Equipment state data analysis method and system applied to automated electrical equipment

Abstract

The invention provides an equipment state data analysis method and system applied to automated electrical equipment, and the method comprises the steps: obtaining a historical operation data set of target equipment, and constructing a multi-level state analysis model based on the historical operation data set, the multi-level state analysis model comprises at least one first analysis level and at least one second analysis level, key operation parameters are extracted from the real-time operation data flow of the target equipment through an operation parameter screening strategy in the first analysis level, and a parameter screening result is generated according to the key operation parameters; inputting a parameter screening result into a state evaluation strategy in a second analysis hierarchy, and generating a state evaluation index of the target equipment in combination with a dynamic threshold range corresponding to the key operation parameter; and outputting a real-time operation state grade of the target equipment according to a comparison result of the state evaluation index and a preset state threshold value. According to the invention, efficient monitoring and accurate early warning of the equipment health state can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method and system for analyzing the state data of equipment applied to automated electricity. Background Art

[0002] The analysis of the state data of automated electrical equipment is one of the cores in the field of industrial operation and maintenance, aiming to identify potential faults and optimize maintenance strategies by real-time monitoring of equipment operation parameters. In the prior art, the state of the equipment is usually evaluated in a binary manner directly based on a preset static threshold. For example, after screening key parameters through fixed rules or manual experience, the selected parameters are compared with a constant threshold to determine the state level of the equipment, or a general machine learning model is used to train all parameters without discrimination to generate a state classification result. However, the static threshold cannot adapt to the dynamic change characteristics of the parameters of the equipment under different operation stages and load conditions, resulting in normal fluctuations being misjudged as abnormal states or real abnormalities being missed. Secondly, the coupling degree between the parameter screening and state evaluation links is too high, and a hierarchical processing mechanism is not established, so that redundant parameters interfere with the evaluation accuracy and the model response speed is restricted by the high-dimensional data processing pressure. Moreover, the traditional method lacks the consideration of the temporal correlation between parameters and the dynamic allocation of weights, and cannot accurately capture the key feature changes during the equipment degradation process, resulting in the maintenance decision lacking foresight and accuracy, and being prone to over-maintenance or maintenance lag problems under complex working conditions, seriously affecting the equipment operation efficiency and increasing the operation and maintenance cost. Summary of the Invention

[0003] In view of this, the present invention provides a method and system for analyzing the state data of equipment applied to automated electricity. The technical solution of the present invention is realized as follows: On the one hand, the present invention provides a method for analyzing the state data of equipment applied to automated electricity, the method comprising: obtaining a historical operation data set of a target device, the historical operation data set including historical data sequences of various equipment operation parameters and corresponding equipment state labels; constructing a multi-level state analysis model based on the historical operation data set, the multi-level state analysis model including at least one first analysis level and at least one second analysis level, wherein each first analysis level corresponds to a running parameter screening strategy, each second analysis level corresponds to a state evaluation strategy, and the second analysis level is after the first analysis level; extracting key running parameters from the real-time operation data stream of the target device through the running parameter screening strategy in the first analysis level, and generating a parameter screening result according to the key running parameters; inputting the parameter screening result into the state evaluation strategy in the second analysis level, and combining with the dynamic threshold range corresponding to the key running parameters, generating a state evaluation index of the target device; and outputting the real-time operation state level of the target device according to the comparison result between the state evaluation index and a preset state threshold.

[0004] On the other hand, the present invention provides a data analysis system, including a memory and a processor. The memory stores a computer program that can run on the processor, and when the processor executes the program, the steps in the above-mentioned method are implemented.

[0005] The method for analyzing the device status data applied to automated electricity of the present invention screens key operating parameters through the first analysis level in the multi-level status analysis model and generates a parameter screening result. Combining with the status evaluation strategy based on the dynamic threshold range in the second analysis level, it generates a status evaluation index for the target device. Finally, by comparing with the preset status threshold, it outputs the real-time operating status level, which can effectively solve the problems of low calculation efficiency and high misjudgment rate caused by poor adaptability of static thresholds and parameter redundancy in traditional device status analysis. Among them, the first analysis level dynamically extracts key operating parameters through parameter correlation screening conditions and stability indicators and assigns weights to ensure that the parameter screening result input to the second analysis level focuses on the core data that significantly affects the device status, reducing the interference of irrelevant parameters. The second analysis level dynamically adjusts the threshold boundary in combination with the historical operating data distribution interval and the real-time operating stage, so that the status evaluation index can accurately reflect the degradation trend of the device in different operating environments. Through the sequential dependence and data collaboration mechanism between the first analysis level and the second analysis level, a closed-loop logic chain from parameter screening to status evaluation is realized, improving the credibility and real-time performance of the analysis result. The comprehensive deviation score and status anomaly level generated based on the dynamic threshold range can accurately locate the potential fault risk points of the device and quantify the degree of anomaly, providing multi-dimensional data support for maintenance decision-making. Finally, through the automated output of the real-time operating status level, the manual inspection cost is significantly reduced and the fault response time is shortened. At the same time, the problems of over-maintenance or under-maintenance caused by fixed thresholds are avoided, realizing the efficient monitoring and accurate early warning of the device health status in complex industrial scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 It is a schematic flowchart of the implementation of a method for analyzing the device status data applied to automated electricity provided by an embodiment of the present invention.

[0007] Figure 2 It is a schematic diagram of the hardware entity of a data analysis system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0008] In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The terms "first / second / third" involved are only used to distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first / second / third" can be interchanged with a specific order or sequence when allowed, so that the embodiments of the present invention described herein can be implemented in an order other than that illustrated or described herein.

[0009] An embodiment of the present invention provides a method for analyzing device status data applied to automated electricity. This method can be executed by a processor of a data analysis system. Among them, the data analysis system can refer to devices with data processing capabilities such as servers, laptops, tablets, desktop computers, etc.

[0010] Figure 1 It is a schematic diagram of the implementation process of a method for analyzing device status data applied to automated electricity provided by an embodiment of the present invention, as Figure 1 shown, this method includes the following steps: Step S100: Obtain the historical operation data set of the target device. The historical operation data set includes historical data sequences of various device operation parameters and corresponding device status labels.

[0011] Device operation parameters refer to various physical quantities or indicators that can reflect the working state of the device during the operation of the device, such as voltage, current, temperature, rotational speed, etc. The historical data sequence is a data set of these device operation parameters recorded in chronological order over a period of time in the past. The device status label is used to classify and identify the operation state of the device at each time point, such as normal, abnormal, faulty, etc.

[0012] In an embodiment of the present invention, obtaining the historical operation data set of the target device is the basis for subsequent device status analysis. The device operation parameters can be collected in real time through various sensors installed on the device and stored in a database. Over time, these data form the historical operation data set. For example, for an electric motor, current sensors, temperature sensors, and rotational speed sensors can be installed to collect its current, temperature, and rotational speed data during operation, and corresponding device status labels, such as normal operation, overload operation, fault, etc., can be marked for the data at each time point.

[0013] Step S200: Construct a multi-level state analysis model based on the historical operation data set. The multi-level state analysis model includes at least one first analysis level and at least one second analysis level. Each first analysis level corresponds to an operation parameter screening strategy, and each second analysis level corresponds to a state evaluation strategy. The second analysis level is after the first analysis level.

[0014] The operation parameter screening strategy refers to a method of screening out key operation parameters that have an impact on device state analysis from a large number of device operation parameters. The state evaluation strategy is a method of evaluating and judging the current state of the device based on the screened key operation parameters.

[0015] Through the operation parameter screening strategy, the first analysis level can remove parameters with less impact on the device state, reduce the complexity of data processing, and improve the analysis efficiency. The second analysis level uses the state evaluation strategy to evaluate the device state in combination with the key operation parameters, and obtains a more accurate state evaluation result.

[0016] Exemplarily, the first analysis level can adopt the method of correlation analysis to screen key operation parameters, calculate the correlation degree between each operation parameter and the device state label, and select the parameters with higher correlation degree as key operation parameters. The second analysis level can adopt any feasible machine learning model, such as support vector machine, neural network, etc., to classify and evaluate the device state according to the key operation parameters.

[0017] Step S300: Extract key operation parameters from the real-time operation data stream of the target device through the operation parameter screening strategy in the first analysis level, and generate a parameter screening result according to the key operation parameters.

[0018] The real-time operation data stream refers to the device operation parameter data generated in real time during the current operation of the target device. The key operation parameters are the parameters screened out from the real-time operation data stream that are of great significance for device state analysis. The parameter screening result is the relevant information containing the key operation parameters, such as the real-time measurement value of the parameter, the parameter weight, etc.

[0019] In this step, by processing the real-time operation data stream through the operation parameter screening strategy, the key operation parameters can be quickly and accurately extracted. These key operation parameters can more effectively reflect the current operation state of the device and provide an important basis for subsequent state evaluation.

[0020] For example, for a transformer in a power system, its real-time operation data stream includes multiple parameters such as voltage, current, oil temperature, and winding temperature. Through the operation parameter screening strategy at the first analysis level, it is found that the voltage and oil temperature have a relatively high correlation with the state of the transformer. Therefore, the voltage and oil temperature are extracted as key operation parameters, and the parameter screening result is generated based on their real-time measurement values and relevant calculations.

[0021] As an implementation manner, step S300 may specifically include the following steps: Step S310: Intercept the operation parameter sequence within the current time window from the real-time operation data stream. The operation parameter sequence includes real-time measurement values of multiple device operation parameters.

[0022] The time window refers to a time period range for selecting data in the real-time operation data stream. The operation parameter sequence is a sequence composed of the real-time measurement values of multiple device operation parameters collected within this time window arranged in chronological order.

[0023] In practical applications, since the real-time operation data stream is continuously generated, in order to facilitate analysis and processing, it is necessary to intercept the data within a suitable time window. For example, the time window can be set to 1 minute, and the real-time measurement values of device operation parameters such as voltage, current, and temperature within the most recent 1 minute are intercepted from the real-time operation data stream to form the operation parameter sequence.

[0024] Step S320: Based on the parameter correlation screening conditions in the first analysis level, calculate the correlation coefficient between each operation parameter among the multiple device operation parameters and the device status label in the historical operation dataset.

