Equipment data adjusting system and method based on artificial intelligence

Abstract

The invention discloses an equipment data adjustment system and method based on artificial intelligence, relates to the technical field of intelligent adjustment, and solves the technical problem that a deviation may exist between an analyzed abnormal condition and an actual condition when a single and rapid processing mode is used for processing data and performing anomaly analysis. Real-time data of equipment are collected through a data sensor; performing feature analysis according to the real-time data to obtain feature data of the equipment; analyzing the operation state of the equipment based on the feature data to obtain an analysis result; generating a notification signal according to the analysis result; analyzing the operation state of the equipment by using the adjustment prediction model to obtain prediction data; adjusting the equipment data by using the prediction data, and monitoring the equipment in real time to obtain feedback data of the equipment; optimizing the equipment data according to the feedback data; the real-time data of the equipment can be analyzed by using a proper analysis mode, and the accuracy of an analysis result is improved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent regulation and relates to a device data regulation technology, specifically an artificial intelligence-based device data regulation system and method. Background Art

[0002] In industrial production, scientific research experiments and daily life, data adjustment of various equipment is indispensable. By reasonably adjusting equipment data, precise control of equipment can be achieved, performance can be optimized, and production efficiency and quality can be improved. By accurately adjusting equipment parameters, equipment can be operated in the best condition, thereby improving its work efficiency and productivity. Adjustment can be optimized for specific performance indicators of equipment to meet production needs. Reasonable equipment data adjustment can reduce equipment wear and failure, thereby extending its service life. By real-time monitoring and adjustment of equipment data, potential problems can be discovered and handled in a timely manner. Adjustment can optimize the operating conditions of equipment, reduce unnecessary energy consumption and wear, and thus reduce maintenance costs.

[0003] The prior art (invention patent application with publication number CN118467212A) discloses a method and system for processing industrial equipment data, comprising the following steps: step S1, formulating a quick analysis and processing method during the industrial equipment data analysis and processing process; step S2, monitoring the industrial equipment data processing results, and judging and analyzing abnormal conditions; step S3, when an analysis abnormality occurs, adjusting the data analysis mode, and performing self-repair of the data analysis abnormality; step S4, outputting abnormal nodes in the quick analysis processing, and manually optimizing the quick analysis and processing method; step S5, adjusting the analysis resource configuration of each industrial equipment according to human resource allocation; the prior art uses a quick and single processing method for analysis when processing data. However, using a single and quick processing method to process the data and then analyze the abnormal conditions of the equipment may cause the analyzed abnormal conditions to deviate from the actual conditions, which is not conducive to handling the abnormal conditions of the equipment.

[0004] The present invention provides an artificial intelligence-based device data adjustment system and method to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an artificial intelligence-based equipment data adjustment system and method, which is used to solve the technical problem that the prior art uses a quick and single processing method to analyze data when processing data. However, using a single and quick processing method to process data and then analyze the abnormal conditions of the equipment may cause the analyzed abnormal conditions to deviate from the actual situation, which is not conducive to handling the abnormal conditions of the equipment.

[0006] To achieve the above-mentioned object, a first aspect of the present invention provides an artificial intelligence-based device data adjustment system, comprising: a data adjustment module, a data acquisition module, and a feedback optimization module;

[0007] Data acquisition module: used to collect real-time data of equipment through data sensors;

[0008] Data adjustment module: used to perform feature analysis based on real-time data to obtain feature data of the equipment; analyze the operating status of the equipment based on the feature data to obtain analysis results; generate notification signals based on the analysis results; analyze the operating status of the equipment using the adjustment prediction model to obtain prediction data;

[0009] Feedback optimization module: used to adjust device data using prediction data, monitor the device in real time, obtain device feedback data; and optimize device data based on the feedback data.

[0010] Preferably, the feature analysis based on real-time data includes:

[0011] Retrieve real-time data and pre-process it; real-time data includes: real-time temperature data, real-time pressure data, real-time vibration data, real-time current data, and real-time voltage data; pre-processing includes: removing missing values and eliminating outliers;

[0012] The timestamps of the sub-data of the real-time data are aligned by using interpolation analysis; a data processing table is obtained; the aligned sub-data are matched with the data processing table to obtain corresponding data analysis measures; and the sub-data of the real-time data are analyzed according to the data analysis measures to obtain feature data.

