rate feature of the temperature-humidity coupling coefficient , the fluctuation range feature ;

[0048] Calculate the change trend feature of the harmonic pollution degree index , the change trend feature of the content of each harmonic, and the harmonic pollution mutation feature ;

[0049] Calculate the change frequency feature of the quantization index of the lubricating oil dielectric constant ;

[0050] Combine various features to generate a targeted feature vector .

[0051] As a further solution of the present invention, the acquisition of the historical data of the operating parameters of the industrial equipment, the health status feature vector, the setting of the equipment health status classification, and the generation of the health diagnosis model training set and the construction of the health diagnosis prediction model based on the historical data of the operating parameters of the industrial equipment, the health status feature vector, and the equipment health status classification specifically include:

[0052] Set the health status classification, and the health status classification includes: healthy level, sub-healthy level, fault critical warning level, fault level;

[0053] Collect the historical data of the operating parameters of the industrial equipment and based on the targeted feature vector Generate a training set for the health diagnosis model based on the collection and calculation of real-time health status feature vectors and health status grading;

[0054] Train the health diagnosis model training set through a support vector machine and a convolutional neural network, and optimize it using the cross-validation method to generate a health diagnosis prediction model .

[0055] As a further aspect of the present invention, the collection and calculation of real-time health status feature vectors, inputting the real-time health status feature vectors into the health diagnosis prediction model, and outputting the predicted health status grading of the device specifically include:

[0056] Input the extracted real-time feature vectors into the health diagnosis prediction model to output the current health status level of the industrial device ;

[0057] Based on the current health status level of the industrial device , match and send corresponding warning action information.

[0058] As a further aspect of the present invention, on the other hand, an industrial device health diagnosis system for the industrial Internet, the system includes:

[0059] A first acquisition module for acquiring the operating parameters of industrial devices monitored by several types of sensors;

[0060] The several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors;

[0061] A first acquisition module for acquiring the temperature and humidity correlation data in the operating parameters of industrial devices;

[0062] A first calculation and generation module for calculating and generating a temperature and humidity correlation feedback value;

[0063] A second acquisition module for acquiring the power voltage correlation data in the operating parameters of industrial devices;

[0064] A second calculation and generation module for calculating and generating a power voltage correlation feedback value;

[0065] A third acquisition module for acquiring the lubricating oil correlation data in the operating parameters of industrial devices;

[0066] A third calculation and generation module for calculating and generating a lubricating oil correlation feedback value;

[0067] A generation module for generating a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value;

[0068] A prediction model module, configured to obtain historical data of industrial equipment operation parameters, health status feature vectors, set the classification of equipment health status, and generate a training set for a health diagnosis model and construct a health diagnosis prediction model based on the historical data of industrial equipment operation parameters, health status feature vectors, and the classification of equipment health status;

[0069] An input / output module, configured to collect and calculate real-time health status feature vectors, input the real-time health status feature vectors into the health diagnosis prediction model, and output the predicted health status classification of the equipment.

[0070] As a further solution of the present invention, the first calculation and generation module specifically includes:

[0071] An extraction and generation unit, configured to extract and generate the temperature normalization value in the industrial equipment operation parameters and the humidity normalization value ;

[0072] A first calculation unit, configured to calculate the temperature-humidity correlation coefficient of the temperature normalization value and the humidity normalization value ; ;

[0073] A calculation and generation unit, configured to calculate and generate a temperature-humidity coupling coefficient based on the temperature-humidity correlation coefficient ; .

[0074] As a further solution of the present invention, the second calculation and generation module specifically includes:

[0075] An extraction unit, configured to extract the current data and the voltage data in the industrial equipment operation parameters;

[0076] A generation unit, configured to generate the current components of each harmonic and the voltage components of the current data and the voltage data based on Fourier transform; ; ;

[0077] A second calculation unit, configured to calculate the total harmonic distortion rate of the current and the total harmonic distortion rate of the current ;

[0078] A matching unit, configured to respectively match the corresponding weight coefficients and based on the total harmonic distortion rate of the current and ;

[0079] A third computing unit for calculating the harmonic pollution degree index .

[0080] An industrial equipment health diagnosis method and system for industrial Internet provided by the present invention. Through multi-sensor data fusion analysis, the method and system convert the implicit associations of multi-modalities into quantifiable health feature vectors, and construct an equipment health assessment system coupling multiple physical fields. The intelligent diagnosis model based on dynamic feature vectors realizes fault trend prediction, and combines with the closed-loop optimization mechanism to continuously improve the diagnosis accuracy, so as to support predictive maintenance decision-making, reduce the risk of unplanned downtime, not only break through the limitations of traditional single-parameter monitoring, but also provide technical support for industrial equipment and realize the full life cycle management of equipment. Description of the Drawings

[0081] Figure 1 is the main flowchart of an industrial equipment health diagnosis method for industrial Internet.