[0025] The parameter correlation screening conditions are pre-set rules or criteria for evaluating the correlation between operation parameters and device status labels. The correlation coefficient is a numerical value used to measure the degree of association between each operation parameter and the device status label. The larger the correlation coefficient, the closer the association between the operation parameter and the device status.

[0026] In this step, by calculating the correlation coefficient, the influence degree of each operation parameter on the device status can be quantitatively evaluated. For example, methods such as Pearson correlation coefficient and mutual information can be used to calculate the correlation coefficient. For example, for an industrial robot, its operation parameters include joint angle, motor current, load torque, etc. By calculating the correlation coefficients between these parameters and the device status labels (such as normal, faulty, etc.) in the historical operation dataset, it can be determined which parameters are more important for the status evaluation of the robot.

[0027] As an implementation manner, step S320 may specifically include the following steps: Step S321: Extract the data distribution characteristics of each device operation parameter under different device status labels from the historical operation dataset. The data distribution characteristics include parameter mean, variance, and distribution shape indicators.

[0028] The parameter mean refers to the average value of all historical data of the device operation parameter under a certain device status label. Variance is a statistic that measures the degree of data dispersion and reflects the fluctuation of data relative to the mean. Distribution shape indicators are used to describe the distribution shape of data, such as skewness, kurtosis, etc.

[0029] In this step, extracting the data distribution characteristics helps to understand the variation law of each device operation parameter under different device states. For example, for the spindle speed parameter of a numerically controlled machine tool, under normal operating conditions, its mean may be stable near a set value and the variance is small; while in the fault state, the mean may change significantly and the variance will also increase. By analyzing these data distribution characteristics, it can provide an important basis for calculating the correlation coefficient in the follow-up.

[0030] Step S322: Calculate the discrimination value of each device operation parameter between different device status labels according to the data distribution characteristics. The discrimination value is determined based on the combination of the parameter mean difference and the variance ratio.

[0031] The discrimination value is an index used to measure the discrimination ability of a device operation parameter between different device status labels. The parameter mean difference reflects the gap between the average values of this parameter under different device states, and the variance ratio takes into account the degree of data dispersion. By combining the parameter mean difference and the variance ratio, a more comprehensive discrimination value can be obtained.

[0032] In actual calculation, the following formula can be used to calculate the discrimination value: Discrimination value = (Mean1 - Mean2) / (Variance1 + Variance2), where Mean1 and Mean2 are the means of this parameter under two different device status labels respectively, and Variance1 and Variance2 are the corresponding variances. For example, for the flow rate parameter of a water pump, under normal operating conditions and fault operating conditions, by calculating its discrimination value, it can be judged whether this parameter can effectively distinguish these two states.

[0033] Step S323: Obtain the parameter change trend curve of each device operation parameter in the historical operation dataset, and extract the number of peak points and the positions of trend inflection points in the parameter change trend curve.

[0034] The parameter change trend curve is a curve formed by connecting the historical data of the device operation parameter in chronological order, which reflects the change trend of this parameter over time. The peak point is the point of local maximum in the curve, and the trend inflection point is the point where the slope of the curve changes.

[0035] Extracting the number of peak points and the positions of trend inflection points helps to discover abnormal changes in the operating parameters of the device. For example, for the power parameter of a wind turbine, if the number of peak points in the parameter change trend curve suddenly increases or the position of the trend inflection point moves abnormally, it may indicate that the generator has failed or been affected by the external environment.

[0036] Step S324: Generate a correlation coefficient by combining the discrimination value, the number of peak points, and the positions of trend inflection points, where the correlation coefficient is positively correlated with the discrimination value and the number of peak points, and negatively correlated with the dispersion of the positions of trend inflection points.

[0037] The generation of the correlation coefficient comprehensively considers multiple factors such as the discrimination value, the number of peak points, and the positions of trend inflection points. The larger the discrimination value, the stronger the discrimination ability of the parameter between different device states, and the greater the impact on the device state. Therefore, the correlation coefficient is positively correlated with the discrimination value. The more the number of peak points, it may indicate that the change of the parameter is more intense, and the correlation with the device state is also closer. So the correlation coefficient is positively correlated with the number of peak points. The smaller the dispersion of the positions of trend inflection points, it indicates that the change trend of the parameter is more stable, and the correlation degree with the device state is higher. Therefore, the correlation coefficient is negatively correlated with the dispersion of the positions of trend inflection points. In actual calculation, a weighted summation method can be used to generate the correlation coefficient: correlation coefficient = a * discrimination value + b * number of peak points - c * dispersion of the positions of trend inflection points, where a, b, and c are pre-set weight coefficients.

[0038] Step S325: Normalize the correlation coefficient and dynamically match and verify the time-series change of the device state label in the historical operation dataset with the normalized correlation coefficient.

[0039] Normalization processing is to convert the correlation coefficient to a preset interval, such as [0, 1], for the convenience of comparison and analysis between different parameters. Dynamic matching verification is to compare the normalized correlation coefficient with the time-series change of the device state label in the historical operation dataset to check whether the correlation coefficient can accurately reflect the change of the device state.

[0040] In this step, common normalization methods can be used for normalization processing, such as min-max normalization, Z-score normalization, etc. Through dynamic matching verification, the accuracy and reliability of the correlation coefficient can be verified. For example, if the normalized correlation coefficient can make corresponding adjustments in a timely manner when the device state changes, it indicates that the correlation coefficient can better reflect the change of the device state and has high accuracy and reliability.

[0041] Step S330: Screen out candidate operating parameters with a correlation coefficient greater than a preset correlation threshold from multiple device operating parameters, and determine a parameter stability index based on the fluctuation range of the real-time measured values of the candidate operating parameters.

[0042] The preset correlation threshold is a critical value set in advance to determine whether the correlation degree between an operating parameter and the device state is strong enough. Candidate operating parameters refer to operating parameters with a correlation coefficient greater than the preset correlation threshold. The parameter stability index is an index used to measure the fluctuation of the real-time measured values of candidate operating parameters. The smaller the fluctuation range, the higher the parameter stability.

[0043] In this step, by screening candidate operating parameters, the amount of data for subsequent processing can be reduced, and the analysis efficiency can be improved. Determining the parameter stability index based on the fluctuation range of the real-time measured values of candidate operating parameters helps to understand the stability of these parameters and provides an important basis for subsequent parameter ranking and selection of key operating parameters. For example, for the operating parameters collected by multiple sensors in an automated production line, candidate operating parameters with a correlation coefficient greater than the preset correlation threshold are screened out, and then the fluctuation range of the real-time measured values of these candidate operating parameters is calculated to determine their parameter stability index.

[0044] Step S340: Rank the candidate operating parameters according to the parameter stability index, and select the preset number of candidate operating parameters ranked at the front as key operating parameters.

[0045] The preset number is the number of key operating parameters to be selected that is set in advance. By ranking the candidate operating parameters according to the parameter stability index, parameters with higher stability can be preferentially selected as key operating parameters.

[0046] In practical applications, parameters with higher stability can more accurately reflect the operating state of the device and reduce errors caused by parameter fluctuations. For example, for multiple monitoring parameters in a power system, after ranking the candidate operating parameters according to the parameter stability index, the top 5 candidate operating parameters are selected as key operating parameters for subsequent device state analysis.

[0047] Step S350: Generate a parameter screening result including parameter weight allocation based on the real-time measured values of the key operating parameters. The parameter weights are dynamically adjusted according to the correlation coefficient and the parameter stability index.

[0048] The parameter weight refers to the proportion of each key operating parameter in the parameter screening result, which is used to reflect the importance of the parameter for device state analysis. Dynamic adjustment means that the parameter weights are adjusted in real time according to the changes in the correlation coefficient and the parameter stability index.

[0049] In this step, by generating a parameter screening result that includes parameter weight allocation, the impacts of various key operating parameters on the device status can be more accurately considered comprehensively. The parameter weights are dynamically adjusted according to the correlation coefficient and the parameter stability index, enabling the parameter screening result to better conform to the actual operating conditions of the device. For example, for a key operating parameter in an industrial control system, the parameter weights are dynamically adjusted according to its correlation coefficient and the parameter stability index, generating a parameter screening result with different weights for subsequent status evaluation.

[0050] As an implementation manner, step S350 may specifically include the following steps: Step S351: Calculate the real-time offset according to the real-time measurement value of the key operating parameter and the historical mean value of the corresponding parameter in the historical operation dataset.

[0051] The real-time offset is the difference between the real-time measurement value of the key operating parameter and the mean value of this parameter in the historical operation dataset. It reflects the deviation degree of the current parameter value relative to the historical average level.

[0052] In this step, calculating the real-time offset helps to discover abnormal changes in the key operating parameters. For example, for the temperature parameter of an air-conditioning system, its historical mean value is 25°C, and the current real-time measurement value is 28°C, then the real-time offset is 3°C. By real-time monitoring the real-time offset, it is possible to timely discover whether the temperature parameter has abnormal fluctuations.

[0053] Step S352: Determine the real-time abnormal probability of each key operating parameter based on the real-time offset and the parameter stability index.

[0054] The real-time abnormal probability refers to the likelihood of a certain key operating parameter being abnormal at the current moment. It is comprehensively calculated based on the real-time offset and the parameter stability index.

[0055] In actual calculation, a probability model can be used to determine the real-time abnormal probability. For example, when the real-time offset is large and the parameter stability index is low, it indicates that the possibility of this parameter being abnormal is large, and the real-time abnormal probability will increase accordingly. For the speed parameter of an elevator, if the real-time offset exceeds the preset range and the parameter stability is poor, then the real-time abnormal probability of this speed parameter will increase.

[0056] Step S353: Generate a dynamic weight allocation ratio according to the real-time abnormal probability and the correlation coefficient, where the dynamic weight allocation ratio is positively correlated with the product of the real-time abnormal probability and the correlation coefficient.

[0057] The dynamic weight allocation ratio refers to the specific weight ratio that each key operating parameter occupies in the parameter screening result. It is positively correlated with the product of the real-time anomaly probability and the correlation coefficient, that is, the larger the real-time anomaly probability and the correlation coefficient, the larger the dynamic weight allocation ratio.