[0013] After preprocessing the real-time data, the present invention aligns the timestamps of the sub-data of the real-time data, which can accurately identify the event sequence and avoid the problem of misjudging the cause of the fault; and aligning the timestamps of each sub-data can make the calculation result more precise, reduce the phase angle deviation, and make the analysis result more accurate.

[0014] Preferably, analyzing the sub-data of the real-time data according to the data analysis measure includes:

[0015] Retrieve the real-time temperature data and real-time pressure data from the real-time data; calculate the change rate and range of the real-time temperature data and real-time pressure data within the set time period, and obtain the temperature change rate, pressure change rate, temperature range and pressure range;

[0016] Retrieve real-time vibration data from real-time data; fit the real-time vibration data within a set time period into a vibration signal using the cubic spline interpolation method; analyze the vibration signal using the fast Fourier transform to obtain the frequency domain characteristics of the vibration signal; the frequency domain characteristics include: main frequency fluctuation rate, sideband energy ratio, and harmonic distortion;

[0017] Retrieving real-time current data and real-time voltage data from real-time data; integrating the real-time current data and real-time voltage data within a set time period into analog signals; analyzing each harmonic component data in the analog signals using fast Fourier transform; calculating each harmonic content rate and total harmonic distortion rate based on each harmonic component data; wherein each harmonic component data includes: each harmonic amplitude and fundamental wave amplitude;

[0018] The temperature change rate and temperature extreme difference are marked as temperature characteristic data; the pressure change rate and pressure extreme difference are marked as pressure characteristic data; the frequency domain characteristics are marked as vibration characteristic data; the harmonic content rate and total harmonic distortion rate are marked as circuit characteristic data; and the characteristic data of the sub-data are integrated to generate characteristic data.

[0019] It should be noted that the calculation formula for total harmonic distortion is: The harmonic order is limited to 2-50; the calculation formula for each harmonic distortion rate is:

[0020] The present invention analyzes sub-data in different real-time data according to data analysis measures. Since the data types of sub-data are different, the impact on the equipment is different; therefore, analyzing according to the data type of the sub-data can make the analysis results more accurate and provide a data basis for subsequent analysis.

[0021] Preferably, the analyzing the operating status of the device based on the characteristic data includes:

[0022] Retrieving temperature characteristic data and pressure characteristic data from the characteristic data; when the temperature characteristic data and the pressure characteristic data are greater than corresponding characteristic thresholds, generating a temperature abnormality signal or a pressure abnormality signal;

[0023] Retrieving vibration characteristic data from the characteristic data; when the vibration characteristic data exceeds a corresponding vibration threshold range, generating a vibration abnormality signal;

[0024] Retrieve circuit characteristic data from the characteristic data; when the total harmonic distortion rate in the circuit characteristic data is greater than the distortion threshold, generate a circuit abnormality signal; otherwise, analyze the content rate of each harmonic; screen out the maximum value among the content rates of each harmonic, and calculate the difference between the maximum value and the content rates of the remaining harmonics; when the difference is greater than the difference threshold, generate a circuit abnormality signal.

[0025] It should be noted that the distortion threshold and difference threshold are set based on actual experience. Based on the total harmonic distortion rate limit for low-voltage equipment in the IEEE 519-2014 standard, the distortion threshold is set to 5%. Simulation verification based on historical data shows that the probability of circuit failure increases significantly when the difference exceeds 8%, so the difference is set to 8%.

[0026] The present invention analyzes the operating status of the equipment according to the analyzed characteristic data, and adopts different processing and analysis methods according to different characteristic data of the equipment, which can make the analysis results more accurate and lay the foundation for subsequent equipment data adjustment.

[0027] Preferably, the method for obtaining the characteristic threshold and the vibration threshold range includes:

[0028] Obtain historical data of the device; remove abnormal data from the historical data to obtain simulated data; the data type of the historical data is consistent with the real-time data; abnormal data in the historical data refers to real-time data under abnormal conditions of the device;

[0029] Sorting the simulation data from large to small according to the numerical value to obtain a simulation sorting table; cross-integrating the first n simulation data and the last n simulation data in the simulation sorting table to generate a simulation data group; simulating the simulation data group using device simulation software to obtain simulation results;

[0030] The simulation data group is analyzed using data analysis measures to obtain simulation feature data; the simulation is repeated according to the simulation results until the simulation boundary point is obtained; the feature threshold and vibration threshold range are set according to the boundary point; the simulation results include normal or abnormal; the simulation boundary point refers to the critical numerical point between normal and abnormal simulation results.

