Industrial Internet Data Real-Time Prediction System and Method

Abstract

The present invention discloses an industrial Internet of Things data real-time prediction system and method, which relates to the technical field of industrial equipment data prediction. The prediction system includes an information collection module, an operating status evaluation module, a status determination module, and an overhaul plan adjustment module. The technical key points are as follows: on the premise of the stability of industrial equipment, the initial overhaul frequency Jp is obtained according to the current operating status evaluation value Ose. On the premise of fluctuations in industrial equipment, it is further determined whether there is manual control. After excluding the influence brought by manual control, the overhaul adjustment frequency Jtp is obtained according to the initial overhaul frequency Jp and the maximum difference of the operating status evaluation value Ose exceeding the standard threshold group extracted within the preset time period S, realizing precise overhaul control of industrial equipment in different states, ensuring to a certain extent that potential faults or abnormalities can be eliminated through overhaul before the industrial equipment fails, and ensuring the stability of the operation of industrial equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment data prediction, and specifically to a real-time prediction system and method for industrial Internet data. Background Technique

[0002] Fault prediction of industrial equipment based on data is a technology based on data analysis and machine learning. By collecting, analyzing, and modeling the operation data of the equipment, the probability or time window of possible equipment failures can be predicted. When making predictions, it is necessary to comprehensively consider the parameters generated by the corresponding industrial equipment. Each parameter is summarized to the prediction system through the Internet, and then comprehensive prediction and analysis processing are carried out. When making predictions, it is usually necessary to establish a model. A fault prediction model is established based on characteristic data. Commonly used models include regression models, support vector machines, neural networks, decision trees, etc. Time series analysis methods such as ARIMA and LSTM can also be used.

[0003] Currently, when detecting abnormalities in the parameters of industrial equipment, traditional fault detection systems set standard range values for each parameter in the industrial equipment, and issue early warning operations when the corresponding range values are exceeded. When predicting potential faults or equipment abnormalities, a prediction model is usually built for processing. However, even after predicting potential faults or abnormalities, for industrial equipment, the maintenance frequency needs to be adjusted. Currently, the adjustment of the maintenance frequency requires manual control. Even if the time point of future abnormalities in industrial equipment can be predicted, due to the different states of industrial equipment, the current maintenance regulation of industrial equipment still cannot achieve precise adjustment. If the maintenance of industrial equipment is not timely enough subsequently, the failure rate of the corresponding industrial equipment will increase significantly, which will also affect the stability of the operation of industrial equipment to a certain extent. Summary of the Invention

[0004] (I) Technical Problems to be Solved

[0005] In view of the deficiencies of the prior art, the present invention provides a real-time prediction system and method for industrial Internet data. Under the condition of judging whether the industrial equipment is stable, corresponding maintenance strategies are given respectively. On the premise that the industrial equipment is stable, the initial maintenance frequency Jp is obtained based on the current operation state evaluation value Ose. On the premise that the industrial equipment fluctuates, it is further judged whether there is manual regulation. Under the condition of excluding the influence brought by manual regulation, the maintenance adjustment frequency Jtp is obtained based on the initial maintenance frequency Jp and the maximum difference between the operation state evaluation values Ose exceeding the standard threshold group within the preset time period S, realizing precise maintenance regulation of industrial equipment in different states. To a certain extent, it ensures that the industrial equipment can eliminate potential fault hazards or abnormalities through maintenance before a failure occurs, and ensures the stability of the operation of industrial equipment, solving the problems raised in the background technique.

[0006] (2) Technical solution

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions:

[0008] An industrial Internet of Things data real-time prediction system, comprising:

[0009] An information collection module, which acquires the parameters during the operation of the corresponding industrial equipment, and the parameters include the operating temperature, pressure value, actual power value, and environmental index;

[0010] An operating status evaluation module, which builds a data analysis model based on the preprocessed parameters to generate an operating status evaluation value Ose. Based on the operating status evaluation values Ose at each moment, a deep learning algorithm is used for modeling to obtain a fault prediction model, and the predicted operating status evaluation value Ose is displayed in the form of a curve graph based on this fault prediction model;

[0011] A status determination module, which compares the predicted operating status evaluation value Ose of the corresponding industrial equipment with a preset standard threshold group according to the curve graph of the predicted operating status evaluation value Ose. When the predicted operating status evaluation value Ose in the future is outside the range of the standard threshold group, an operation instruction for secondary determination is triggered;

[0012] An overhaul plan adjustment module, which executes the operation of secondary determination, determines whether it is a manual adjustment. If so, it observes the operation of the industrial equipment within a predetermined time H. If not, it continues to adjust the overhaul frequency based on the initial overhaul strategy.