[0058] In this step, by generating the dynamic weight allocation ratio, the weights of each key operating parameter can be more reasonably allocated, making the parameter screening result more accurately reflect the operating state of the device. For example, for multiple key operating parameters in a sewage treatment system, the dynamic weight allocation ratio is generated according to their real-time anomaly probability and correlation coefficient, so that the parameters with a higher probability of anomaly and a higher correlation with the device state have a higher weight in the parameter screening result.

[0059] Step S354: Weightedly fuse the real-time measurement values of the key operating parameters with the dynamic weight allocation ratio to generate a parameter screening result containing weighted parameters.

[0060] Weighted fusion means multiplying the real-time measurement values of the key operating parameters by the corresponding dynamic weight allocation ratio, and then adding these products to obtain a comprehensive result. The parameter screening result containing weighted parameters can more comprehensively reflect the impact of each key operating parameter on the device state.

[0061] In practical applications, weighted fusion can highlight the role of key operating parameters with larger weights while reducing the impact of parameters with smaller weights. For example, for the key operating parameters collected by multiple sensors in a smart home system, their real-time measurement values are weightedly fused with the dynamic weight allocation ratio to generate a parameter screening result containing weighted parameters for subsequent device state evaluation.

[0062] Step S355: Generate a parameter screening code based on the weighted parameter values in the parameter screening result. The parameter screening code is used to identify the priority order of different key operating parameters in the second analysis level.

[0063] The parameter screening code is a way to encode key operating parameters. It determines the priority order of different key operating parameters in the second analysis level according to the weighted parameter values in the parameter screening result.

[0064] In this step, generating the parameter screening code helps to more efficiently process key operating parameters in the second analysis level. For example, for multiple key operating parameters in an automated warehousing system, the parameter screening code is generated according to their weighted parameter values, so that these parameters can be processed in the priority order in the second analysis level, improving the efficiency and accuracy of state evaluation.

[0065] Step S400: Input the parameter screening result into the state evaluation strategy in the second analysis level, and combine it with the dynamic threshold range corresponding to the key operating parameters to generate the state evaluation index of the target device.

[0066] The state evaluation strategy is a method used to evaluate and judge the operating state of a device based on the input parameter screening result. The dynamic threshold range refers to the reasonable value range of the key operating parameters at different device operating stages and conditions, and it will be adaptively adjusted as factors such as device operating time and environment change. The state evaluation index is a set of indexes used to comprehensively reflect the current operating state of the target device, such as the abnormal level, abnormal duration, etc.

[0067] In this step, inputting the parameter screening result into the state evaluation strategy and analyzing it in combination with the dynamic threshold range can more accurately evaluate the operating state of the target device. For example, for a wind turbine generator, input the parameter screening result into the state evaluation strategy, and at the same time combine the dynamic threshold range of key operating parameters such as generator temperature and speed to generate the state evaluation index of this wind turbine generator, such as whether there is an abnormality and the severity of the abnormality.

[0068] As an implementation manner, step S400 may specifically include the following steps: Step S410: Parse the real-time measurement value of the key operating parameter and the corresponding parameter weight distribution ratio from the parameter screening result.

[0069] In the parameter screening result, the real-time measurement value of the key operating parameter and the weight distribution ratio corresponding to each parameter are included. Parsing this information is for accurately using these data for state evaluation subsequently.

[0070] In this step, by parsing the parameter screening result, the real-time measurement value of the key operating parameter and the parameter weight distribution ratio can be extracted. For example, for the parameter screening result of a chemical production device, parse the real-time measurement values of key operating parameters such as temperature and pressure and their corresponding parameter weight distribution ratios to provide basic data for subsequent state evaluation.

[0071] Step S420: Determine the dynamic threshold range of the key operating parameter according to the parameter distribution interval under different device state labels in the historical operation dataset, and the dynamic threshold range is adaptively adjusted according to the device operation time.

[0072] The parameter distribution interval refers to the value range of the key operating parameter under different device state labels. The dynamic threshold range is determined according to the parameter distribution interval and will be adaptively adjusted as the device operation time changes to adapt to the dynamic changes of the device operating state.

[0073] In this step, by analyzing the parameter distribution intervals under different device status tags in the historical operation dataset, the reasonable value ranges of key operation parameters in different states can be determined. As the device operation time increases, the performance and operation status of the device may change, so the dynamic threshold range also needs to be adjusted accordingly. For example, for the battery parameters of an electric vehicle, according to the parameter distribution intervals under different charging states and driving states in the historical operation dataset, the dynamic threshold ranges of key operation parameters such as battery voltage and current are determined and adaptively adjusted as the battery usage time and charge-discharge cycles increase.

[0074] As an implementation manner, in step S420, to determine the dynamic threshold range of key operation parameters, it may specifically include the following steps: Step S421: Extract the historical maximum and minimum values of the key operation parameters in different device operation stages from the historical operation dataset.

[0075] The historical maximum and minimum values refer to the maximum and minimum values reached by the key operation parameters in different device operation stages in the historical operation dataset. They reflect the value range of this parameter in the historical operation process.

[0076] In this step, extracting the historical maximum and minimum values of the key operation parameters helps to understand the change range of this parameter in different operation stages. For example, for the temperature parameter of a steel smelting furnace, extracting the historical maximum and minimum values in different operation stages such as the melting stage and the refining stage from the historical operation dataset provides a reference for subsequent determination of the dynamic threshold range.

[0077] Step S422: Calculate the initial threshold range based on the historical maximum and minimum values, and determine the current operation stage based on the real-time operation time of the target device.

[0078] The initial threshold range is a preliminary threshold range calculated based on the historical maximum and minimum values, which provides a basis for subsequent dynamic correction. The current operation stage is determined according to the real-time operation time of the target device and the operation rules of the device. Different operation stages may correspond to different parameter value ranges.

[0079] In actual calculation, the initial threshold range can take the interval between the historical maximum and minimum values. For example, for the water level parameter of an industrial boiler, the initial threshold range calculated according to the historical maximum and minimum values is [water level minimum value, water level maximum value]. Then, according to the real-time operation time of the boiler, it is judged which operation stage such as the startup stage, the normal operation stage or the shutdown stage the current is in.

[0080] Step S423: Obtain the moving average and moving variance of the real-time measured values of the key operation parameters within the current operation stage.

[0081] The sliding average refers to the average value of the real-time measurement values of key operating parameters within a sliding window during the current operating phase. The sliding variance is a statistic that measures the degree of dispersion of these real-time measurement values within the sliding window.

[0082] In this step, obtaining the sliding average and sliding variance helps to understand the real-time changes of key operating parameters during the current operating phase. For example, for the feed speed parameter of a numerically controlled machine tool, by calculating its sliding average and sliding variance during the current operating phase, the stability and fluctuation of this parameter can be judged.

[0083] Step S424: Dynamically correct the initial threshold range according to the sliding average and sliding variance to generate a dynamic threshold range including a correction offset.

[0084] The correction offset refers to the amount by which the upper and lower limit values of the threshold range need to be adjusted when dynamically correcting the initial threshold range. By dynamically correcting the initial threshold range according to the sliding average and sliding variance, the dynamic threshold range can be made more in line with the current operating state of the device.

[0085] In this step, the process of dynamic correction can adjust the upper and lower limit values of the initial threshold range according to the changes in the sliding average and sliding variance. For example, if the sliding average changes significantly, it indicates that the average level of the parameter has changed, and the threshold range needs to be adjusted accordingly; if the sliding variance increases, it indicates that the fluctuation of the parameter intensifies, and the threshold range also needs to be appropriately adjusted.

[0086] As an implementation manner, step S424 may specifically include the following steps: Step S4241: Calculate the difference between the sliding average and the median of the initial threshold range as the mean offset.

[0087] The mean offset refers to the difference between the sliding average and the median of the initial threshold range, which reflects the degree of deviation of the current parameter's average level from the initial threshold range.

[0088] In this step, calculating the mean offset helps to determine the direction and magnitude of the adjustment required for the initial threshold range. For example, for the pressure parameter of a refrigeration system, calculate the difference between its sliding average and the median of the initial threshold range. If the difference is positive, it indicates that the average level of the current pressure is higher than the median of the initial threshold range, and the threshold range needs to be adjusted accordingly upwards.

[0089] Step S4242: Determine the variance correction coefficient according to the ratio of the sliding variance to the historical variance.

[0090] The variance correction coefficient is a coefficient calculated based on the ratio of the sliding variance to the historical variance. It is used to adjust the width of the threshold range to adapt to changes in the parameter fluctuation situation.

[0091] In this step, by comparing the magnitudes of the sliding variance and the historical variance, it can be determined whether there has been a change in the parameter fluctuation situation. If the sliding variance is greater than the historical variance, it indicates that the parameter fluctuation has intensified, and the variance correction coefficient needs to be increased to expand the threshold range; conversely, if the sliding variance is less than the historical variance, it indicates that the parameter fluctuation has decreased, and the variance correction coefficient needs to be decreased to narrow the threshold range.

[0092] Step S4243: Multiply the mean offset by the variance correction coefficient to generate a dynamic correction amount.

[0093] The dynamic correction amount is a value calculated based on the mean offset and the variance correction coefficient. It is used to adjust the upper and lower limit values of the initial threshold range.

[0094] In this step, multiplying the mean offset by the variance correction coefficient can comprehensively consider the deviation degree of the parameter average level and the change in the fluctuation situation to obtain a reasonable dynamic correction amount. For example, for the oil temperature parameter of a power transformer, based on the calculated mean offset and variance correction coefficient, multiplying them to obtain the dynamic correction amount, which is used to adjust the dynamic threshold range of the oil temperature.

[0095] Step S4244: Add the dynamic correction amount to the upper and lower limit values of the initial threshold range respectively to generate the corrected dynamic threshold range.

[0096] In this step, by adding the dynamic correction amount to the upper and lower limit values of the initial threshold range respectively, the corrected dynamic threshold range can be obtained. For example, for the speed parameter of an elevator, the initial threshold range is [lower speed limit, upper speed limit], and the dynamic correction amount is Δv, then the corrected dynamic threshold range is [lower speed limit + Δv, upper speed limit + Δv].

[0097] Step S4245: Perform boundary constraint processing on the corrected dynamic threshold range so that the upper limit value of the dynamic threshold range does not exceed a preset percentage of the historical maximum value, and the lower limit value is not lower than a preset percentage of the historical minimum value.