[0031] It should be noted that the characteristic threshold and vibration threshold range are implemented through a dynamic update mechanism. Specifically, the latest real-time data is incorporated into the historical data set every 24 hours, the simulation data group is regenerated, the boundary points are simulated and calculated, and the threshold is updated.

[0032] The present invention generates simulation data based on the historical data of the equipment, and divides the simulation data in the order of maximum value and minimum value to obtain simulation data group, uses data analysis measures to analyze the simulation data group to obtain simulation feature data, and repeatedly simulates the simulation data group until the simulation boundary point is obtained; it is possible to perform repeated experiments to simulate the operation of the equipment and find the boundary value between normal and abnormal, which is conducive to the rationality of setting the feature threshold and vibration range.

[0033] Preferably, the analyzing the operating status of the equipment using the adjustment prediction model includes:

[0034] Retrieve abnormal signals; integrate feature data corresponding to timestamps of abnormal signals to generate an adjustment prediction sequence; wherein abnormal signals include: temperature abnormal signals, pressure abnormal signals, vibration abnormal signals, and circuit abnormal signals;

[0035] Call the adjustment prediction model; input the adjustment prediction sequence into the adjustment prediction model to obtain the corresponding parameter label, and match the corresponding adjustment data and prediction results according to the parameter label; wherein the adjustment prediction model is constructed based on the artificial intelligence model; the parameter label is set to a positive integer.

[0036] It should be noted that parameter label 1 corresponds to the temperature adjustment data, label 2 corresponds to the pressure adjustment data, and label 3 corresponds to the shutdown maintenance instruction.

[0037] The present invention predicts the equipment based on the characteristic data of the timestamp corresponding to the abnormal signal to obtain prediction data; the model constructed by using the artificial intelligence model predicts the operating status of the equipment, which can quickly analyze the operating status of the equipment for a period of time in the future, which is conducive to timely adjustment of the equipment and avoid affecting the production rate of products due to equipment failure.

[0038] Preferably, the regulation prediction model is constructed based on an artificial intelligence model, including:

[0039] Acquire a standard data set; wherein the standard data set includes standard input data consistent with the content attributes of the adjustment prediction sequence and standard output data consistent with the content attributes of the parameter label;

[0040] Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence; use the test set to test the artificial intelligence model and obtain test indicators;

[0041] When the test index is greater than the index threshold, the artificial intelligence model is marked as an adjusted prediction model; otherwise, the artificial intelligence model is optimized through cross-validation and hyperparameter tuning technology until the test index of the artificial intelligence model is greater than the index threshold; among them, the artificial intelligence model includes: convolutional neural network model or long short-term memory neural network model.

[0042] It should be noted that artificial intelligence models are trained using supervised learning, unsupervised learning or reinforcement learning methods; the division ratio of the standard data set and the indicator threshold are set by expert assessment; the test indicators include: accuracy, recall rate and F1 score; when retraining the artificial intelligence model, the proportion of the standard data set needs to be re-divided.

[0043] Preferably, the adjusting the device data using the prediction data includes:

[0044] Retrieve prediction data; obtain the remaining working time of the equipment; the prediction data includes: adjustment data, predicted failure time and failure type;

[0045] The equipment data is adjusted according to the adjustment data in the predicted data, and the equipment's inspection and maintenance time is analyzed; when the predicted failure time is less than the equipment's remaining working time, the equipment is stopped and a fault detection signal is generated; otherwise, the corresponding inspection and maintenance time is set according to the predicted failure time.

[0046] The present invention adjusts the equipment data according to the adjustment data, so that the equipment can operate in the best state, and performs fault inspection and maintenance on the equipment according to the predicted failure time, which can reduce the probability of abnormalities occurring in the equipment during operation, which is conducive to ensuring the normal operation of the equipment, reducing the number of failures, and extending the service life of the equipment.