[0013] Furthermore, the operating temperature is obtained by uniformly arranging a plurality of temperature sensors at positions on the surface of the corresponding industrial equipment in the operating area, and the average value of each temperature value at the same moment is the required operating temperature.

[0014] Furthermore, the environmental index represents the data of the environment in the workshop where the corresponding industrial equipment is located, and the calculation method of the environmental index is as follows:

[0015]

[0016] In the formula, Hz represents the environmental index, Tw represents the environmental temperature value in the workshop where the corresponding industrial equipment is located, Ts represents the environmental humidity value in the workshop where the corresponding industrial equipment is located, α and β are respectively the preset proportionality coefficients of the environmental temperature value and the environmental humidity value, and α > β > 0.

[0017] Furthermore, the preprocessing process of the parameters is: cleaning each data in the parameters, and then performing dimensionless processing on each parameter.

[0018] Furthermore, the formula for generating the operating status evaluation value Ose is as follows:

[0019]

[0020] Wherein, Wd represents the operating temperature, Yl represents the pressure value, U represents the actual output pressure, I represents the actual output current, (UI) represents the actual power value, Hz represents the environmental index, G is a constant correction factor, e is the base of the natural logarithm function, a1, a2, a3, and a4 are the influence factors of the operating temperature, pressure value, actual power value, and environmental index respectively, and a3 > a1 > a2 > a4 > 0.

[0021] Furthermore, the process of using a deep learning algorithm for modeling to obtain a fault prediction model is as follows:

[0022] S101. Collect a historical data set containing the operating parameters of industrial equipment and the corresponding operating status evaluation value Ose, perform feature extraction on the collected data, and convert the time series data into a format suitable for RNN input;

[0023] S102. Based on the historical data set, construct an RNN model. After selecting the LSTM structure, divide the historical data set into a training set and a validation set. During this period, use the cross-validation method, input the training set into the RNN model for training, and adjust the model parameters to minimize the loss function;

[0024] S103. Input the validation set into the trained model for validation, optimize the model according to the validation results, and then save the trained model;

[0025] S104. Input the operating status evaluation value Ose obtained by real-time calculation into the model for prediction calculation, obtain the predicted value of the operating status evaluation value Ose of the industrial equipment at a future moment, and display the predicted operating status evaluation value Ose in the form of a curve graph.

[0026] Furthermore, after comparing the operating status evaluation value Ose of the corresponding industrial equipment in the future with a preset standard threshold group, if the operating status evaluation value Ose in the future is within the range of the standard threshold group, then formulate a maintenance strategy based on the current operating status evaluation value Ose. The preset standard threshold group is the first standard threshold oml1 and the second standard threshold oml2 respectively, and oml1 < oml2.

[0027] Furthermore, the process of formulating a maintenance strategy based on the current operating status evaluation value Ose is as follows:

[0028] S201. Obtain the operating status evaluation value Ose at different moments within T time t , where t represents the number of the operating status evaluation value at different moments within T time, t = 1, 2, 3, 4,..., n, and n is a positive integer;

[0029] S202. Calculate the average value of the operation status evaluation values at different moments within the T time and the operation status evaluation value Ose t , and calculate the operation status fluctuation coefficient Osfc. The calculation formula is as follows:

[0030]

[0031] S203. Calculate and obtain the maintenance frequency based on the operation status fluctuation coefficient Osfc. The calculation formula is as follows:

[0032]

[0033] In the formula, Jp represents the maintenance frequency, St represents the usage duration of the corresponding industrial equipment, b1 and b2 are respectively the preset proportionality coefficients of the usage duration and the operation status fluctuation coefficient of the corresponding industrial equipment, and b2 > b1 > 0, C is a constant correction coefficient, and int is the rounding function.