[0082] Figure 2 is the flowchart of obtaining the temperature and humidity correlation data in the operating parameters of industrial equipment, calculating and generating the temperature and humidity correlation feedback value in an industrial equipment health diagnosis method for industrial Internet.

[0083] Figure 3 is the flowchart of obtaining the power voltage correlation data in the operating parameters of industrial equipment, calculating and generating the power voltage correlation feedback value in an industrial equipment health diagnosis method for industrial Internet.

[0084] Figure 4 is the flowchart of obtaining the lubricating oil correlation data in the operating parameters of industrial equipment, calculating and generating the lubricating oil correlation feedback value in an industrial equipment health diagnosis method for industrial Internet.

[0085] Figure 5 is the flowchart of generating a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value and the lubricating oil correlation feedback value in an industrial equipment health diagnosis method for industrial Internet.

[0086] Figure 6 is the flowchart of obtaining the historical data of the operating parameters of industrial equipment, the health status feature vector, setting the equipment health status grading, generating a health diagnosis model training set based on the historical data of the operating parameters of industrial equipment, the health status feature vector and the equipment health status grading, and constructing a health diagnosis prediction model in an industrial equipment health diagnosis method for industrial Internet.

[0087] Figure 7It is a flow chart for collecting and calculating real-time health status feature vectors in an industrial equipment health diagnosis method for the industrial Internet, inputting the real-time health status feature vectors into a health diagnosis prediction model, and outputting the predicted health status grading of the equipment.

[0088] Figure 8 It is the main structure diagram of an industrial equipment health diagnosis system for the industrial Internet.

[0089] Figure 9 It is the structural block diagram of the first calculation and generation module in an industrial equipment health diagnosis system for the industrial Internet.

[0090] Figure 10 It is the structural block diagram of the second calculation and generation module in an industrial equipment health diagnosis system for the industrial Internet. Specific implementation mode

[0091] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0092] The following describes the specific implementation of the present invention in detail with reference to specific embodiments.

[0093] An industrial equipment health diagnosis method and system for the industrial Internet provided by the present invention solve the technical problems in the background art.

[0094] As Figure 1 shown, it is the main flow chart of an industrial equipment health diagnosis method for the industrial Internet provided by an embodiment of the present invention. The industrial equipment health diagnosis method for the industrial Internet includes:

[0095] Step S100: Collect the operation parameters of industrial equipment monitored by several types of sensors;

[0096] The several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors;

[0097] Step S200: Obtain the temperature and humidity correlation data in the operation parameters of industrial equipment, calculate and generate a temperature and humidity correlation feedback value;

[0098] Step S300: Obtain the power voltage correlation data in the operation parameters of industrial equipment, calculate and generate a power voltage correlation feedback value;

[0099] Step S400: Obtain the lubricating oil correlation data in the operation parameters of industrial equipment, calculate and generate a lubricating oil correlation feedback value;

[0100] Step S500: Generate a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value;

[0101] Step S600: Obtain the historical data of the industrial equipment operation parameters and the health status feature vector, set the equipment health status grading, and generate a health diagnosis model training set based on the historical data of the industrial equipment operation parameters, the health status feature vector, and the equipment health status grading, and construct a health diagnosis prediction model;

[0102] Step S700: Collect and calculate the real-time health status feature vector, input the real-time health status feature vector into the health diagnosis prediction model, and output the predicted health status grading of the equipment;