[0098] The boundary constraint processing is to ensure that the dynamic threshold range does not exceed a reasonable range. The preset percentage is a pre-set proportional value used to limit the upper and lower limit values of the dynamic threshold range.

[0099] In this step, through boundary constraint processing, the rationality and reliability of the dynamic threshold range can be ensured. For example, for the temperature parameter of an automated production line, set the preset percentage of the historical maximum value to 110% and the preset percentage of the historical minimum value to 90%. Perform boundary constraint processing on the corrected dynamic threshold range so that the upper limit value does not exceed 110% of the historical maximum value, and the lower limit value is not lower than 90% of the historical minimum value.

[0100] Step S425: Match and verify the dynamic threshold range with the historical change rate of the key operating parameter so that the change rate of the dynamic threshold range does not exceed the preset rate limit.

[0101] The historical change rate refers to the change speed of the key operating parameter during the historical operation process. The preset rate limit is a critical value preset in advance to limit the change speed of the dynamic threshold range.

[0102] In this step, by matching and verifying the dynamic threshold range with the historical change rate of the key operating parameter, the change of the dynamic threshold range can be prevented from being too drastic, ensuring the stability and reliability of the dynamic threshold range. For example, for the pressure parameter of a chemical reactor, compare the change rate of the dynamic threshold range with the historical change rate of the pressure parameter to ensure that the change rate of the dynamic threshold range does not exceed the preset rate limit.

[0103] Step S430: Calculate the deviation degree between the real-time measurement value and the upper and lower limit values of the dynamic threshold range, and generate a comprehensive deviation score in combination with the parameter weight distribution ratio.

[0104] The deviation degree refers to the absolute value of the difference between the real-time measurement value of the key operating parameter and the upper or lower limit value of the dynamic threshold range. The comprehensive deviation score is a score value comprehensively calculated based on the deviation degree and the parameter weight distribution ratio, and is used to reflect the abnormal degree of the key operating parameter.

[0105] In this step, by calculating the comprehensive deviation score, the abnormal situation of the key operating parameter can be evaluated more comprehensively. For example, for the voltage parameter of a generator, calculate the deviation degree between its real-time measurement value and the upper and lower limit values of the dynamic threshold range, and then generate a comprehensive deviation score in combination with the weight distribution ratio of this parameter. If the comprehensive deviation score is high, it indicates that the abnormal degree of this voltage parameter is large.

[0106] Step S440: Determine the status abnormal level corresponding to the key operating parameter according to the comparison result between the comprehensive deviation score and the preset deviation threshold.

[0107] The preset deviation threshold is a critical value set in advance, which is used to determine whether the abnormal degree of the key operating parameters has reached the set level. The status abnormality level is divided into different levels according to the comparison result of the comprehensive deviation score and the preset deviation threshold, such as mild abnormality, moderate abnormality, severe abnormality, etc.

[0108] In this step, by comparing the comprehensive deviation score with the preset deviation threshold, the status abnormality level corresponding to the key operating parameters can be accurately determined. For example, for the flow rate parameter of a water pump, if the comprehensive deviation score is greater than the preset deviation threshold and the deviation degree is large, it is determined that the status abnormality level corresponding to the flow rate parameter is severe abnormality.

[0109] Step S450: Integrate the status abnormality levels of all key operating parameters to generate an overall status evaluation index for the target device. The overall status evaluation index includes the distribution of abnormality levels and the duration of abnormality.

[0110] The distribution of abnormality levels refers to the distribution of the status abnormality levels corresponding to each key operating parameter, which reflects the abnormal degree of the target device in different aspects. The duration of abnormality refers to the continuous time that the target device is in an abnormal state.

[0111] In this step, by integrating the status abnormality levels of all key operating parameters, an overall status evaluation index for the target device can be obtained. For example, for a production line in an automated factory, integrate the status abnormality levels of each key operating parameter (such as temperature, pressure, speed, etc.) to generate an overall status evaluation index for the production line, including the distribution of abnormality levels (such as which parameters are mildly abnormal, which parameters are moderately abnormal, etc.) and the duration of abnormality, so as to comprehensively understand the operating status of the production line.

[0112] Step S500: Output the real-time operating status level of the target device according to the comparison result between the status evaluation index and the preset status threshold.

[0113] The preset status threshold is a set of critical values set in advance, which is used to divide the status evaluation index into different levels. The real-time operating status level is determined according to the comparison result between the status evaluation index and the preset status threshold, which reflects the current operating status of the target device, such as normal, warning, fault, etc.

[0114] In this step, by comparing the status evaluation indicators with the preset status thresholds, the real-time operation status level of the target device can be accurately output. For example, for a wind turbine in a wind farm, comparing its status evaluation indicators (such as abnormal level distribution, abnormal duration, etc.) with the preset status thresholds, if the status evaluation indicators are within the normal range, the real-time operation status level is output as normal; if it exceeds the warning threshold but does not reach the failure threshold, the real-time operation status level is output as warning; if it exceeds the failure threshold, the real-time operation status level is output as failure.

[0115] As an implementation manner, the method provided by the embodiment of the present invention may further include: Step S210: Add a third analysis level to the multi-level status analysis model, and the third analysis level is after the second analysis level.

[0116] The third analysis level is a newly added analysis level in the multi-level status analysis model, which is used to further analyze and predict the status of the device. Adding the third analysis level after the second analysis level can use the status evaluation indicators output by the second analysis level for more in-depth analysis.

[0117] In this step, adding the third analysis level can enhance the function of the multi-level status analysis model, enabling it to provide more comprehensive device status information. For example, for the server devices in a large data center, adding the third analysis level on the basis of the original first analysis level and second analysis level can predict the future status of the servers and discover potential fault hazards in advance.

[0118] Step S220: Through the status prediction strategy in the third analysis level, predict the future status evolution path of the target device according to the historical change trend of the status evaluation indicators.

[0119] The status prediction strategy is a method for predicting the future status of the target device according to the historical change trend of the status evaluation indicators. The future status evolution path refers to the process of status changes that the target device may experience in a future period of time.

[0120] In this step, through the status prediction strategy, the future status of the target device can be predicted using historical data, providing a reference for the maintenance and management of the device. For example, for a structural health monitoring system of a bridge, through the status prediction strategy in the third analysis level, according to the historical change trend of the bridge status evaluation indicators, predict the structural safety status evolution path of the bridge in a future period of time and take corresponding maintenance measures in advance.

[0121] As an implementation manner, step S220 may specifically include the following steps: Step S221: Extract time series data from the state evaluation metrics. The time series data includes the state evaluation metric values of the target device at different time points and the real-time measurement values of the corresponding key operating parameters.

[0122] Time series data is a set of data arranged in chronological order, which reflects the changes in the state evaluation metric values and the real-time measurement values of the key operating parameters of the target device over time.

[0123] In this step, the purpose of extracting time series data is to be able to use this data for state prediction in the follow-up. For example, for the state evaluation metrics of an aeroengine, extract the state evaluation metric values at different time points in the past period and the real-time measurement values of the corresponding key operating parameters (such as temperature, pressure, etc.) to form time series data.

[0124] Step S222: Perform multi-scale decomposition on the time series data through the state prediction strategy in the third analysis level to generate the decomposed time series data including the long-term trend component, the periodic fluctuation component, and the short-term noise component.

[0125] Multi-scale decomposition is a method of decomposing time series data into different scale components. The long-term trend component reflects the long-term change trend of the time series data, the periodic fluctuation component reflects the periodic change characteristics of the data, and the short-term noise component is the random fluctuation part in the data.

[0126] In this step, different characteristics in the time series data can be separated through multi-scale decomposition, which is convenient for subsequent analysis and processing of different components. For example, for the load forecasting problem of a power system, perform multi-scale decomposition on the load time series data to obtain the long-term trend component, the periodic fluctuation component, and the short-term noise component, and analyze and predict these components respectively, which can improve the forecasting accuracy.

[0127] Step S223: Identify the overall degradation direction of the target device based on the long-term trend component, and extract the operation state change period of the target device in combination with the periodic fluctuation component.

[0128] The overall degradation direction refers to the direction in which the performance of the target device gradually deteriorates during long-term operation. The operation state change period refers to the time interval during which the operation state of the target device changes periodically according to a rule.

[0129] In this step, by analyzing the long-term trend component, the overall degradation situation of the target device can be understood, and by combining the periodic fluctuation component, the change period of the device operation state can be extracted. For example, for the vibration monitoring data of a mechanical device, by analyzing the long-term trend component, it is found that the vibration amplitude of the device gradually increases, indicating that there is an overall degradation trend of the device; by combining the periodic fluctuation component, the change period of the device vibration can be extracted to understand the periodic change law of the device operation state.

[0130] Step S224: Generate a set of prediction sub-models according to the overall degradation direction and the change period of the operation state. The set of prediction sub-models includes multiple prediction sub-models that match different degradation stages and periodic fluctuation characteristics.

[0131] The set of prediction sub-models is a group of sub-models used to predict the future state of the target device, and each sub-model is designed for different degradation stages and periodic fluctuation characteristics.

[0132] In this step, generating a set of prediction sub-models according to the overall degradation direction and the change period of the operation state can improve the accuracy and pertinence of state prediction. For example, for the problem of predicting the battery life of an electric vehicle, according to the overall degradation direction of the battery and the charge-discharge cycle fluctuation characteristics, a set of multiple prediction sub-models that match different degradation stages and periodic fluctuation characteristics are generated, such as the sub-model for the early degradation stage, the sub-model for the middle degradation stage, etc.

[0133] Step S225: Input the decomposed time series data into each prediction sub-model in the set of prediction sub-models to generate future state prediction results under different confidence levels, and perform cross-validation on the future state prediction results.

[0134] Cross-validation is a method for evaluating the accuracy and reliability of a prediction model. By dividing the data set into multiple subsets and alternately using different subsets for training and validation.

[0135] In this step, inputting the decomposed time series data into each prediction sub-model to generate future state prediction results under different confidence levels and performing cross-validation can screen out the most accurate and reliable prediction results. For example, for the problem of predicting the quality of an industrial production process, input the decomposed time series data into each prediction sub-model in the set of prediction sub-models to generate product quality prediction results under different confidence levels, and then evaluate the accuracy of these results through cross-validation and select the optimal prediction result.