[0034] Furthermore, under the condition of determining that it is manually regulated, observe the operation of the industrial equipment within the predetermined time H. If, after the predetermined time H, the operation status evaluation value Ose is within the range of the standard threshold group, then continue to execute the initial maintenance strategy; otherwise, continue to adjust the maintenance frequency on the basis of the initial maintenance strategy;

[0035] The process of continuing to adjust the maintenance frequency on the basis of the initial maintenance strategy is as follows:

[0036] S301. Extract the maximum difference between the operation status evaluation value Ose and the standard threshold group within the preset time period S;

[0037] S302. Based on the maximum difference and the maintenance frequency obtained under the initial maintenance strategy, re - construct the data analysis model to generate the maintenance adjustment frequency. The formula is as follows:

[0038] Jtp = int(c1 * Jp + ln(c2 * Cs max + 1) 2 )

[0039] In the formula, Jtp represents the maintenance adjustment frequency, Cs max represents the maximum difference, c1 and c2 are respectively the influence factors of the maintenance adjustment frequency and the maximum difference, c2 > c1 > 0, 1 > c1 > 0, and int is the rounding function.

[0040] The industrial Internet data real - time prediction method includes the following steps: S1. Obtain the parameters during the operation of the corresponding industrial equipment, and the parameters include the operating temperature, pressure value, actual power value, and environmental index;

[0041] S2. Based on the pre - processed parameters, build a data analysis model to generate the operation status evaluation value Ose. Based on the operation status evaluation values Ose at each moment, use deep - learning algorithms to build a model to obtain a fault prediction model. Based on this fault prediction model, display the predicted operation status evaluation value Ose in the form of a curve graph;

[0042] S3. Based on the curve graph of the predicted operation status evaluation value Ose, compare the future operation status evaluation value Ose of the corresponding industrial equipment with the preset standard threshold group. If the future operation status evaluation value Ose is within the range of the standard threshold group, formulate a maintenance strategy according to the current operation status evaluation value Ose. The preset standard threshold group is the first standard threshold oml1 and the second standard threshold oml2 respectively, and oml1 < oml2. If the future operation status evaluation value Ose is outside the range of the standard threshold group, trigger an operation instruction for secondary determination;

[0043] S4. Execute the operation of secondary determination to determine whether it is a manual regulation. If so, observe the operation of the industrial equipment within the predetermined time H. Under the condition of determining that it is a manual regulation, observe the operation of the industrial equipment within the predetermined time H. If the operation status evaluation value Ose is within the range of the standard threshold group after the predetermined time H, continue to execute the initial maintenance strategy. Otherwise, adjust the maintenance frequency on the basis of the initial maintenance strategy.

[0044] (III) Beneficial effects

[0045] The present invention provides an industrial Internet data real - time prediction system and method, which have the following beneficial effects:

[0046] 1. When predicting the abnormal situation of industrial equipment, different from the traditional fault detection system that sets standard range values for each parameter in the industrial equipment and gives an early warning operation when the corresponding range value is exceeded, this application comprehensively considers various factors during the operation of industrial equipment, including equipment parameters and the surrounding environment, so as to obtain a comprehensive evaluation value, that is, the operation status evaluation value Ose. By establishing a fault prediction model, the curve distribution of the operation status evaluation value Ose can be analyzed. To a certain extent, it can accurately and effectively obtain whether there are potential fault or abnormal problems in the corresponding industrial equipment, and give solutions and maintenance strategies in time, reflecting the flexibility of the system design;

[0047] 2. When the present invention determines whether an industrial device is in a stable condition, corresponding maintenance strategies are given respectively. On the premise that the industrial device is stable, the initial maintenance frequency Jp is obtained based on the current operation status evaluation value Ose. On the premise that the industrial device fluctuates, it is continuously determined whether there is manual regulation. Under the condition of excluding the influence brought by manual regulation, the maintenance adjustment frequency Jtp is obtained based on the initial maintenance frequency Jp and the maximum difference of the operation status evaluation value Ose exceeding the standard threshold group extracted within the preset time period S, realizing precise maintenance regulation of industrial devices in different states, and ensuring to a certain extent that potential faults or abnormalities of industrial devices can be eliminated through maintenance before a failure occurs, and ensuring the stability of the operation of industrial devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the module structure of the industrial Internet data real-time prediction system in the present invention;

[0049] Figure 2 It is a schematic diagram of the overall process of the industrial Internet data real-time prediction method in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0051] Embodiment 1: Please refer to Figure 1 , this embodiment provides an industrial Internet data real-time prediction system. The prediction system includes an information collection module, an operation status evaluation module, a status determination module, and a maintenance plan adjustment module that run in sequence. And the entire prediction system evaluates the parameters generated during the operation of industrial devices. These parameters are summarized through the Internet and are uniformly processed, analyzed, or predicted by the prediction system;

[0052] The information collection module obtains the parameters during the operation of the corresponding industrial device;