[0103] In the application of this embodiment, a variety of sensors are used, including temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors, to comprehensively and real-time collect the operating parameters of industrial equipment. The temperature and humidity sensors are deployed at key environmental positions where the equipment is located. The power voltage sensors monitor the power parameters during the operation of the equipment, covering key indicators such as current and voltage. The lubricating oil parameter sensors focus on collecting relevant parameters of the lubricating oil, including temperature, pressure, viscosity, and the dielectric constant of the lubricating oil. For the temperature and humidity correlation data, the temperature and humidity correlation feedback value is calculated and generated. The coupling effect between the two is emphatically analyzed. In a high-humidity environment, changes in temperature may cause condensation on the surface of the equipment, thus accelerating the corrosion and damage of the equipment. For the power voltage correlation data, by analyzing the fluctuations, harmonic components, etc. of the current and voltage, the power voltage correlation feedback value is calculated. This value can reflect the power stability, energy consumption situation of the equipment, and whether there are potential electrical fault hazards. For the lubricating oil correlation data, based on the change of the dielectric constant, the lubricating oil correlation feedback value is calculated and generated, which can reflect the aging, pollution degree of the lubricating oil and its impact on the lubrication effect of the equipment. Based on the above three correlation feedback values, a health state feature vector is further generated. The feature vector synthesizes information from multiple aspects such as temperature and humidity, power voltage, and lubricating oil, and can comprehensively and accurately reflect the current health state of the equipment. At the same time, historical data of the operating parameters of industrial equipment are collected, combined with the health state feature vector, and a detailed equipment health state grading standard is set. The equipment health state grading includes multiple levels such as healthy, sub-healthy, fault warning, and fault, and each level corresponds to different equipment operating states and characteristics. By matching the historical data, the health state feature vector with the equipment health state grading, a health diagnosis model training set is generated. This training set contains a large amount of sample data. Based on the health diagnosis model training set, a health diagnosis prediction model is constructed using machine learning or deep learning algorithms. In actual application, the real-time health state feature vector is continuously collected and calculated, and input into the trained health diagnosis prediction model. The model outputs the predicted health state grading of the equipment according to the input real-time data.

[0104] As Figure 2 shown, as a preferred embodiment of the present invention, the obtaining of the temperature and humidity correlation data in the operating parameters of industrial equipment, calculating and generating the temperature and humidity correlation feedback value specifically includes:

[0105] Step S201: Extract and generate the temperature normalization value in the operating parameters of industrial equipment and the humidity normalization value ;

[0106] Step S202: Calculate the temperature and humidity correlation coefficient of the temperature normalization value and the humidity normalization value ;

[0107] The temperature - humidity correlation coefficient is calculated as follows:

[0108] ;

[0109] In the formula, is the time value, is the number of data samples, is the mean value of the temperature normalization value, is the mean value of the humidity normalization value;

[0110] Step S203: Based on the temperature - humidity correlation coefficient , calculate and generate the temperature - humidity coupling coefficient ;

[0111] The temperature - humidity coupling coefficient is calculated as follows:

[0112] ;

[0113] In the formula, and are weight coefficients, and , is the time weight factor, is the time interval.

[0114] When this embodiment is applied, temperature and humidity data are extracted from the operating parameters of industrial equipment and converted into temperature normalization values and humidity normalization values. The influence brought by different dimensions is eliminated, so that the data are on the same scale. On this basis, in - depth analysis is carried out on the temperature normalization value and the humidity normalization value, and the temperature - humidity correlation coefficient of the two is calculated. The calculation of the correlation coefficient adopts the Pearson correlation coefficient algorithm. By quantifying the degree of association between the two in the data change trend, the linear correlation between temperature and humidity is measured. In order to more comprehensively reflect the coupling effect of temperature and humidity, further based on the temperature - humidity correlation coefficient , calculate and generate the temperature - humidity coupling coefficient . This coefficient not only considers the linear correlation between temperature and humidity, but also incorporates information such as the amplitude and rate of change of both, so as to realize the quantitative evaluation of the comprehensive influence of temperature and humidity.

[0115] As Figure 3 shown, as a preferred embodiment of the present invention, the obtaining of the power voltage correlation data in the operating parameters of industrial equipment and calculating and generating the power voltage correlation feedback value specifically includes:

[0116] Step S301: Extract the current data and voltage data ;

[0117] Step S302: Based on Fourier transform, generate the current components and voltage components of each harmonic from the current data and voltage data ;

[0118] Step S303: Calculate the total harmonic distortion rate of current and the total harmonic distortion rate of current ;

[0119] The calculation process of the total harmonic distortion rate of current is as follows:

[0120] ;

[0121] The calculation process of the total harmonic distortion rate of current is as follows:

[0122] ;

[0123] In the formula, is the harmonic order, is the fundamental current, is the fundamental voltage;

[0124] Step S304: Based on the total harmonic distortion rate of current and the total harmonic distortion rate of current , respectively match the corresponding weight coefficients and ;

[0125] Step S305: Calculate the harmonic pollution degree index ;

[0126] The calculation process of the harmonic pollution degree index is as follows:

[0127] ;

[0128] In the formula, and are weight coefficients, and ;

[0129] When this embodiment is applied, the current data and voltage data in the operating parameters of the industrial equipment are extracted. Subsequently, with the help of Fourier transform, the current and voltage data in the time domain are converted to the frequency domain to generate the current components and voltage components . Based on these components, calculate the total harmonic distortion rate of the current and the total harmonic distortion rate of the voltage , which measure the degree to which the current and voltage waveforms deviate from the ideal sine wave. To more accurately evaluate the impact of harmonic pollution on equipment, for and , corresponding weight coefficients are respectively matched. The determination of the weight coefficients is based on factors such as equipment type and operating environment. Finally, calculate the harmonic pollution degree index , forming the power voltage correlation feedback value.