[0047] Preferably, the optimizing the device data according to the feedback data includes:

[0048] Retrieve feedback data; the feedback data includes: number of failures and percentage of downtime; mark the number of failures and percentage of downtime in the feedback data as GC and TZ respectively;

[0049] By formula Calculate the feedback coefficient of the device data; where α and β are proportional coefficients greater than 0, and DL is the dimension removal coefficient;

[0050] When the feedback coefficient is less than the feedback threshold, the device data is continuously fed back; otherwise, the device data is re-analyzed and adjusted.

[0051] It should be noted that the feedback threshold and proportional coefficient are set based on actual experience. When the precision of the equipment is high, the feedback threshold is set lower, and vice versa. When the proportion of the equipment's downtime has a greater impact on the equipment, the corresponding proportional coefficient β is set higher. When the fault has a greater impact on the equipment, the corresponding proportional coefficient α is set higher. The dimension removal coefficient is used to remove the dimension in the formula. The downtime ratio is the ratio of the equipment's downtime during working hours to the total working time.

[0052] The present invention calculates the feedback coefficient of the device based on the feedback data of the device, can evaluate the operating status of the device, and when the set device data is not suitable for the operation of the device, can provide timely feedback and re-analyze the device data, and can continuously monitor the operating status of the device, which is beneficial to ensure the normal use of the device and extend the service life of the device.

[0053] A second aspect of the present invention provides an artificial intelligence-based device data adjustment method, comprising:

[0054] Step S1: Collecting real-time data of the device through data sensors;

[0055] Step S2: Perform feature analysis based on real-time data to obtain feature data of the device; analyze the operating status of the device based on the feature data to obtain analysis results;

[0056] Step S3: Generate a notification signal based on the analysis results; use the adjustment prediction model to analyze the operating status of the equipment to obtain prediction data;

[0057] Step S4: Use the predicted data to adjust the device data, monitor the device in real time, and obtain feedback data from the device; optimize the device data based on the feedback data.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1. After preprocessing the real-time data, the present invention aligns the timestamps of the sub-data of the real-time data, which can accurately identify the event sequence and avoid the problem of misjudging the cause of the fault. In addition, aligning the timestamps of each sub-data can make the calculation result more accurate, reduce the phase angle deviation, and make the analysis result more accurate. The sub-data in different real-time data are analyzed according to data analysis measures. Since different data types of sub-data have different effects on the equipment, therefore, analyzing according to the data type of the sub-data can make the analysis result more accurate and provide a data basis for subsequent analysis. The operating status of the equipment is analyzed according to the analyzed characteristic data. Different processing and analysis methods are adopted according to different characteristic data of the equipment, which can make the analysis result more accurate and lay the foundation for subsequent equipment data adjustment. Simulation data is generated based on the historical data of the equipment, and the simulation data is divided in the order of maximum value and minimum value to obtain a simulation data group. The simulation data group is analyzed using the data analysis measures to obtain simulation characteristic data, and the simulation data group is repeatedly simulated until the simulation boundary point is obtained. Repeated experiments can be performed to simulate the operation of the equipment, and the boundary value between normal and abnormal can be found, which is conducive to the rationality of setting the characteristic threshold and vibration range.

[0060] 2. The present invention predicts the equipment according to the characteristic data of the timestamp corresponding to the abnormal signal to obtain the predicted data; the model constructed by the artificial intelligence model is used to predict the operating status of the equipment, which can quickly analyze the operating status of the equipment in the future, which is conducive to timely adjustment of the equipment to avoid the impact of equipment failure on the production rate of products; the equipment data is adjusted according to the adjustment data, so that the equipment can operate in the best state, and the equipment is repaired and maintained according to the predicted failure time, which can reduce the probability of abnormalities in the equipment during operation, which is conducive to ensuring the normal operation of the equipment, reducing the number of failures, and extending the service life of the equipment; the feedback coefficient of the equipment is calculated according to the feedback data of the equipment, so as to evaluate the operating status of the equipment. When the set equipment data is not suitable for the operation of the equipment, timely feedback can be given and the equipment data can be re-analyzed, and the operating status of the equipment can be continuously monitored, which is conducive to ensuring the normal use of the equipment and extending the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0062] Figure 1 A schematic diagram of the overall steps of the system of the present invention;

[0063] Figure 2 Schematic diagram of the data processing and analysis steps of the present invention;

[0064] Figure 3 Schematic diagram of the device state prediction and feedback optimization steps of the present invention;

[0065] Figure 4 Schematic diagram of the specific steps of the method of the present invention. DETAILED DESCRIPTION