[0053] Among them, the corresponding industrial devices include large-scale production equipment in industrial workshops, sensor devices attached to or configured on large-scale production equipment, which are used to collect various environmental parameters and index data, including temperature, pressure, humidity, vibration, light intensity, etc.; for example, temperature sensors, pressure sensors, and vibration sensors, energy equipment, that is, the energy required to supply industrial workshops and production equipment, such as electricity, steam, and gas, for example, generators, boilers, and gas pumps;

[0054] The parameters corresponding to the operation of industrial equipment include at least the operating temperature, pressure value, actual power value, and environmental index;

[0055] The operating temperature refers to the heat value emitted by industrial equipment during operation. Too high or too low temperature may affect the stability and reliability of industrial equipment, and temperature changes may cause problems such as thermal expansion and contraction of equipment, material fatigue, and changes in lubricant performance. Therefore, the operating temperature is data that affects the operating status of industrial equipment. For the operating temperature, several temperature sensors are set at the positions of the corresponding industrial equipment surface in the operating area to obtain the average value of each temperature value at the same time, which is the required operating temperature.

[0056] The pressure value indicates the internal pressure value of industrial equipment during operation. Pressure exceeding the tolerance range of the equipment will increase the stress and risk of the equipment, which may cause equipment failure or damage. The pressure value can be obtained in the same way as the operating temperature, so it will not be elaborated here.

[0057] The actual power value indicates the power value of the corresponding industrial equipment when it is actually working. The actual power value is calculated as the product of the actual output voltage and the actual output current. The actual power value also needs to be kept within a certain range to ensure the stability of the corresponding industrial equipment's operating state.

[0058] The environmental index represents the data of the environment in the factory where the corresponding industrial equipment is located, reflecting the environmental conditions of the corresponding industrial equipment, and the calculation method of the environmental index is as follows:

[0059]

[0060] Wherein, Hz represents the environmental index, Tw represents the ambient temperature value in the factory where the corresponding industrial equipment is located, Ts represents the ambient humidity value in the factory where the corresponding industrial equipment is located, α and β are the preset proportional coefficients of the ambient temperature value and the ambient humidity value, respectively, and α>β>0; the collection of the ambient temperature value and the ambient humidity value is obtained in real time by installing a temperature and humidity sensor in the factory.

[0061] It should be noted that among the parameters of the corresponding industrial equipment during operation, if any of the operating temperature, pressure value, actual power value, and environmental index exceeds the corresponding standard range value, the system will issue an early warning to remind the staff to perform targeted maintenance. This application considers the possible hidden dangers of failure of the corresponding industrial equipment when any data does not exceed the corresponding standard range value.

[0062] Specifically, when predicting abnormal conditions of industrial equipment, different from setting standard range values for each parameter in industrial equipment by traditional fault detection systems and giving early warning operations when the corresponding range values are exceeded, this application comprehensively considers various factors during the operation of industrial equipment, including equipment parameters and the surrounding environment, so as to obtain a comprehensive evaluation value, that is, the operating state evaluation value Ose. By establishing a fault prediction model, the curve distribution of the operating state evaluation value Ose is analyzed, and to a certain extent, it can accurately and effectively obtain whether there are potential faults or abnormalities in the corresponding industrial equipment, and timely give solutions and maintenance strategies, reflecting the flexibility of the system design.

[0063] The operating state evaluation module builds a data analysis model based on the preprocessed parameters, generates the operating state evaluation value Ose reflecting the operating state of the corresponding industrial equipment, and uses deep learning algorithms to build a model based on the operating state evaluation values Ose at each moment, that is, historical data, to obtain a fault prediction model. Based on this fault prediction model, the predicted operating state evaluation value Ose is displayed in the form of a curve graph for users to view and monitor in real time;

[0064] Among them, the preprocessing process of the parameters is to clean each data in the parameters, and then perform dimensionless processing on each parameter to remove the units of each data to ensure that the influence of units on the calculation can be excluded during subsequent calculations;

[0065] The formula for generating the operating state evaluation value Ose is as follows:

[0066]

[0067] In the formula, Wd represents the operating temperature, Yl represents the pressure value, U represents the actual output pressure, I represents the actual output current, (UI) represents the actual power value, Hz represents the environmental index, G is a constant correction coefficient, and its specific value can be adjusted and set by the user or generated by fitting an analysis function, and the value range of G is 0 to 1. e is a mathematical constant, which is the base of the natural logarithm function. a1, a2, a3, and a4 are the influence factors of the operating temperature, pressure value, actual power value, and environmental index respectively, a3 > a1 > a2 > a4 > 0, and 1 > a1 > 0, 1 > a2 > 0, 1 > a3 > 0, 1 > a4 > 0;