[0130] As Figure 4 shown, as a preferred embodiment of the present invention, the obtaining of the lubricating oil correlation data in the operating parameters of industrial equipment, calculating and generating the lubricating oil correlation feedback value specifically includes:

[0131] Step S401: Extract the dielectric constant of the lubricating oil in the operating parameters of the industrial equipment;

[0132] Step S402: Calculate and generate the mean value and the standard deviation of the dielectric constant of the lubricating oil;

[0133] Step S403: Calculate and generate the deviation degree of the dielectric constant of the lubricating oil;

[0134] The calculation process of the deviation degree of the dielectric constant of the lubricating oil is:

[0135] ;

[0136] Step S404: Based on the preset deviation degree threshold , calculate and generate the quantization index of the dielectric constant of the lubricating oil;

[0137] The calculation process of the quantization index of the dielectric constant of the lubricating oil is:

[0138] ;

[0139] It should be understood that first, use the lubricating oil parameter sensor to accurately extract the dielectric constant of the lubricating oil from the operating parameters of the industrial equipment. The dielectric constant of the lubricating oil is the core index reflecting its insulation performance and chemical stability. Collect the dielectric constant data of the lubricating oil for a period of time, and use statistical methods to calculate its mean value and the standard deviation . The mean value can reflect the overall level of the dielectric constant of the lubricating oil, and the standard deviation Then, the degree of data dispersion is measured, and the combination of the two can help judge the stability of the dielectric constant of the lubricating oil. On this basis, the deviation degree of the dielectric constant of the lubricating oil is calculated. . Based on the preset deviation degree threshold, a quantization index of the dielectric constant of the lubricating oil is generated. When the deviation degree of the dielectric constant exceeds the preset threshold, it indicates that the lubricating oil may be contaminated, oxidized or aged, and its performance has deteriorated.

[0140] For example Figure 5 As shown, as a preferred embodiment of the present invention, the generation of the health state feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value specifically includes:

[0141] Step S501: Calculate the change rate feature of the temperature and humidity coupling coefficient and the fluctuation range feature ;

[0142] Step S502: Calculate the change trend feature of the harmonic pollution degree index , the change trend feature of the content of each harmonic , and the harmonic pollution mutation feature ;

[0143] Step S503: Calculate the change frequency feature of the quantization index of the lubricating oil dielectric constant ;

[0144] Step S504: Combine the features to generate a targeted feature vector ;

[0145] When this embodiment is applied, in the stage of generating the health state feature vector, the key information in each correlation feedback value is further mined. For the temperature and humidity coupling coefficient , calculate its change rate feature and fluctuation range feature. The former reflects the change speed of the temperature and humidity coupling effect over time, and the latter reflects its fluctuation amplitude within a certain time. For the harmonic pollution degree index , analyze its change trend feature to judge whether the harmonic pollution is gradually increasing or decreasing; calculate the change trend feature of the content of each harmonic to clarify the change of the influence of different frequency harmonics on the equipment; at the same time, capture the harmonic pollution mutation feature to timely discover sudden harmonic events that may cause equipment failures. For the quantization index of the lubricating oil dielectric constant , calculate its change frequency feature, and through this feature, the speed of lubricating oil performance deterioration can be judged. Finally, the above features are organically combined to generate a targeted feature vector , this vector comprehensively reflects the operating status of the device. The generated targeted feature vector is used as the core data of the training set of the health diagnosis model, and a high-precision health diagnosis prediction model is constructed with the help of machine learning or deep learning algorithms.

[0146] As Figure 6 shown, as a preferred embodiment of the present invention, the steps of obtaining historical data of industrial equipment operating parameters, health status feature vectors, setting the health status classification of the equipment, generating a training set for the health diagnosis model based on the historical data of industrial equipment operating parameters, health status feature vectors and the health status classification of the equipment, and constructing a health diagnosis prediction model specifically include:

[0147] Step S601: Set the health status classification;

[0148] The health status classification includes: healthy level, sub-healthy level, fault critical warning level, and fault level;

[0149] Step S602: Collect historical data of industrial equipment operating parameters and generate a training set for the health diagnosis model based on the targeted feature vector and the health status classification;

[0150] Step S603: Train the training set of the health diagnosis model through a support vector machine and a convolutional neural network, and use the cross-validation method for optimization to generate a health diagnosis prediction model ;

[0151] In the application of this embodiment, the health status of the device is divided into four levels: healthy level, sub-healthy level, fault critical warning level, and fault level. Collect historical data of industrial equipment operating parameters, combine the targeted feature vector with the health status classification to construct a training set for the health diagnosis model, and use the support vector machine and convolutional neural network algorithms to train the training set of the health diagnosis model. During the training process, the cross-validation method is used to optimize the model. By repeatedly dividing the training set and the validation set, and training and evaluating the model repeatedly, the overfitting phenomenon is effectively avoided, and the generalization ability and diagnosis accuracy of the model are improved. After multiple rounds of training and optimization, a health diagnosis prediction model is generated .