[0136] Step S226: According to the cross-validation results, screen out the target prediction sub-model with the smallest deviation from the value of the latest state evaluation index in the real-time operation data stream of the target device.

[0137] The target prediction sub-model is the sub-model selected from the set of prediction sub-models that is most suitable for predicting the current state of the target device, and it has the smallest deviation from the latest state evaluation metric value in the real-time operation data stream of the target device.

[0138] In this step, by screening the target prediction sub-model, the accuracy of state prediction can be improved. For example, for the performance prediction problem of a communication base station, according to the cross-validation results, the target prediction sub-model with the smallest deviation from the latest state evaluation metric value in the real-time operation data stream of the base station is selected for subsequent base station performance prediction.

[0139] Step S227: Use the target prediction sub-model to perform trend correction on the short-term noise component in the time series data to generate a future state evolution path that includes the impact of the corrected noise.

[0140] Trend correction refers to processing the short-term noise component to remove the random fluctuations in it and make it more conform to the overall change trend. The future state evolution path that includes the impact of the corrected noise is the change path of the future state of the target device obtained after considering the corrected impact of the short-term noise component.

[0141] In this step, by performing trend correction on the short-term noise component, the future state evolution path can be made smoother and more accurate. For example, for a weather forecasting problem, use the target prediction sub-model to perform trend correction on the short-term noise component in the weather data to generate a future weather change path that includes the impact of the corrected noise, improving the accuracy of weather forecasting.

[0142] Step S228: Correlate and match the future state evolution path with the dynamic threshold range of the key operating parameters in the real-time operation data stream to identify the abnormal prediction intervals in the future state evolution path that exceed the dynamic threshold range.

[0143] Correlate and match refers to comparing the future state evolution path with the dynamic threshold range of the key operating parameters to find the part of the future state evolution path that exceeds the dynamic threshold range. The abnormal prediction interval refers to the time period in the future state evolution path that exceeds the dynamic threshold range.

[0144] In this step, through correlation and matching, the time period when the target device may have abnormalities can be discovered in advance, providing a basis for preventive maintenance. For example, for a chemical production device, correlate and match its future state evolution path with the dynamic threshold ranges of key operating parameters such as temperature and pressure to identify the abnormal prediction intervals in the future state evolution path where the temperature or pressure exceeds the dynamic threshold range, and take measures in advance to avoid accidents.

[0145] Step S229: Generate preventive maintenance suggestions including maintenance trigger conditions and maintenance parameter ranges based on the start time point, end time point of the anomaly prediction interval, and the corresponding critical operating parameter types.

[0146] The maintenance trigger condition refers to the condition under which equipment maintenance is required, which is usually related to the start time point of the anomaly prediction interval and the abnormal conditions of critical operating parameters. The maintenance parameter range refers to the reasonable value range of the parameters that need to be adjusted during equipment maintenance.

[0147] In this step, generating preventive maintenance suggestions based on the start time point, end time point of the anomaly prediction interval, and the corresponding critical operating parameter types can plan the equipment maintenance work in advance and improve the reliability and service life of the equipment. For example, for a power transformer, generate preventive maintenance suggestions including maintenance trigger conditions (such as performing maintenance when the temperature exceeds a certain threshold) and maintenance parameter ranges (such as the temperature range for replacing the cooling oil) according to the start time point of the anomaly prediction interval and the abnormal conditions of the corresponding temperature parameters.

[0148] Step S230: Extract the key turning points in the future state evolution path and generate preventive maintenance suggestions in combination with the real-time measurement values of critical operating parameters.

[0149] The key turning point refers to the point where the state changes relatively violently in the future state evolution path, which usually indicates an important change in the equipment state. Generating preventive maintenance suggestions in combination with the real-time measurement values of critical operating parameters can make the maintenance suggestions more in line with the actual operating conditions of the equipment.

[0150] In this step, extracting key turning points and generating preventive maintenance suggestions in combination with real-time measurement values can arrange the equipment maintenance work more accurately, avoiding unnecessary maintenance and the occurrence of failures. For example, for an aeroengine, extract the key turning points in its future state evolution path, combine the real-time measurement values of critical operating parameters such as the engine temperature and pressure, and generate targeted preventive maintenance suggestions, such as performing engine overhaul and debugging before the key turning point.

[0151] As an implementation manner, step S230 may specifically include the following steps: Step S231: Identify the interval segments in the future state evolution path where the trend change rate of the predicted state curve exceeds the preset rate threshold, and extract the start time point and end time point of the interval segment as the initial turning points.

[0152] The predicted state curve is a graphical representation of the future state evolution path, which reflects the change of the target device's future state over time. The trend change rate refers to the slope of the predicted state curve at a certain point, which indicates the speed of state change. The preset rate threshold is a pre-set critical value used to determine whether the trend change rate is too large. The initial turning points refer to the starting time point and the ending time point of the interval segment in the predicted state curve where the trend change rate exceeds the preset rate threshold.

[0153] In this step, by identifying the interval segment in the predicted state curve where the trend change rate exceeds the preset rate threshold, the part with relatively drastic state change in the future state evolution path can be found, and its starting time point and ending time point are extracted as the initial turning points. For example, for the operation state prediction curve of an elevator, identify the interval segment in the curve where the trend change rate exceeds the preset rate threshold, and extract the starting time point and the ending time point of this interval segment as the initial turning points. These turning points may indicate important changes in the elevator operation state.

[0154] Step S232: Based on the fluctuation amplitude and fluctuation direction of the predicted state curve between the initial turning points, calculate the trend correlation degree between adjacent initial turning points, and screen out the initial turning points with a trend correlation degree less than the preset correlation threshold as candidate turning points.

[0155] The trend correlation degree is an index used to measure the similarity degree of the predicted state curve between adjacent initial turning points, and it is calculated according to the fluctuation amplitude and fluctuation direction. The preset correlation threshold is a pre-set critical value used to determine whether the trend correlation degree is small enough. The candidate turning points refer to the initial turning points with a trend correlation degree less than the preset correlation threshold. The state changes between these turning points may be more independent and important.

[0156] In this step, by calculating the trend correlation degree between adjacent initial turning points and screening out the initial turning points with a trend correlation degree less than the preset correlation threshold as candidate turning points, the range of key turning points can be further narrowed. For example, for the motion state prediction curve of an industrial robot, calculate the trend correlation degree between adjacent initial turning points, and screen out the initial turning points with a trend correlation degree less than the preset correlation threshold as candidate turning points. These candidate turning points may correspond to important changes in the robot motion state.

[0157] Step S233: According to the distribution density of the candidate turning points in the future state evolution path and the time interval between adjacent candidate turning points, fuse the second derivative change characteristics of the predicted state curve, and determine the candidate turning points that meet the density condition and time interval condition as the key turning points.

[0158] The distribution density refers to the degree of concentration of candidate turning points in the future state evolution path. The time interval refers to the time distance between adjacent candidate turning points. The second derivative change feature reflects the curvature change of the predicted state curve. The density condition and the time interval condition are preset criteria for screening out the true key turning points.

[0159] In this step, by comprehensively considering the distribution density of candidate turning points, the time interval, and the second derivative change feature of the predicted state curve, the candidate turning points that meet the conditions are determined as key turning points, which can more accurately find the most important state change points in the future state evolution path. For example, for the power prediction curve of a wind turbine generator, according to the distribution density of candidate turning points and the time interval between adjacent candidate turning points, combined with the second derivative change feature of the curve, the candidate turning points that meet the density condition and the time interval condition are determined as key turning points, and these key turning points may correspond to the fault occurrence points or performance change points of the wind turbine generator.

[0160] Step S234: Obtain the real-time measurement values of key operating parameters within the target time window corresponding to the key turning points, and extract the number of anomalies and the duration of anomalies where the real-time measurement values exceed the dynamic threshold range within the target time window.

[0161] The target time window refers to a time period centered on the key turning point for obtaining the real-time measurement values of key operating parameters. The number of anomalies refers to the number of times the real-time measurement values of key operating parameters exceed the dynamic threshold range within the target time window. The duration of anomalies refers to the total duration during which the key operating parameters are in an abnormal state.

[0162] In this step, by obtaining the real-time measurement values of key operating parameters within the target time window corresponding to the key turning points and extracting the number of anomalies and the duration of anomalies, the abnormal conditions of key operating parameters near the key turning points can be understood. For example, for the temperature parameter of a chemical reactor, obtain the real-time temperature measurement values within the target time window corresponding to the key turning points, and extract the number of anomalies and the duration of anomalies where the temperature exceeds the dynamic threshold range within this time window. These information can provide a basis for judging the operating state of the reactor and taking corresponding maintenance measures.

[0163] Step S235: Determine the state anomaly level of the key operating parameters at the key turning points according to the number of anomalies and the duration of anomalies, and generate a maintenance type identifier in combination with the trend change rate of the key turning points.

[0164] The status anomaly level is divided into different levels according to the number of anomalies and the duration of anomalies, such as mild anomaly, moderate anomaly, severe anomaly, etc. The maintenance type identifier is a code used to identify the type of maintenance to be performed, and it is generated based on the status anomaly level of the key operating parameters and the trend change rate of the key turning points.

[0165] In this step, the status anomaly level is determined according to the number of anomalies and the duration of anomalies, and the maintenance type identifier is generated in combination with the trend change rate, which can provide clear guidance for subsequent maintenance work. For example, for the oil temperature parameter of a power transformer, its status anomaly level is determined to be severe anomaly according to the number of anomalies and the duration of anomalies near the key turning point, and the maintenance type identifier such as "urgent maintenance" is generated in combination with the trend change rate of the key turning point, indicating that the transformer needs to be repaired immediately.

[0166] Step S236: Match the maintenance action sequence in the preset maintenance strategy library based on the maintenance type identifier, and sort the maintenance action sequence according to the status anomaly level.

[0167] The preset maintenance strategy library is a database containing the maintenance action sequences corresponding to various maintenance types. The maintenance action sequence refers to a series of maintenance actions that need to be performed to complete a certain type of maintenance. The priority sorting is to sort the maintenance action sequence according to the status anomaly level, so that the maintenance action sequence with a higher status anomaly level has a higher priority.