[0066] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] See also Figure 1 , a first aspect of the present invention provides an artificial intelligence-based device data adjustment system, comprising: a data adjustment module, a data acquisition module, and a feedback optimization module;

[0068] Data acquisition module: used to collect real-time data of equipment through data sensors;

[0069] Data adjustment module: used to perform feature analysis based on real-time data to obtain feature data of the equipment; analyze the operating status of the equipment based on the feature data to obtain analysis results; generate notification signals based on the analysis results; analyze the operating status of the equipment using the adjustment prediction model to obtain prediction data;

[0070] Feedback optimization module: used to adjust device data using prediction data, monitor the device in real time, obtain device feedback data; and optimize device data based on the feedback data.

[0071] See also Figure 2 , real-time data of the equipment is collected through data sensors and preprocessed. Among them, real-time data includes: real-time temperature data, real-time pressure data, real-time vibration data, real-time current data and real-time voltage data. Preprocessing includes: removing missing values and eliminating outliers; aligning the timestamps of sub-data of real-time data by interpolation analysis; obtaining a data processing table; matching the aligned sub-data with the data processing table to obtain corresponding data analysis measures.

[0072] Retrieve the real-time temperature data and real-time pressure data from the real-time data; calculate the change rate and range of the real-time temperature data and real-time pressure data within the set time period, and obtain the temperature change rate, pressure change rate, temperature range and pressure range;

[0073] Retrieve real-time vibration data from real-time data; fit the real-time vibration data within a set time period into a vibration signal using the cubic spline interpolation method; use fast Fourier transform to analyze the vibration signal to obtain the frequency domain characteristics of the vibration signal; wherein, the frequency domain characteristics include: main frequency fluctuation rate, sideband energy ratio and harmonic distortion; retrieve real-time current data and real-time voltage data from real-time data; integrate the real-time current data and real-time voltage data within a set time period into analog signals respectively; use fast Fourier transform to analyze the harmonic component data in the analog signal; calculate the harmonic content rate and total harmonic distortion rate based on the harmonic component data; wherein, the harmonic component data include: harmonic amplitude and fundamental amplitude; mark the temperature change rate and temperature extreme difference as temperature characteristic data; mark the pressure change rate and pressure extreme difference as pressure characteristic data; mark the frequency domain characteristics as vibration characteristic data; mark the harmonic content rate and total harmonic distortion rate as circuit characteristic data; integrate the characteristic data of the sub-data to generate characteristic data.

[0074] Retrieve the temperature characteristic data and pressure characteristic data from the characteristic data; when the temperature characteristic data and the pressure characteristic data are greater than the corresponding characteristic thresholds, generate a temperature anomaly signal or a pressure anomaly signal; retrieve the vibration characteristic data from the characteristic data; when the vibration characteristic data exceeds the corresponding vibration threshold range, generate a vibration anomaly signal; retrieve the circuit characteristic data from the characteristic data; when the total harmonic distortion rate in the circuit characteristic data is greater than the distortion threshold, generate a circuit anomaly signal; otherwise, analyze the content rate of each harmonic; screen out the maximum value of each harmonic content rate, and calculate the difference between the maximum value and the remaining harmonic content rates; when the difference is greater than the difference threshold, generate a circuit anomaly signal.

[0075] Acquire historical data of the equipment; remove abnormal data in the historical data to obtain simulation data; wherein the data type of the historical data is consistent with the real-time data; abnormal data of the historical data refers to real-time data under abnormal conditions of the equipment; sort the simulation data from large to small according to the numerical value to obtain a simulation sorting table; cross-integrate the first n simulation data and the last n simulation data in the simulation sorting table to generate a simulation data group; use equipment simulation software to simulate the simulation data group to obtain simulation results; use data analysis measures to analyze the simulation data group to obtain simulation feature data; repeat the simulation according to the simulation results until the simulation boundary point is obtained; set the feature threshold and vibration threshold range according to the boundary point; wherein the simulation result includes normal or abnormal; the simulation boundary point refers to the critical numerical point between normal and abnormal simulation results.

[0076] It should be noted that the value of n is 5% of the total amount of historical data (for example, when there are 1000 groups of historical data, n = 50). If the data is unevenly distributed, a dynamic adjustment strategy is adopted, n = max (5%, 50 groups).