[0152] As Figure 7 shown, as a preferred embodiment of the present invention, the steps of collecting and calculating the real-time health status feature vector, inputting the real-time health status feature vector into the health diagnosis prediction model, and outputting the predicted health status classification of the device specifically include:

[0153] Step S701: Input the extracted real-time feature vector into the health diagnosis prediction model and output the current health status level of the industrial equipment ;

[0154] Step S702: Based on the current health status level of the industrial equipment , match and send corresponding warning action information;

[0155] When this embodiment is applied, the real-time operation data of the industrial equipment is continuously collected, and according to the data processing flow described above, the real-time health status feature vector is calculated. The extracted real-time feature vector is input into the health diagnosis and prediction model . The model outputs the current health status level of the industrial equipment according to the rules and patterns obtained from training; according to the current health status level of the industrial equipment, the system automatically matches and sends corresponding warning action information. When the equipment is in the healthy level, routine monitoring is maintained; when it is in the sub-healthy level, a prompt message is sent, suggesting strengthening the equipment inspection; when it is in the fault critical warning level, a warning signal is sent to remind relevant personnel to prepare for equipment maintenance; when it is in the fault level, an emergency alarm is immediately triggered to notify the maintenance personnel to conduct fault troubleshooting and repair, and at the same time, the production system is linked to take corresponding emergency measures, such as suspending the equipment operation to prevent the fault from expanding.

[0156] As Figure 8 shown, as another preferred embodiment of the present invention, on the other hand, an industrial equipment health diagnosis system for the industrial Internet, the system includes:

[0157] The first acquisition module 100 is used to acquire the operation parameters of the industrial equipment monitored by several types of sensors;

[0158] The several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors;

[0159] The first acquisition module 200 is used to acquire the temperature and humidity correlation data in the operation parameters of the industrial equipment;

[0160] The first calculation and generation module 300 is used to calculate and generate the temperature and humidity correlation feedback value;

[0161] The second acquisition module 400 is used to acquire the power voltage correlation data in the operation parameters of the industrial equipment;

[0162] The second calculation and generation module 500 is used to calculate and generate the power voltage correlation feedback value;

[0163] The third acquisition module 600 is used to acquire the lubricating oil correlation data in the operation parameters of the industrial equipment;

[0164] The third calculation and generation module 700 is used to calculate and generate the lubricating oil correlation feedback value;

[0165] A generation module 800, configured to generate a health status feature vector based on a humidity correlation feedback value, a power voltage correlation feedback value, and a lubricating oil correlation feedback value;

[0166] A prediction model module 900, configured to obtain historical data of industrial equipment operation parameters and a health status feature vector, set a classification of the equipment health status, generate a health diagnosis model training set based on the historical data of industrial equipment operation parameters, the health status feature vector, and the classification of the equipment health status, and construct a health diagnosis prediction model;

[0167] An input / output module 1000, configured to collect and calculate a real-time health status feature vector, input the real-time health status feature vector into the health diagnosis prediction model, and output a predicted health status classification of the equipment.

[0168] When this embodiment is applied, a first collection module 100 collects industrial equipment operation parameters monitored by several types of sensors, a first acquisition module 200 acquires humidity-temperature correlation data in the industrial equipment operation parameters, a first calculation and generation module 300 calculates and generates a humidity-temperature correlation feedback value, a second acquisition module 400 acquires power voltage correlation data in the industrial equipment operation parameters, a second calculation and generation module 500 calculates and generates a power voltage correlation feedback value, a third acquisition module 600 acquires lubricating oil correlation data in the industrial equipment operation parameters, a third calculation and generation module 700 calculates and generates a lubricating oil correlation feedback value, a generation module 800 generates a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value, obtains historical data of industrial equipment operation parameters and the health status feature vector, sets a classification of the equipment health status, generates a health diagnosis model training set based on the historical data of industrial equipment operation parameters, the health status feature vector, and the classification of the equipment health status, a prediction model module 900 constructs a health diagnosis prediction model, collects and calculates a real-time health status feature vector, and an input / output module 1000 inputs the real-time health status feature vector into the health diagnosis prediction model and outputs a predicted health status classification of the equipment.