[0168] In this step, matching the maintenance action sequence in the preset maintenance strategy library based on the maintenance type identifier and sorting the priority according to the status anomaly level can ensure that appropriate maintenance measures can be taken in a timely manner when the equipment has anomalies. For example, for the equipment failure of an automated production line, the corresponding maintenance action sequence such as replacing parts and adjusting parameters is matched from the maintenance strategy library according to the maintenance type identifier, and then these maintenance action sequences are sorted according to the status anomaly level, and the maintenance action sequence with a higher status anomaly level is preferentially executed.

[0169] Step S237: Generate preventive maintenance suggestions including maintenance type, maintenance time window, and the execution order of maintenance actions according to the time distribution characteristics of the key turning points and the priority sorting of the maintenance action sequence.

[0170] The time distribution characteristic refers to the distribution of the key turning points on the time axis, such as the interval time, occurrence frequency, etc. The maintenance time window refers to the appropriate time period for performing maintenance work. The execution order of maintenance actions refers to the execution order of each maintenance action in the maintenance action sequence.

[0171] In this step, according to the time distribution characteristics of key turning points and the priority ranking of maintenance action sequences, preventive maintenance suggestions are generated, which can reasonably arrange the maintenance work of equipment, improve the maintenance efficiency and the reliability of the equipment. For example, for the server equipment in a large data center, according to the time distribution characteristics of key turning points and the priority ranking of maintenance action sequences, preventive maintenance suggestions including maintenance types (such as regular inspections, fault repairs, etc.), maintenance time windows (such as weekend idle time), and the execution order of maintenance actions (check hardware first, then update software, etc.) are generated.

[0172] As an implementation manner, the method provided by the embodiments of the present invention may further include, after step S500: Step S600: Set a feedback optimization mechanism in the multi-level state analysis model. The feedback optimization mechanism iteratively updates the operation parameter screening strategy and the state evaluation strategy according to the output result of the real-time operation state level.

[0173] The feedback optimization mechanism is a mechanism for optimizing the multi-level state analysis model. It adjusts the operation parameter screening strategy and the state evaluation strategy according to the output result of the real-time operation state level to improve the accuracy and adaptability of the model.

[0174] In this step, setting the feedback optimization mechanism enables the multi-level state analysis model to continuously learn and improve, and adapt to the changes in the equipment operation state. For example, for the equipment state analysis model of a smart grid, through the feedback optimization mechanism, according to the output result of the real-time operation state level, the operation parameter screening strategy is adjusted to select more appropriate key operation parameters; the state evaluation strategy is updated to improve the accuracy of state evaluation.

[0175] Step S700: Obtain the matching error between the real-time operation state level and the actual equipment maintenance record, and calculate the model optimization weight according to the matching error.

[0176] The matching error refers to the degree of difference between the real-time operation state level and the actual equipment maintenance record. The model optimization weight is a weight value calculated according to the matching error, and is used to adjust the update degree of the operation parameter screening strategy and the state evaluation strategy.

[0177] In this step, obtaining the matching error and calculating the model optimization weight can quantify the inaccuracy degree of the model and provide a basis for the optimization of the model. For example, for the state analysis model of an industrial equipment, by comparing the real-time operation state level with the actual equipment maintenance record, the matching error is calculated, and then the model optimization weight is calculated according to the matching error. When the matching error is large, the model optimization weight is increased to strengthen the update of the operation parameter screening strategy and the state evaluation strategy.

[0178] Step S800: Based on the model optimization weights, adjust the parameter correlation screening conditions in the first analysis level and the dynamic threshold range generation rules in the second analysis level to generate rules.

[0179] The parameter correlation screening conditions are the conditions used to screen key operating parameters in the first analysis level, and the dynamic threshold range generation rules are the rules used to determine the dynamic threshold range of key operating parameters in the second analysis level. Adjusting these conditions and rules based on the model optimization weights can make the multi-level state analysis model more accurately screen key operating parameters and determine the dynamic threshold range.

[0180] In this step, adjusting the parameter correlation screening conditions and the dynamic threshold range generation rules according to the model optimization weights can gradually improve the performance of the model. For example, for a state analysis model of an elevator device, adjust the parameter correlation screening conditions in the first analysis level according to the model optimization weights, and select more relevant operating parameters as key operating parameters; adjust the dynamic threshold range generation rules in the second analysis level to make the dynamic threshold range more in line with the actual operating conditions of the elevator.

[0181] Step S900: Apply the adjusted parameter correlation screening conditions and the dynamic threshold range generation rules to the newly received real-time operation data stream to generate an updated multi-level state analysis model.

[0182] In this step, apply the adjusted parameter correlation screening conditions and the dynamic threshold range generation rules to the newly received real-time operation data stream. Through the processing and analysis of the new data, generate an updated multi-level state analysis model. For example, for a state analysis model of a sewage treatment plant device, apply the adjusted parameter correlation screening conditions and the dynamic threshold range generation rules to the newly received real-time operation data stream, re-screen the parameters such as sewage flow and water quality and determine the thresholds to generate an updated multi-level state analysis model.

[0183] Step S1000: Use the updated multi-level state analysis model to perform retrospective verification on the historical operation data set to ensure that the prediction accuracy rate of the updated multi-level state analysis model is not lower than the preset accuracy rate threshold.

[0184] Retrospective verification refers to using the updated multi-level state analysis model to analyze and predict the historical operation data set, and comparing the prediction results with the actual device state labels to verify the accuracy of the model. The preset accuracy rate threshold is a critical value set in advance to measure whether the prediction accuracy rate of the model meets the requirements.

[0185] In this step, through retrospective verification, it can be ensured that the performance of the updated multi-level state analysis model is improved. For example, for the state analysis model of an aerospace device, the updated model is used to perform retrospective verification on the historical operation data set to check whether the prediction accuracy of the model is not lower than the preset accuracy threshold. If the requirement is not met, the parameters and rules of the model are continuously adjusted until the requirement is satisfied.

[0186] As an implementation manner, the method provided by the embodiments of the present invention may further include: Step S1100: Dynamically adjust the boundary value range of the preset state threshold based on the historical distribution characteristics of the real-time operation state levels.

[0187] The historical distribution characteristics of the real-time operation state levels refer to the characteristics such as the frequencies and distribution intervals of different real-time operation state levels in the past period of time. Dynamically adjusting the boundary value range of the preset state threshold means adjusting the upper limit value and the lower limit value of the preset state threshold according to these historical distribution characteristics to make it more in line with the actual operation situation of the device.

[0188] In this step, dynamically adjusting the boundary value range of the preset state threshold based on the historical distribution characteristics of the real-time operation state levels can improve the accuracy of state evaluation. For example, for the device state analysis of a solar photovoltaic power generation system, according to the historical distribution characteristics of the real-time operation state levels, it is found that the frequency of a certain state level appears too high or too low, indicating that the boundary value range of the preset state threshold may be unreasonable. At this time, the boundary value range of the preset state threshold is dynamically adjusted to make the state evaluation more accurately reflect the actual operation state of the device.

[0189] Step S1200: Obtain multiple real-time operation state levels output by the target device within a preset time period, and count the distribution intervals of the state evaluation index values corresponding to each level.

[0190] The preset time period is a pre-set time range for obtaining the real-time operation state levels of the target device. The distribution interval of the state evaluation index values refers to the value range of the state evaluation index under each real-time operation state level.

[0191] In this step, obtaining multiple real-time operation state levels of the target device within a preset time period and counting the distribution intervals of the state evaluation index values corresponding to each level can understand the distribution rules of the state evaluation index under different state levels. For example, for the device state analysis of an intelligent building, obtaining multiple real-time operation state levels output by the building devices within one month and counting the distribution intervals of the state evaluation index values such as temperature and humidity corresponding to each level provides data support for subsequent adjustment of the preset state threshold.

[0192] Step S1300: Match the distribution interval of the state evaluation index values with the occurrence frequency of parameter anomaly events under the same device state label in the historical operation dataset to determine the sensitive threshold interval that needs to be adjusted preferentially among the preset state thresholds.

[0193] The occurrence frequency of parameter anomaly events refers to the number of times that parameters appear abnormally under the same device state label in the historical operation dataset. The sensitive threshold interval refers to the interval in the preset state thresholds that has a relatively high correlation with the occurrence frequency of parameter anomaly events, and the adjustment of these intervals may have a greater impact on the accuracy of state evaluation.

[0194] In this step, by matching the distribution interval of the state evaluation index values with the occurrence frequency of parameter anomaly events to determine the sensitive threshold interval, the preset state thresholds can be adjusted in a targeted manner. For example, for the state analysis of an industrial robot, match the distribution interval of the state evaluation index values with the occurrence frequency of parameter anomaly events under the same device state label in the historical operation dataset. It is found that the occurrence frequency of parameter anomaly events of a certain state evaluation index is relatively high in a certain interval, and this interval is the sensitive threshold interval, and the preset state threshold of this interval needs to be adjusted preferentially.

[0195] Step S1400: Generate threshold offset correction coefficients for different real-time operation state levels according to the fluctuation characteristics of the state evaluation index values within the sensitive threshold interval.

[0196] The fluctuation characteristics refer to the change amplitude, change frequency and other characteristics of the state evaluation index values within the sensitive threshold interval. The threshold offset correction coefficient is a coefficient used to adjust the boundary value of the preset state threshold. It is generated according to the fluctuation characteristics of the state evaluation index values within the sensitive threshold interval, and different real-time operation state levels may correspond to different threshold offset correction coefficients.

[0197] In this step, by generating the threshold offset correction coefficient according to the fluctuation characteristics of the state evaluation index values within the sensitive threshold interval, the preset state threshold can be adjusted more precisely. For example, for the state analysis of the equipment in a power system, generate threshold offset correction coefficients for different real-time operation state levels according to the fluctuation characteristics of the state evaluation index values such as voltage and current within the sensitive threshold interval. When the state evaluation index values fluctuate greatly, increase the threshold offset correction coefficient to expand the range of the boundary values of the preset state threshold.

[0198] Step S1500: Perform a superposition operation on the threshold offset correction coefficient and the current boundary value of the preset state threshold to generate a dynamically adjusted preset state threshold.