[0068] It should be noted that for parameters such as operating temperature, pressure value, and actual power value generated during the operation of industrial equipment itself, they need to be corrected by a constant correction factor after comprehensive calculation to ensure the accuracy of the comprehensive data. For environmental parameters such as environmental indices that affect industrial equipment, additional calculation and processing are required, and finally, they are accumulated with the values of other comprehensive parameters to obtain an evaluation value calculated by comprehensively considering various factors, that is, the operating state evaluation value Ose;

[0069] The system calculates and processes to obtain the corresponding operating state evaluation value Ose at each moment, and then obtains a data set composed of each operating state evaluation value Ose, which is the historical data;

[0070] The process of using a deep learning algorithm to build a fault prediction model is as follows:

[0071] S101. Collect the historical data set containing the operating parameters of industrial equipment and the corresponding operating state evaluation value Ose, and perform feature extraction on the collected data, such as time series feature extraction, including any method among time window, sliding window, and moving average, to convert the time series data into a format suitable for RNN input;

[0072] S102. Based on the historical data, construct an RNN model, select variant structures such as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit), and divide the historical data set into a training set and a validation set. During this period, use the cross-validation method, input the training set into the RNN model for training, and adjust the model parameters to minimize the loss function so that the model can better fit the historical data;

[0073] S103. Input the validation set into the trained model for validation, and evaluate the performance and prediction accuracy of the model; according to the validation results, optimize the model, such as adjusting hyperparameters and network structure, to improve the generalization ability of the model, and then save the trained model for subsequent use;

[0074] S104. Input the operating state evaluation value Ose obtained by real-time calculation into the model for prediction calculation, obtain the prediction of the operating state evaluation value Ose of the industrial equipment at future moments, and display the predicted operating state evaluation value Ose in any form of chart, dashboard, or curve graph for users to view and monitor in real time.

[0075] The status determination module compares the future operating status evaluation value Ose of the corresponding industrial equipment with a preset standard threshold group based on the curve graph of the predicted operating status evaluation value Ose. If the future operating status evaluation value Ose is within the range of the standard threshold group, it indicates that the operating status of the corresponding industrial equipment is stable, and a maintenance strategy is formulated according to the current operating status evaluation value Ose. If the future operating status evaluation value Ose is outside the range of the standard threshold group, it indicates that the operating status of the corresponding industrial equipment fluctuates, that is, there are potential faults, and an operation instruction for secondary determination is triggered.

[0076] Among them, the preset standard threshold group is the first standard threshold oml1 and the second standard threshold oml2 respectively, and oml1 < oml2. The future operating status evaluation value Ose being within the range of the standard threshold group means oml1 ≤ Ose ≤ oml2, and the future operating status evaluation value Ose being outside the range of the standard threshold group means Ose < oml1 or Ose > oml2. In the curve graph, the X-axis represents different moments, the Y-axis represents specific evaluation values, the first standard threshold oml1 and the second standard threshold oml2 are two parallel horizontal baseline distributions in the curve graph respectively, and the operating status evaluation value Ose at each moment forms a corresponding curve.

[0077] The process of formulating a maintenance strategy according to the current operating status evaluation value Ose is as follows:

[0078] S201. Obtain the operating status evaluation value Ose at different moments within T time t , where t represents the number of the operating status evaluation value at different moments within T time, t = 1, 2, 3, 4, ……, n, and n is a positive integer;

[0079] S202. According to the average value of the operating status evaluation values at different moments within T time and the operating status evaluation value Ose t , calculate the operating status fluctuation coefficient Osfc, and the calculation formula is as follows:

[0080]

[0081] S203. Calculate and obtain the maintenance frequency based on the operating status fluctuation coefficient Osfc, and the calculation formula is as follows:

[0082]

[0083] Where, Jp represents the maintenance frequency, St represents the usage duration of the corresponding industrial equipment, b1 and b2 are respectively the preset proportionality coefficients of the usage duration and the operation state fluctuation coefficient of the corresponding industrial equipment, and b2 > b1 > 0, 1 > b2 > 0, 1 > b1 > 0, C is a constant correction coefficient, and its specific value can be adjusted and set by the user or generated by fitting an analysis function, and the value range of C is 0 to 1, int is the rounding function;

[0084] It should be noted that if the usage duration of the corresponding industrial equipment is longer, more frequent maintenance is required. If the operation state fluctuation coefficient Osfc is larger, the operation state of the equipment fluctuates more, and more frequent maintenance is also required. Therefore, both the usage duration of the corresponding industrial equipment and the operation state fluctuation coefficient are proportional to the maintenance frequency. Here, the maintenance strategy is formulated according to the current operation state evaluation value Ose as the initial maintenance strategy.