[0169] As Figure 9 shown, as another preferred embodiment of the present invention, the first calculation and generation module 300 specifically includes:

[0170] An extraction and generation unit 301, configured to extract and generate a temperature normalization value in the industrial equipment operation parameters and a humidity normalization value ;

[0171] A first calculation unit 302, configured to calculate a humidity-temperature correlation coefficient of the temperature normalization value and the humidity normalization value ; ;

[0172] The calculation and generation unit 303 is configured to calculate and generate a temperature-humidity coupling coefficient based on the temperature-humidity correlation coefficient . .

[0173] In the application of this embodiment, the extraction and generation unit 301 extracts and generates the temperature normalization value in the operating parameters of the industrial equipment and the humidity normalization value . The first calculation unit 302 calculates the temperature-humidity correlation coefficient of the temperature normalization value and the humidity normalization value . Based on the temperature-humidity correlation coefficient , the calculation and generation unit 303 calculates and generates the temperature-humidity coupling coefficient . .

[0174] As Figure 10 shown, as another preferred embodiment of the present invention, the second calculation and generation module 500 specifically includes:

[0175] An extraction unit 501 is configured to extract the current data and the voltage data in the operating parameters of the industrial equipment;

[0176] A generation unit 502 generates the current components of each harmonic and the voltage components of each harmonic by performing a Fourier transform on the current data and the voltage data ;

[0177] A second calculation unit 503 is configured to calculate the total harmonic distortion rate of the current and the total harmonic distortion rate of the current ;

[0178] A matching unit 504 is configured to respectively match the corresponding weight coefficients and based on the total harmonic distortion rate of the current and ;

[0179] A third calculation unit 505 is configured to calculate the harmonic pollution degree index .

[0180] In the application of this embodiment, the extraction unit 501 extracts the current data and the voltage data in the operating parameters of the industrial equipment, and the generation unit 502 generates the current components of each harmonic and the voltage components of each harmonic by performing a Fourier transform on the current data Sum voltage component , the second calculation unit 503 calculates the total current harmonic distortion rate and the total current harmonic distortion rate , based on the total current harmonic distortion rate and the total current harmonic distortion rate , the matching unit matches the corresponding weight coefficients and , the third calculation unit 505 calculates the harmonic pollution degree index .

[0181] In the above embodiments of the present invention, an industrial equipment health diagnosis method for the industrial Internet is provided, and an industrial equipment health diagnosis system for the industrial Internet is provided. A variety of types of sensors are used, including temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors, to comprehensively and real-time collect the operating parameters of industrial equipment. The temperature and humidity sensors are deployed at key environmental positions where the equipment is located. The power voltage sensors monitor the power parameters during the operation of the equipment, covering key indicators such as current and voltage. The lubricating oil parameter sensors focus on collecting relevant parameters of the lubricating oil, such as temperature, pressure, viscosity, and the dielectric constant of the lubricating oil. For the temperature and humidity correlation data, the temperature and humidity correlation feedback value is calculated and generated. The coupling effect between the two is emphatically analyzed. In a high-humidity environment, changes in temperature may cause condensation on the surface of the equipment, thereby accelerating the corrosion and damage of the equipment. For the power voltage correlation data, by analyzing the fluctuations, harmonic components, etc. of the current and voltage, the power voltage correlation feedback value is calculated. This value can reflect the power stability, energy consumption situation of the equipment, and whether there are potential electrical fault hazards. For the lubricating oil correlation data, based on the change of the dielectric constant, the lubricating oil correlation feedback value is calculated and generated, which can reflect the aging, pollution degree of the lubricating oil, and the impact on the lubrication effect of the equipment. Based on the above three correlation feedback values, a health status feature vector is further generated. The feature vector synthesizes information from multiple aspects such as temperature and humidity, power voltage, and lubricating oil, and can comprehensively and accurately reflect the current health status of the equipment. At the same time, historical data of the operating parameters of industrial equipment are collected, combined with the health status feature vector, and a detailed equipment health status grading standard is set. The equipment health status grading includes multiple levels such as healthy, sub-healthy, fault warning, and fault, and each level corresponds to different equipment operating states and characteristics. By matching the historical data, the health status feature vector with the equipment health status grading, a health diagnosis model training set is generated. This training set contains a large amount of sample data. Based on the health diagnosis model training set, a health diagnosis prediction model is constructed using machine learning or deep learning algorithms. In practical applications, the real-time health status feature vector is continuously collected and calculated, and input into the trained health diagnosis prediction model. The model outputs the predicted health status grading of the equipment according to the input real-time data. This method and system convert the implicit correlations of multi-modal, etc. into quantifiable health feature vectors through multi-sensor data fusion analysis, and construct an equipment health assessment system coupled with multiple physical fields. The intelligent diagnosis model based on the dynamic feature vector realizes the prediction of fault trends, and continuously improves the diagnosis accuracy in combination with the closed-loop optimization mechanism, so as to support predictive maintenance decisions, reduce the risk of unplanned downtime, not only break through the limitations of traditional single-parameter monitoring, but also provide technical support for industrial equipment and realize the full life cycle management of equipment.