[0077] For example, suppose the threshold of the temperature characteristic data is analyzed. After a series of simulations, it is found that the rate of change in the temperature characteristic data is normal at 22% and abnormal at 23%. The corresponding characteristic threshold can be set to 22%. Similarly, the simulation boundary points can be screened out through a series of simulation experiments, thereby making the set threshold more consistent with the actual operation of the equipment.

[0078] It should be noted that the simulation data are sorted from large to small according to the numerical value to obtain a simulation sorting table; for example: assuming that there are 100 groups of simulation data in the simulation sorting table, the corresponding simulation data in the order of group 1, group 100, group 2, and group 99 are cross-integrated to generate simulation data groups; multiple groups of control groups are formed to facilitate simulation.

[0079] It should be noted that using simulation to analyze the characteristic threshold and vibration threshold range of characteristic data can accurately set the value of the threshold range, which is conducive to improving the accuracy of the analysis results.

[0080] See also Figure 3 , retrieve the abnormal signal; integrate the feature data of the timestamp corresponding to the abnormal signal to generate an adjustment prediction sequence; wherein, the abnormal signal includes: temperature abnormal signal, pressure abnormal signal, vibration abnormal signal and circuit abnormal signal; call the adjustment prediction model; input the adjustment prediction sequence into the adjustment prediction model, obtain the corresponding parameter label, and match the corresponding adjustment data and prediction result according to the parameter label; wherein, the adjustment prediction model is constructed based on the artificial intelligence model; the parameter label is set to a positive integer.

[0081] Retrieve prediction data; obtain the remaining working time of the equipment; wherein the prediction data includes: adjustment data, predicted failure time and failure type; adjust the equipment data according to the adjustment data in the prediction data, and analyze the equipment's inspection and maintenance time; when the predicted failure time is less than the equipment's remaining working time, stop the equipment and generate a fault detection signal; otherwise, set the corresponding inspection and maintenance time according to the predicted failure time.

[0082] Monitor the equipment in real time and obtain feedback data from the equipment; the feedback data includes: the number of failures and the proportion of downtime; the number of failures and the proportion of downtime in the feedback data are marked as GC and TZ respectively; through the formula Calculate the feedback coefficient of the device data; where α and β are proportional coefficients greater than 0, and DL is the dimension removal coefficient; when the feedback coefficient is less than the feedback threshold, continue to feed back the device data; otherwise, re-analyze and adjust the device data.

[0083] It should be noted that the greater the proportion of equipment downtime, the greater the corresponding feedback coefficient; the more times the equipment fails, the greater the corresponding feedback coefficient; therefore, the feedback coefficient is positively correlated with the number of failures and the proportion of downtime; if the adjusted equipment causes the equipment to freeze, etc., resulting in equipment failure or shutdown, it means that the adjusted equipment is not suitable for normal operation of the equipment, then the equipment data can be re-analyzed and adjusted in time, and the equipment data can be continuously optimized to make the adjusted equipment data more suitable for equipment operation.

[0084] It is worth noting that the regulation prediction model is built based on an artificial intelligence model, including:

[0085] Acquire a standard data set; wherein the standard data set includes standard input data consistent with the content attributes of the adjustment prediction sequence and standard output data consistent with the content attributes of the parameter label;

[0086] Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence; use the test set to test the artificial intelligence model and obtain test indicators;

[0087] When the test index is greater than the index threshold, the artificial intelligence model is marked as an adjusted prediction model; otherwise, the artificial intelligence model is optimized through cross-validation and hyperparameter tuning technology until the test index of the artificial intelligence model is greater than the index threshold; among them, the artificial intelligence model includes: convolutional neural network model or long short-term memory neural network model.

[0088] It should be noted that the standard dataset is divided into training set, validation set and test set according to the ratio of 7:2:1. When the data volume is less than 1000 groups, 5-fold cross validation is used;

[0089] It should be noted that artificial intelligence models are trained using supervised learning, unsupervised learning or reinforcement learning methods; the division ratio of the standard data set and the indicator threshold are set by expert assessment; the test indicators include: accuracy, recall rate and F1 score; when retraining the artificial intelligence model, the proportion of the standard data set needs to be re-divided.