[0085] The maintenance plan adjustment module performs a secondary determination operation to determine whether it is a manual regulation. If so, observe the operation of the industrial equipment within the predetermined time H. If not, continue to adjust the maintenance frequency based on the initial maintenance strategy;

[0086] Among them, determining whether it is a manual regulation is detected by installing a camera. The behavior patterns of the equipment operation are recorded through the camera, including the behaviors of the equipment operators and the adjustment of parameters, and the behavior patterns of the images captured by the camera are analyzed to determine whether there are signs of manual regulation of the equipment;

[0087] Under the condition of determining that it is a manual regulation, observe the operation of the industrial equipment within the predetermined time H. If the operation state evaluation value Ose is within the range of the standard threshold group after the predetermined time H, continue to execute the initial maintenance strategy. Otherwise, continue to adjust the maintenance frequency based on the initial maintenance strategy;

[0088] The process of continuing to adjust the maintenance frequency based on the initial maintenance strategy is as follows:

[0089] S301. Extract the maximum difference between the operation state evaluation value Ose exceeding the standard threshold group within the preset time period S;

[0090] S302. Based on the maximum difference and the maintenance frequency obtained under the initial maintenance strategy, re - establish a data analysis model to generate the adjusted maintenance frequency. The formula is as follows:

[0091] Jtp = int(c1 * Jp+ln(c2 * Cs max + 1) 2 )

[0092] Where, Jtp represents the adjusted maintenance frequency, Cs maxDenote the maximum difference, where c1 and c2 are the influence factors of the maintenance adjustment frequency and the maximum difference respectively, c2 > c1 > 0, 0 < c1 < 1, 0 < c2 < 1, and int is the rounding function.

[0093] Specifically, when judging whether the industrial equipment is in a stable condition, corresponding maintenance strategies are given respectively. On the premise that the industrial equipment is stable, the initial maintenance frequency Jp is obtained based on the current operation status evaluation value Ose. On the premise that the industrial equipment fluctuates, it is further judged whether there is manual regulation. After excluding the influence brought by manual regulation, the maintenance adjustment frequency Jtp is obtained based on the initial maintenance frequency Jp and the maximum difference of the operation status evaluation values Ose exceeding the standard threshold group within the preset time period S. This realizes the precise maintenance regulation of industrial equipment in different states, and to a certain extent ensures that potential faults or abnormalities can be eliminated through maintenance before the industrial equipment fails, ensuring the stability of the operation of the industrial equipment.

[0094] Embodiment 2: Please refer to Figure 2 , based on Embodiment 1, this embodiment also provides a method for real-time prediction of industrial Internet data, including the following specific steps: S1. Obtain the parameters during the operation of the corresponding industrial equipment, and the parameters include operating temperature, pressure value, actual power value, and environmental index;

[0095] S2. Based on the preprocessed parameters, build a data analysis model to generate the operation status evaluation value Ose. Based on the operation status evaluation values Ose at each moment, use a deep learning algorithm to build a model to obtain a fault prediction model, and display the predicted operation status evaluation value Ose in the form of a curve graph;

[0096] S3. Based on the curve graph of the predicted operation status evaluation value Ose, compare the future operation status evaluation value Ose of the corresponding industrial equipment with the preset standard threshold group. If the future operation status evaluation value Ose is within the range of the standard threshold group, formulate a maintenance strategy according to the current operation status evaluation value Ose. The preset standard threshold group is respectively the first standard threshold oml1 and the second standard threshold oml2, and oml1 < oml2. If the future operation status evaluation value Ose is outside the range of the standard threshold group, trigger the operation instruction for secondary determination;

[0097] S4. Execute the operation of secondary determination to judge whether it is manual regulation. If so, observe the operation of the industrial equipment within the predetermined time H. Under the condition of determining that it is manual regulation, observe the operation of the industrial equipment within the predetermined time H. If the operation status evaluation value Ose is within the range of the standard threshold group after the predetermined time H, continue to execute the initial maintenance strategy. Otherwise, adjust the maintenance frequency on the basis of the initial maintenance strategy.