[0182] In order to enable the smooth operation of the above-mentioned method and system, in addition to the various modules described above, the system may further include more or fewer components than those described above, or combine certain components, or different components. For example, it may include input / output devices, network access devices, buses, processors, and memories, etc.

[0183] The so-called processor may be a central processing unit, or may also be other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc. The above-mentioned processor is the control center of the above-mentioned system, and connects each part through various interfaces and lines.

[0184] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0185] The above-described embodiments only represent several implementation manners of the present invention, and their descriptions are relatively specific and detailed, but should not be construed as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

[0186] The above is only the preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An industrial equipment health diagnosis method for the industrial Internet, characterized in that, The method includes: Collecting the operation parameters of industrial equipment monitored by several types of sensors; the several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors; Obtaining the temperature and humidity correlation data in the operation parameters of the industrial equipment, calculating and generating a temperature and humidity correlation feedback value; Obtaining the power voltage correlation data in the operation parameters of the industrial equipment, calculating and generating a power voltage correlation feedback value; Obtaining the lubricating oil correlation data in the operation parameters of the industrial equipment, calculating and generating a lubricating oil correlation feedback value; Generating a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value; Obtaining the historical data of the operation parameters of the industrial equipment and the health status feature vector, setting the equipment health status grading, and generating a health diagnosis model training set based on the historical data of the operation parameters of the industrial equipment, the health status feature vector, and the equipment health status grading, and constructing a health diagnosis prediction model; Collecting and calculating the real-time health status feature vector, inputting the real-time health status feature vector into the health diagnosis prediction model, and outputting the predicted health status grading of the equipment.

2. The industrial equipment health diagnosis method for industrial Internet according to claim 1, characterized in that The obtaining of the temperature and humidity correlation data in the operation parameters of the industrial equipment, calculating and generating a temperature and humidity correlation feedback value specifically includes: Extract and generate the temperature normalization value in the operating parameters of industrial equipment and the humidity normalization value ; Calculate the normalized temperature value and the normalized humidity value of the temperature-humidity correlation coefficient ; The temperature and humidity correlation coefficient is calculated as follows: ； In the formula, is the time value, is the number of data samples, is the average value of temperature normalization values, is the average value of humidity normalization values; Based on the temperature-humidity correlation coefficient , calculate and generate the temperature-humidity coupling coefficient ; The temperature-humidity coupling coefficient The calculation process is as follows: ； In the formula, and are weight coefficients, and , is the time weight factor, is the time interval.

3. The industrial equipment health diagnosis method for the industrial Internet according to claim 2, wherein The obtaining of the power voltage correlation data in the operation parameters of the industrial equipment, calculating and generating a power voltage correlation feedback value specifically includes: Extract the current data and voltage data from the operating parameters of industrial equipment and voltage data ; The current data and voltage data Based on the Fourier transform, generate the current components of each harmonic and voltage components ; Calculate the total harmonic distortion rate of current and the total harmonic distortion rate of sum current ; The total current harmonic distortion rate is calculated as follows: ； The total harmonic distortion rate of the current The calculation process is as follows: ； In the formula, is the harmonic order, is the fundamental current, is the fundamental voltage; Based on the total harmonic distortion rate of current and the total harmonic distortion rate of current , respectively match the corresponding weight coefficients and ; Calculate the harmonic pollution degree index ; The harmonic pollution degree index The calculation process is as follows: ； In the formula, and are weight coefficients, and .