[0199] In this step, the threshold offset correction coefficient is superimposed on the current boundary value of the preset state threshold to obtain the dynamically adjusted preset state threshold. For example, for the analysis of the device state of a communication base station, the current boundary value of the preset state threshold is [lower threshold, upper threshold], the threshold offset correction coefficient is Δ, and the dynamically adjusted preset state threshold is [lower threshold + Δ, upper threshold + Δ].

[0200] Step S1600: Reclassify the subsequent input real-time operation state levels according to the dynamically adjusted preset state threshold, and output the updated real-time operation state levels and the corresponding threshold adjustment logs.

[0201] Reclassification means re-dividing and determining the subsequent input real-time operation state levels according to the dynamically adjusted preset state threshold. The threshold adjustment log is a log file that records the process and results of the preset state threshold adjustment, and it contains information such as the adjustment time, adjustment parameters, and adjustment amplitude.

[0202] In this step, reclassifying the real-time operation state levels according to the dynamically adjusted preset state threshold can make the state assessment more accurately reflect the actual operation state of the device. At the same time, outputting the threshold adjustment log can facilitate the subsequent review and analysis of the threshold adjustment process. For example, for the analysis of the device state of an automated production line, reclassify the subsequent input real-time operation state levels according to the dynamically adjusted preset state threshold, and output the updated real-time operation state levels and the corresponding threshold adjustment logs to promptly detect changes in the device state and evaluate the effect of the threshold adjustment.

[0203] Step S1700: Correlate and verify the threshold adjustment log with the operation and maintenance records of the target device, and filter out the abnormal threshold intervals in the dynamically adjusted preset state threshold that do not match the actual device failure events.

[0204] Correlation verification means comparing the threshold adjustment log with the operation and maintenance records of the target device to check whether the dynamically adjusted preset state threshold matches the actual device failure events. The abnormal threshold interval refers to the interval in the dynamically adjusted preset state threshold that does not match the actual device failure events, and these intervals may lead to inaccurate state assessment.

[0205] In this step, filtering out the abnormal threshold intervals through correlation verification can further optimize the preset state threshold. For example, for the analysis of the device state of a wind farm, correlate and verify the threshold adjustment log with the operation and maintenance records of the wind turbines, and find that the dynamically adjusted preset state threshold of a certain state assessment index does not match the actual device failure events within a certain interval. This interval is the abnormal threshold interval, and the threshold of this interval needs to be adjusted again.

[0206] Step S1800: Based on the adjustment time nodes of the abnormal threshold interval and the corresponding state evaluation index values, reverse-correct the calculation logic of the threshold offset correction coefficient to generate a preset state threshold after secondary correction.

[0207] The reverse-correction of the calculation logic of the threshold offset correction coefficient means adjusting and improving the method for generating the threshold offset correction coefficient according to the adjustment time nodes of the abnormal threshold interval and the corresponding state evaluation index values. The preset state threshold after secondary correction is a more accurate preset state threshold obtained after reverse-correcting the calculation logic of the threshold offset correction coefficient.

[0208] In this step, based on the adjustment time nodes of the abnormal threshold interval and the corresponding state evaluation index values, reverse-correct the calculation logic of the threshold offset correction coefficient to generate a preset state threshold after secondary correction, which can improve the accuracy and reliability of the preset state threshold. For example, for the state analysis of a chemical production device, according to the adjustment time nodes of the abnormal threshold interval and the corresponding state evaluation index values such as temperature and pressure, reverse-correct the calculation logic of the threshold offset correction coefficient to generate a preset state threshold after secondary correction, making the state evaluation more in line with the actual operating conditions of the device.

[0209] Step S1900: Perform secondary reclassification on the real-time operating state level through the preset state threshold after secondary correction, and output the finally calibrated real-time operating state level and the calibration parameter set.

[0210] The secondary reclassification means re-dividing and determining the real-time operating state level according to the preset state threshold after secondary correction. The calibration parameter set is a set containing the preset state threshold after secondary correction and other relevant calibration parameters, which records the results of the final calibration.

[0211] In this step, by outputting the finally calibrated real-time operating state level and the calibration parameter set through secondary reclassification, the accuracy of the state evaluation can be ensured to reach a high level. For example, for the server devices in a large data center, use the preset state threshold after secondary correction to perform secondary reclassification on the real-time operating state level, output the finally calibrated real-time operating state level, such as normal, warning, fault, etc., and organize the preset state threshold after secondary correction and related calibration parameters (such as adjustment coefficients, boundary values, etc.) into a calibration parameter set. This can not only provide accurate device state information for the operation and maintenance personnel of the data center, but also provide a reliable basis for subsequent device management and maintenance decisions.

[0212] As an implementation manner, the method provided by the embodiment of the present invention may further include: Step S2000: Generate a device maintenance trigger instruction sequence based on the finally calibrated real-time operation status level.

[0213] The device maintenance trigger instruction sequence is a set of instructions used to trigger device maintenance operations. Different finally calibrated real-time operation status levels correspond to different maintenance requirements, so corresponding maintenance trigger instructions can be generated based on these levels. For example, when the real-time operation status level is "normal", the generated instruction may be regular inspections; when the level is "warning", the instruction may be to conduct a preliminary inspection and parameter adjustment of the device; when the level is "fault", the instruction may be to immediately stop the machine for repair, etc. In practical applications, for a mechanical device on an automated production line, based on its finally calibrated real-time operation status level, the system will generate an instruction sequence including specific maintenance operations and execution order, such as first performing a power-off operation on the device and then troubleshooting faults.

[0214] Step S2100: Extract the duration of the real-time operation status level corresponding to each instruction in the device maintenance trigger instruction sequence and the time interval between adjacent instructions.

[0215] The duration of the real-time operation status level refers to the length of time the device is in a certain real-time operation status level, and the time interval between adjacent instructions refers to the time difference between the execution of two adjacent instructions in the device maintenance trigger instruction sequence. Extracting this information helps to reasonably arrange the time and order of maintenance work. For example, in a power system, for the maintenance trigger instruction sequence of a generator, extracting the duration of the real-time operation status level corresponding to each instruction can determine the operating stability of the generator in different states; extracting the time interval between adjacent instructions can optimize the maintenance plan and avoid over-concentration or too long intervals of maintenance work.

[0216] For the extraction of the duration of the real-time operation status level, it can be analyzed by using the historical data of the device status monitoring system. The device status monitoring system will record the real-time operation status level of the device at different time points. By sorting and counting these records, the duration of each status level can be obtained. For example, for an elevator device in an intelligent building, its status monitoring system will record the operation status of the elevator (normal, faulty, etc.) in real time. By analyzing these records, the duration of the elevator in the faulty state can be determined to evaluate the severity and impact range of the fault.

[0217] Step S2200: Based on the duration and time interval, match the execution order of maintenance actions in the preset maintenance strategy library to generate a preliminary maintenance strategy plan.

[0218] The preset maintenance strategy library is a database containing various device maintenance strategies and the order of action execution. It is classified and stored according to the type, status, and maintenance requirements of the devices. Based on the extracted duration of the real-time operating status level and the time interval between adjacent instructions, the order of maintenance action execution that matches them is searched in the preset maintenance strategy library, thereby generating a preliminary maintenance strategy plan. For example, for a wind turbine generator, according to the duration of its real-time operating status level and the time interval between adjacent instructions, the corresponding maintenance strategy is found in the preset maintenance strategy library. When the device is in a warning state and the duration is long, the maintenance action order of first inspecting the device and then debugging the components is executed to generate a preliminary maintenance strategy plan.

[0219] Step S2300: Detect conflicts between the execution time windows of each maintenance action in the preliminary maintenance strategy plan and the production schedule of the target device, and filter out the maintenance actions to be adjusted with time conflicts.

[0220] The execution time window refers to the time range during which each maintenance action can be executed, and the production schedule of the target device stipulates the production tasks and operation arrangements of the device in different time periods. Conducting conflict detection can avoid conflicts between maintenance work and production tasks, ensuring the normal operation of the device and the smooth execution of the production plan. For example, for the production line equipment in an automobile manufacturing plant, the execution time window of a certain maintenance action in the preliminary maintenance strategy plan overlaps with the production peak period of the production line. Through conflict detection, this time conflict can be discovered, and this maintenance action is marked as a maintenance action to be adjusted.

[0221] Step S2400: Reassign the execution priority of the maintenance actions according to the conflict type of the maintenance actions to be adjusted and the urgency of the real-time operating status level.

[0222] The conflict type can be divided into time conflict, resource conflict, etc., and the urgency of the real-time operating status level reflects the severity of the device problem. Reassigning the execution priority of the maintenance actions according to these factors can ensure that the maintenance work can be carried out efficiently and orderly. For example, for the substation equipment in a power system, a certain maintenance action to be adjusted needs to be rescheduled due to a time conflict, and the real-time operating status level of the device is "fault" with a high degree of urgency. Then the execution priority of this maintenance action should be increased accordingly and arranged for execution first.

[0223] Step S2500: Reorganize the execution order of the maintenance actions in the preliminary maintenance strategy plan based on the reassigned priority to generate an optimized dynamic maintenance strategy.

[0224] According to the reallocated priority of maintenance action execution, adjust and reorganize the execution order of maintenance actions in the preliminary maintenance strategy plan, so that the maintenance work can be carried out more reasonably and efficiently. For example, for the equipment of an automated warehousing system, after reallocating the execution priority of maintenance actions, arrange the maintenance actions with a high degree of urgency in advance and reasonably combine the relevant maintenance actions to generate an optimized dynamic maintenance strategy.

[0225] Step S2600: Split the optimized dynamic maintenance strategy into multiple independently executable maintenance subtasks, and assign corresponding state evaluation index monitoring conditions to each maintenance subtask.

[0226] Splitting the optimized dynamic maintenance strategy into multiple independently executable maintenance subtasks facilitates the operation and management by maintenance personnel. Assigning corresponding state evaluation index monitoring conditions to each maintenance subtask can monitor the execution situation of the maintenance subtasks and the state changes of the equipment in real time. For example, for the maintenance of a ship power system, split the optimized dynamic maintenance strategy into multiple maintenance subtasks such as engine maintenance, fuel circuit inspection, and electrical system maintenance, and assign corresponding state evaluation index monitoring conditions to each subtask, such as engine speed, oil temperature, voltage, etc.