[0098] In the application, several formulas involved are calculated by taking their numerical values after dimensionless treatment. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation, and the values in the formula are set by those skilled in the art according to the actual situation.

[0099] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution.

[0100] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0101] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application.

Claims

1. Industrial Internet data real-time prediction system, characterized by: include: The information collection module obtains the parameters of the corresponding industrial equipment during operation, and the parameters include operating temperature, pressure value, actual power value, and environmental index; The operation status evaluation module builds a data analysis model based on the preprocessed parameters to generate the operation status evaluation value Ose. Based on the operation status evaluation value Ose at each moment, a deep learning algorithm is used to build a model to obtain a fault prediction model. Based on the fault prediction model, the predicted operation status evaluation value Ose is displayed in the form of a curve graph; The state determination module compares the future operation state evaluation value Ose of the corresponding industrial equipment with the preset standard threshold value group according to the curve graph of the predicted operation state evaluation value Ose, and triggers the operation instruction of secondary determination when the future operation state evaluation value Ose is outside the range of the standard threshold value group; If the future operation status evaluation value Ose is within the range of the standard threshold group, a maintenance strategy is formulated according to the current operation status evaluation value Ose, and the standard threshold groups are set to be the first standard threshold oml1 and the second standard threshold oml2, and oml1<oml2; The process of formulating maintenance strategies based on the current operating status evaluation value Ose is as follows: S201. Obtain the operating status evaluation value Ose at different times within T time. t , t represents the number of the running status evaluation value at different times within time T, t = 1, 2, 3, 4, ..., n, n is a positive integer; S202, according to the average value of the running status evaluation value at different times within T time And the operating status evaluation value Ose t , calculate the operating status fluctuation coefficient Osfc, the calculation formula is as follows: S203, calculating and obtaining the maintenance frequency according to the operating status fluctuation coefficient Osfc, the calculation formula is as follows: In the formula, Jp represents the maintenance frequency, St represents the usage time of the corresponding industrial equipment, b1 and b2 are the preset proportional coefficients of the usage time of the corresponding industrial equipment and the operating status fluctuation coefficient, respectively, and b2>b1>0, C is the constant correction coefficient, and int is the rounding function; The maintenance plan adjustment module performs secondary judgment to determine whether it is human control; Under the condition that it is determined to be human control, the operation status of the industrial equipment is observed within the predetermined time H. If the operation status evaluation value Ose is within the range of the standard threshold group after the predetermined time H, the initial maintenance strategy is continued to be executed. Otherwise, the maintenance frequency is continued to be adjusted based on the initial maintenance strategy. The process of further adjusting the maintenance frequency based on the initial maintenance strategy is as follows: S301, extracting the maximum difference between the operating status evaluation value Ose and the standard threshold value group within a preset time period S; S302. Based on the maximum difference and the maintenance frequency obtained under the initial maintenance strategy, a data analysis model is constructed for the second time to generate the maintenance adjustment frequency. The formula used is as follows: Jtp=int(c1*Jp+ln(c2*Cs max +1) 2 ) In the formula, Jtp represents the maintenance adjustment frequency, Cs max Indicates the maximum difference, c1 and c2 are the influencing factors of the maintenance adjustment frequency and the maximum difference respectively, c2>c1>0, 1>c1>0, and int is the rounding function.

2. The industrial Internet data real-time prediction system according to claim 1 is characterized in that: The operating temperature is obtained by evenly arranging a number of temperature sensors at locations on the surface of the corresponding industrial equipment in the operating area, and obtaining the average value of each temperature value at the same time, which is the required operating temperature.

3. The industrial Internet data real-time prediction system according to claim 2 is characterized in that: The environmental index represents the data of the environment in the factory where the industrial equipment is located, and the calculation method of the environmental index is as follows: In the formula, Hz represents the environmental index, Tw represents the ambient temperature value in the factory where the corresponding industrial equipment is located, Ts represents the ambient humidity value in the factory where the corresponding industrial equipment is located, α and β are preset proportional coefficients of the ambient temperature value and the ambient humidity value, respectively, and α>β>0.

4. The industrial Internet data real-time prediction system according to claim 3 is characterized in that: The preprocessing process of the parameters is: cleaning the data in the parameters, and then performing dimensionless processing on each parameter.