[0090] See also Figure 4 A second aspect of the present invention provides an artificial intelligence-based device data adjustment method, comprising:

[0091] Step S1: Collecting real-time data of the device through data sensors;

[0092] Step S2: Perform feature analysis based on real-time data to obtain feature data of the device; analyze the operating status of the device based on the feature data to obtain analysis results;

[0093] Step S3: Generate a notification signal based on the analysis results; use the adjustment prediction model to analyze the operating status of the equipment to obtain prediction data;

[0094] Step S4: Use the predicted data to adjust the device data, monitor the device in real time, and obtain feedback data from the device; optimize the device data based on the feedback data.

[0095] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0096] The working principle of the present invention is as follows: the present invention collects real-time data of the equipment through data sensors; performs feature analysis based on the real-time data to obtain feature data of the equipment; analyzes the operating status of the equipment based on the feature data to obtain analysis results; generates a notification signal according to the analysis results; uses an adjustment prediction model to analyze the operating status of the equipment to obtain prediction data; uses the prediction data to adjust the equipment data, monitors the equipment in real time, and obtains feedback data of the equipment; and optimizes the equipment data according to the feedback data.

[0097] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An artificial intelligence-based device data adjustment system, characterized in that: include: Data conditioning module, as well as data acquisition module and feedback optimization module; Data acquisition module: used to collect real-time data of equipment through data sensors; Data conditioning module: used to perform feature analysis based on real-time data to obtain feature data of the device; Analyze the operating status of the equipment based on the characteristic data to obtain analysis results; generate a notification signal based on the analysis results; analyze the operating status of the equipment using the adjustment prediction model to obtain prediction data; Feedback optimization module: used to adjust equipment data using forecast data, monitor equipment in real time, and obtain feedback data from equipment; Optimize device data based on feedback data.

2. The artificial intelligence-based device data adjustment system according to claim 1, characterized in that: The feature analysis based on real-time data includes: Retrieve real-time data and pre-process it; real-time data includes: real-time temperature data, real-time pressure data, real-time vibration data, real-time current data, and real-time voltage data; pre-processing includes: removing missing values and eliminating outliers; The timestamps of the sub-data of the real-time data are aligned by using interpolation analysis; a data processing table is obtained; the aligned sub-data are matched with the data processing table to obtain corresponding data analysis measures; and the sub-data of the real-time data are analyzed according to the data analysis measures to obtain feature data.

3. The artificial intelligence-based device data adjustment system according to claim 2, characterized in that: Analyzing the sub-data of the real-time data according to the data analysis measures includes: Retrieve the real-time temperature data and real-time pressure data from the real-time data; calculate the change rate and range of the real-time temperature data and real-time pressure data within the set time period, and obtain the temperature change rate, pressure change rate, temperature range and pressure range; Retrieve real-time vibration data from real-time data; fit the real-time vibration data within a set time period into a vibration signal using the cubic spline interpolation method; analyze the vibration signal using the fast Fourier transform to obtain the frequency domain characteristics of the vibration signal; the frequency domain characteristics include: main frequency fluctuation rate, sideband energy ratio, and harmonic distortion; Retrieving real-time current data and real-time voltage data from real-time data; integrating the real-time current data and real-time voltage data within a set time period into analog signals; analyzing each harmonic component data in the analog signals using fast Fourier transform; calculating each harmonic content rate and total harmonic distortion rate based on each harmonic component data; wherein each harmonic component data includes: each harmonic amplitude and fundamental wave amplitude; The temperature change rate and temperature extreme difference are marked as temperature characteristic data; the pressure change rate and pressure extreme difference are marked as pressure characteristic data; the frequency domain characteristics are marked as vibration characteristic data; the harmonic content rate and total harmonic distortion rate are marked as circuit characteristic data; and the characteristic data of the sub-data are integrated to generate characteristic data.

4. The artificial intelligence-based device data adjustment system according to claim 1, characterized in that: The analyzing the operating status of the device based on the characteristic data includes: Retrieving temperature characteristic data and pressure characteristic data from the characteristic data; when the temperature characteristic data and the pressure characteristic data are greater than corresponding characteristic thresholds, generating a temperature abnormality signal or a pressure abnormality signal; Retrieving vibration characteristic data from the characteristic data; when the vibration characteristic data exceeds a corresponding vibration threshold range, generating a vibration abnormality signal; Retrieve circuit characteristic data from the characteristic data; when the total harmonic distortion rate in the circuit characteristic data is greater than the distortion threshold, generate a circuit abnormality signal; otherwise, analyze the content rate of each harmonic; screen out the maximum value among the content rates of each harmonic, and calculate the difference between the maximum value and the content rates of the remaining harmonics; when the difference is greater than the difference threshold, generate a circuit abnormality signal.