4. The industrial equipment health diagnosis method for industrial Internet according to claim 3, characterized in that, The obtaining of the lubricating oil correlation data in the operation parameters of the industrial equipment, calculating and generating a lubricating oil correlation feedback value specifically includes: Extracting the dielectric constant of the lubricating oil in the operation parameters of the industrial equipment; Calculate and generate the mean value of the dielectric constant of the lubricating oil and the standard deviation value ; Calculate and generate the deviation of the dielectric constant of the lubricating oil ; The deviation degree of the dielectric constant of the lubricating oil The calculation process is as follows: ； Based on a preset deviation threshold , calculate and generate a quantization index for the dielectric constant of the lubricating oil ; The quantitative index of the dielectric constant of the lubricating oil The calculation process is as follows: 。 5. The industrial equipment health diagnosis method for the industrial Internet according to claim 4, characterized in that The generating of a health status feature vector based on the humidity correlation feedback value, the power voltage correlation feedback value, and the lubricating oil correlation feedback value specifically includes: Calculate the coupling coefficient of temperature and humidity Characteristics of the rate of change and characteristics of the fluctuation range ; Calculation of harmonic pollution degree index Trend characteristics of and trend characteristics of the content of each harmonic as well as harmonic pollution mutation characteristics ; Calculation of the quantization index of the dielectric constant of lubricating oil Frequency characteristics of the change ; Combine various features to generate a targeted feature vector .

6. The industrial equipment health diagnosis method for industrial Internet according to claim 5, characterized in that The obtaining of the historical data of the operation parameters of the industrial equipment and the health status feature vector, setting the equipment health status grading, and generating a health diagnosis model training set based on the historical data of the operation parameters of the industrial equipment, the health status feature vector, and the equipment health status grading, and constructing a health diagnosis prediction model specifically includes: Setting the health status grading, and the health status grading includes: healthy level, sub-healthy level, fault critical warning level, and fault level; Collect historical data of industrial equipment operation parameters and generate a training set for a health diagnosis model based on targeted feature vectors and health status classification; Train the training set of the health diagnosis model using support vector machines and convolutional neural networks, and optimize it using the cross-validation method to generate a health diagnosis prediction model .

7. The industrial equipment health diagnosis method for industrial Internet according to claim 1, characterized in that The collecting and calculating of the real-time health status feature vector, inputting the real-time health status feature vector into the health diagnosis prediction model, and outputting the predicted health status grading of the equipment specifically includes: Input the extracted real-time feature vectors into the health diagnosis and prediction model to output the current health status level of the industrial equipment ; Based on the current health status level of industrial equipment , match and send corresponding warning action information.

8. An industrial equipment health diagnosis system for the industrial Internet, characterized in that, Applying the industrial equipment health diagnosis method for industrial Internet according to any one of claims 1-7, the system includes: A first collection module for collecting the operation parameters of industrial equipment monitored by several types of sensors; The several types of sensors include: temperature and humidity sensors, power voltage sensors, and lubricating oil parameter sensors; A first obtaining module for obtaining the temperature and humidity correlation data in the operation parameters of the industrial equipment; A first calculation and generation module for calculating and generating a temperature and humidity correlation feedback value; A second obtaining module for obtaining the power voltage correlation data in the operation parameters of the industrial equipment; A second calculation and generation module for calculating and generating a power voltage correlation feedback value; A third obtaining module for obtaining the lubricating oil correlation data in the operation parameters of the industrial equipment; The third calculation and generation module is used to calculate and generate the lubricating oil associated feedback value; The generation module is used to generate a health status feature vector based on the humidity associated feedback value, the power voltage associated feedback value, and the lubricating oil associated feedback value; The prediction model module is used to obtain the historical data of the industrial equipment operation parameters and the health status feature vector, set the equipment health status grading, generate a health diagnosis model training set based on the historical data of the industrial equipment operation parameters, the health status feature vector, and the equipment health status grading, and construct a health diagnosis prediction model; The input / output module is used to collect and calculate the real-time health status feature vector, input the real-time health status feature vector into the health diagnosis prediction model, and output the predicted health status grading of the equipment.

9. The industrial equipment health diagnosis system for industrial Internet according to claim 8, wherein The first calculation and generation module specifically includes: An extraction and generation unit for extracting and generating the temperature normalization value among the operating parameters of industrial equipment and the humidity normalization value ; A first calculation unit for calculating a temperature normalization value and a humidity normalization value of the temperature-humidity correlation coefficient ; A calculation and generation unit, configured to calculate and generate a temperature-humidity coupling coefficient based on a temperature-humidity correlation coefficient 。 。 10. The industrial equipment health diagnosis system for industrial Internet according to claim 8, wherein The second calculation and generation module specifically includes: An extraction unit for extracting current data and voltage data from the operating parameters of industrial equipment and voltage data ; A generating unit that generates current data and voltage data Based on Fourier transform, generates current components of each harmonic and voltage components ; A second calculation unit for calculating the total harmonic distortion rate of the current and the total harmonic distortion rate of the current ; A matching unit, configured to match corresponding weight coefficients respectively based on the total harmonic distortion rate of current and the total harmonic distortion rate of current , and ; The third computing unit is used to calculate the harmonic pollution degree index .

Images