5. The industrial Internet data real-time prediction system according to claim 4 is characterized in that: The formula for generating the operating status evaluation value Ose is as follows: Wherein, Wd represents the operating temperature, Yl represents the pressure value, U represents the actual output pressure, I represents the actual output current, (UI) represents the actual power value, Hz represents the environmental index, G is the constant correction coefficient, e is the base of the natural logarithm function, a1, a2, a3, and a4 are the influencing factors of the operating temperature, pressure value, actual power value, and environmental index, respectively, and a3>a1>a2>a4>0.

6. The industrial Internet data real-time prediction system according to claim 5 is characterized in that: The process of modeling using deep learning algorithms to obtain a fault prediction model is as follows: S101, collect historical data sets containing industrial equipment operating parameters and corresponding operating status evaluation values ​​Ose, perform feature extraction on the collected data, and convert the time series data into a format suitable for RNN input; S102, based on the historical data set, construct an RNN model, select the LSTM structure, divide the historical data set into a training set and a validation set, use the cross-validation method, input the training set into the RNN model for training, and adjust the model parameters to minimize the loss function; S103, inputting the verification set into the trained model for verification, tuning the model according to the verification result and saving the trained model; S104, input the operation status evaluation value Ose obtained by real-time calculation into the model for prediction calculation, obtain the predicted value of the operation status evaluation value Ose corresponding to the industrial equipment at the future moment, and display the predicted operation status evaluation value Ose in the form of a curve graph.

7. A method for real-time prediction of industrial Internet data, using any system described in claims 1 to 6, characterized in that: The method comprises the following steps: S1, obtaining the parameters of the corresponding industrial equipment during operation, and the parameters include operating temperature, pressure value, actual power value, and environmental index; S2. Build a data analysis model based on the preprocessed parameters to generate an operation status evaluation value Ose. Based on the operation status evaluation value Ose at each moment, use a deep learning algorithm to build a model to obtain a fault prediction model. Based on the fault prediction model, display the predicted operation status evaluation value Ose in the form of a curve graph; S3. According to the curve chart of the predicted operating status evaluation value Ose, the future operating status evaluation value Ose of the corresponding industrial equipment is compared with the preset standard threshold value group. If the future operating status evaluation value Ose is within the range of the standard threshold value group, a maintenance strategy is formulated according to the current operating status evaluation value Ose. The standard threshold value groups are respectively the first standard threshold value oml1 and the second standard threshold value oml2, and oml1<oml2; if the future operating status evaluation value Ose is outside the range of the standard threshold value group, a secondary judgment operation instruction is triggered; The process of formulating maintenance strategies based on the current operating status evaluation value Ose is as follows: S201. Obtain the operating status evaluation value Ose at different times within T time. t , t represents the number of the running status evaluation value at different times within time T, t = 1, 2, 3, 4, ..., n, n is a positive integer; S202, according to the average value of the running status evaluation value at different times within T time And the operating status evaluation value Ose t , calculate the operating status fluctuation coefficient Osfc, the calculation formula is as follows: S203, calculating and obtaining the maintenance frequency according to the operating status fluctuation coefficient Osfc, the calculation formula is as follows: In the formula, Jp represents the maintenance frequency, St represents the usage time of the corresponding industrial equipment, b1 and b2 are the preset proportional coefficients of the usage time of the corresponding industrial equipment and the operating status fluctuation coefficient, respectively, and b2>b1>0, C is the constant correction coefficient, and int is the rounding function; S4, perform a secondary determination operation to determine whether it is human control. If so, observe the operation of the industrial equipment within a predetermined time H. If it is determined to be human control, observe the operation of the industrial equipment within a predetermined time H. If the operation status evaluation value Ose is within the range of the standard threshold value group after the predetermined time H, continue to execute the initial maintenance strategy. Otherwise, continue to adjust the maintenance frequency based on the initial maintenance strategy. The process of further adjusting the maintenance frequency based on the initial maintenance strategy is as follows: S301, extracting the maximum difference between the operating status evaluation value Ose and the standard threshold value group within a preset time period S; S302. Based on the maximum difference and the maintenance frequency obtained under the initial maintenance strategy, a data analysis model is constructed for the second time to generate the maintenance adjustment frequency. The formula used is as follows: Jtp=int(c1*Jp+ln(c2*Cs max +1) 2 ) In the formula, Jtp represents the maintenance adjustment frequency, Cs max Indicates the maximum difference, c1 and c2 are the influencing factors of the maintenance adjustment frequency and the maximum difference respectively, c2>c1>0, 1>c1>0, and int is the rounding function.

Images