5. The artificial intelligence-based device data adjustment system according to claim 4, characterized in that: The method for obtaining the characteristic threshold and the vibration threshold range includes: Obtain historical data of the device; remove abnormal data from the historical data to obtain simulated data; the data type of the historical data is consistent with the real-time data; abnormal data in the historical data refers to real-time data under abnormal conditions of the device; Sorting the simulation data from large to small according to the numerical value to obtain a simulation sorting table; cross-integrating the first n simulation data and the last n simulation data in the simulation sorting table to generate a simulation data group; simulating the simulation data group using device simulation software to obtain simulation results; The simulation data group is analyzed using data analysis measures to obtain simulation feature data; the simulation is repeated according to the simulation results until the simulation boundary point is obtained; the feature threshold and vibration threshold range are set according to the boundary point; the simulation results include normal or abnormal; the simulation boundary point refers to the critical numerical point between normal and abnormal simulation results.

6. The artificial intelligence-based device data adjustment system according to claim 4, characterized in that: The use of the adjustment prediction model to analyze the operating status of the equipment includes: Retrieve abnormal signals; integrate feature data corresponding to timestamps of abnormal signals to generate an adjustment prediction sequence; wherein abnormal signals include: temperature abnormal signals, pressure abnormal signals, vibration abnormal signals, and circuit abnormal signals; Call the adjustment prediction model; input the adjustment prediction sequence into the adjustment prediction model to obtain the corresponding parameter label, and match the corresponding adjustment data and prediction results according to the parameter label; wherein the adjustment prediction model is constructed based on the artificial intelligence model; the parameter label is set to a positive integer.

7. The artificial intelligence-based device data adjustment system according to claim 6, characterized in that: The regulation prediction model is constructed based on an artificial intelligence model and includes: Acquire a standard data set; wherein the standard data set includes standard input data consistent with the content attributes of the adjustment prediction sequence and standard output data consistent with the content attributes of the parameter label; Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence; use the test set to test the artificial intelligence model and obtain test indicators; When the test index is greater than the index threshold, the artificial intelligence model is marked as an adjusted prediction model; otherwise, the artificial intelligence model is optimized through cross-validation and hyperparameter tuning technology until the test index of the artificial intelligence model is greater than the index threshold; among them, the artificial intelligence model includes: convolutional neural network model or long short-term memory neural network model.

8. The artificial intelligence-based device data adjustment system according to claim 6, characterized in that: The adjusting the device data by using the prediction data includes: Retrieve prediction data; obtain the remaining working time of the equipment; the prediction data includes: adjustment data, predicted failure time and failure type; The equipment data is adjusted according to the adjustment data in the predicted data, and the equipment's inspection and maintenance time is analyzed; when the predicted failure time is less than the equipment's remaining working time, the equipment is stopped and a fault detection signal is generated; otherwise, the corresponding inspection and maintenance time is set according to the predicted failure time.

9. The artificial intelligence-based device data adjustment system according to claim 1, characterized in that: The optimizing of the device data according to the feedback data includes: Retrieving feedback data; wherein the feedback data includes: the number of failures and the proportion of downtime; and calculating the feedback coefficient by mapping the number of failures and the proportion of downtime to the feedback coefficient; When the feedback coefficient is less than the feedback threshold, the device data is continuously fed back; otherwise, the device data is re-analyzed and adjusted.

10. An artificial intelligence-based device data adjustment method, applied to an artificial intelligence-based device data adjustment system according to any one of claims 1 to 9, characterized in that: include: Step S1: Collecting real-time data of the device through data sensors; Step S2: Perform feature analysis based on real-time data to obtain feature data of the device; Analyze the operating status of the equipment based on the characteristic data to obtain analysis results; Step S3: Generate a notification signal based on the analysis results; use the adjustment prediction model to analyze the operating status of the equipment to obtain prediction data; Step S4: Use the predicted data to adjust the device data, monitor the device in real time, and obtain feedback data from the device; Optimize device data based on feedback data.

Images