[0227] Step S2700: Continuously monitor whether the state evaluation index monitoring conditions meet the preset maintenance task activation threshold, and trigger the execution instruction of the corresponding maintenance subtask when it is met.

[0228] The preset maintenance task activation threshold is a critical value set in advance, which is used to judge whether a certain maintenance subtask needs to be executed. Continuously monitor the state evaluation index monitoring conditions. When the index value reaches or exceeds the preset maintenance task activation threshold, trigger the execution instruction of the corresponding maintenance subtask to ensure that the maintenance work can be carried out in a timely manner. For example, for the maintenance of an air conditioning system, continuously monitor the cooling efficiency of the air conditioner. When the cooling efficiency is lower than the preset maintenance task activation threshold, trigger the execution instruction of the maintenance subtask for cleaning and debugging the air conditioner.

[0229] Step S2800: Continuously collect the data of the key operating parameter changes of the target equipment during the execution of the maintenance subtask, and generate a maintenance effect evaluation index.

[0230] During the execution of the maintenance subtasks, continuously collect the data on the changes in the key operating parameters of the target device, such as temperature, pressure, rotational speed, etc. Through the analysis and processing of this data, generate the maintenance effect evaluation indicators. These indicators can reflect the execution effect of the maintenance subtasks and the improvement of the device status. For example, for a maintenance subtask of a machine tool, during the maintenance process, continuously collect the data on the changes in the key operating parameters such as the machining accuracy and vibration frequency of the machine tool, and generate the maintenance effect evaluation indicators such as the improvement rate of machining accuracy and the reduction rate of vibration frequency.

[0231] Step S2900: According to the difference value between the maintenance effect evaluation indicator and the expected maintenance target, adjust the trigger conditions and execution order of the unexecuted maintenance subtasks.

[0232] The difference value between the maintenance effect evaluation indicator and the expected maintenance target reflects the gap between the execution effect of the maintenance subtasks and the expectation. Based on this difference value, adjust the trigger conditions and execution order of the unexecuted maintenance subtasks so that the maintenance work can better meet the actual needs of the device. For example, for the maintenance of an industrial robot, after completing some maintenance subtasks, it is found that there is a difference between the maintenance effect evaluation indicator and the expected maintenance target, such as the improvement amplitude of the robot's motion accuracy does not reach the expectation. At this time, the trigger conditions of the unexecuted maintenance subtasks can be adjusted, such as reducing the execution threshold of some maintenance subtasks, or adjusting the execution order to give priority to the maintenance subtasks that are more helpful for improving the motion accuracy.

[0233] Step S3000: When all the maintenance subtasks are executed, update the real-time operation status level based on the latest status evaluation indicator value and feedback it to the dynamic adjustment process of the preset status threshold.

[0234] When all the maintenance subtasks are executed, by collecting the latest status evaluation indicator values, such as the performance parameters and operation stability of the device, re-evaluate the real-time operation status level of the device according to these indicator values. Feed back the updated real-time operation status level to the dynamic adjustment process of the preset status threshold to further optimize the preset status threshold and improve the accuracy of status evaluation. For example, for a wind turbine generator, after completing all the maintenance subtasks, collect the status evaluation indicator values such as the power output and vibration condition of the generator set, re-evaluate its real-time operation status level, such as updating from the "warning" status to the "normal" status, and feed back this updated information to the dynamic adjustment module of the preset status threshold to adjust the relevant thresholds.

[0235] By continuously and cyclically executing the above steps, that is, updating the real-time operation status level according to the latest status evaluation index value and feeding it back to the dynamic adjustment process of the preset status threshold, the multi-level status analysis model can be continuously optimized and improved, better adapt to the changes in the operation status of the device, improve the accuracy and reliability of device status analysis, and provide more effective support for the maintenance and management of the device. For example, in the device management system of a large manufacturing enterprise, through continuous feedback and adjustment, the accuracy rate of device fault warning has been continuously improved, the maintenance cost of the device has been continuously reduced, and the production efficiency has been improved.

[0236] Figure 2 The following is a schematic diagram of the hardware entity of a data analysis system provided by an embodiment of the present invention. As Figure 2 shown, the hardware entity of the data analysis system 1000 includes: a processor 1001 and a memory 1002. Among them, the memory 1002 stores a computer program that can run on the processor 1001, and when the processor 1001 executes the program, it implements the steps in the method of any of the above embodiments.

[0237] The memory 1002 stores a computer program that can run on the processor. The memory 1002 is configured to store instructions and applications executable by the processor 1001, and can also cache data to be processed or already processed by the processor 1001 and each module in the data analysis system 1000 (for example, image data, audio data, voice communication data, and video communication data), and can be implemented by flash memory (FLASH) or random access memory (Random Access Memory, RAM).

[0238] When the processor 1001 executes the program, it implements the steps of the method for device status data analysis applied to automated electrical equipment in any of the above items. The processor 1001 generally controls the overall operation of the data analysis system 1000.

[0239] The above is only the implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention.

Claims

1. A method for analyzing the device status data applied to automated electricity, characterized in that, The method includes: Obtaining a historical operation dataset of a target device, where the historical operation dataset includes historical data sequences of multiple device operation parameters and corresponding device status labels; Constructing a multi-level status analysis model based on the historical operation dataset, where the multi-level status analysis model includes at least one first analysis level and at least one second analysis level, and each first analysis level corresponds to an operation parameter screening strategy, each second analysis level corresponds to a status evaluation strategy, and the second analysis level is after the first analysis level; Extracting key operation parameters from the real-time operation data stream of the target device through the operation parameter screening strategy in the first analysis level, and generating a parameter screening result according to the key operation parameters; Inputting the parameter screening result into the status evaluation strategy in the second analysis level, and combining with the dynamic threshold range corresponding to the key operation parameters to generate a status evaluation index of the target device; Outputting the real-time operation status level of the target device according to the comparison result between the status evaluation index and a preset status threshold.

2. The method according to claim 1, wherein The extracting key operation parameters from the real-time operation data stream of the target device through the operation parameter screening strategy in the first analysis level, and generating a parameter screening result according to the key operation parameters includes: Intercepting an operation parameter sequence within a current time window from the real-time operation data stream, where the operation parameter sequence includes real-time measurement values of multiple device operation parameters; Calculating the correlation coefficient between each operation parameter among the multiple device operation parameters and the device status label in the historical operation dataset based on the parameter correlation screening condition in the first analysis level; Screening out candidate operation parameters with a correlation coefficient greater than a preset correlation threshold from the multiple device operation parameters, and determining a parameter stability index according to the fluctuation amplitude of the real-time measurement values of the candidate operation parameters; Sorting the candidate operation parameters according to the parameter stability index, and selecting a preset number of candidate operation parameters ranked at the front as the key operation parameters; Generating a parameter screening result including parameter weight assignment based on the real-time measurement values of the key operation parameters, where the parameter weight is dynamically adjusted according to the correlation coefficient and the parameter stability index.

3. The method according to claim 2, wherein The calculating the correlation coefficient between each operation parameter among the multiple device operation parameters and the device status label in the historical operation dataset based on the parameter correlation screening condition in the first analysis level includes: Extracting the data distribution characteristics of each device operation parameter under different device status labels from the historical operation dataset, where the data distribution characteristics include parameter mean, variance, and distribution form index; Calculating the discrimination value between different device status labels of each device operation parameter according to the data distribution characteristics, where the discrimination value is determined based on the combination of parameter mean difference and variance ratio; Obtaining the parameter change trend curve of each device operation parameter in the historical operation dataset, and extracting the number of peak points and the position of trend inflection points in the parameter change trend curve; Generate the correlation coefficient by combining the discrimination value, the number of peak points, and the position of the trend inflection point, where the correlation coefficient is positively correlated with the discrimination value and the number of peak points, and negatively correlated with the dispersion of the trend inflection point position; Normalize the correlation coefficient and dynamically match and verify the normalized correlation coefficient with the temporal variation of the device status labels in the historical operation dataset.

4. The method according to claim 3, characterized in that, The generation of the parameter screening result including parameter weight assignment based on the real-time measurement values of the key operation parameters includes: Calculate the real-time offset based on the real-time measurement values of the key operation parameters and the historical mean values of the corresponding parameters in the historical operation dataset; Determine the real-time anomaly probability of each key operation parameter based on the real-time offset and the parameter stability index; Generate a dynamic weight assignment ratio according to the real-time anomaly probability and the correlation coefficient, where the dynamic weight assignment ratio is positively correlated with the product of the real-time anomaly probability and the correlation coefficient; Perform weighted fusion on the real-time measurement values of the key operation parameters and the dynamic weight assignment ratio to generate a parameter screening result including weighted parameters; Generate a parameter screening code according to the weighted parameter values in the parameter screening result, and the parameter screening code is used to identify the priority order of different key operation parameters in the second analysis level.

5. The method according to claim 1, wherein The input of the parameter screening result into the status evaluation strategy in the second analysis level, combined with the dynamic threshold range corresponding to the key operation parameters, to generate the status evaluation index of the target device includes: Parse the real-time measurement values of the key operation parameters and the corresponding parameter weight assignment ratios from the parameter screening result; Determine the dynamic threshold range of the key operation parameters according to the parameter distribution intervals under different device status labels in the historical operation dataset, and the dynamic threshold range is adaptively adjusted according to the real-time operation time of the device; Calculate the deviation degrees of the real-time measurement values from the upper and lower limit values of the dynamic threshold range, and generate a comprehensive deviation score in combination with the parameter weight assignment ratio; Determine the status anomaly level corresponding to the key operation parameter according to the comparison result between the comprehensive deviation score and the preset deviation threshold; Integrate the status anomaly levels of all key operation parameters to generate the overall status evaluation index of the target device, and the overall status evaluation index includes the anomaly level distribution and the anomaly duration.

6. The method according to claim 5, characterized in that, The determination of the dynamic threshold range of the key operation parameter includes: Extract the historical maximum and minimum values of the key operation parameter in different device operation stages from the historical operation dataset; Calculate the initial threshold range based on the historical maximum and minimum values, and determine the current operation stage based on the real-time operation time of the target device; Obtain the moving average and moving variance of the real-time measurement values of the key operation parameter within the current operation stage; Dynamically correct the initial threshold range according to the moving average and moving variance to generate a dynamic threshold range including a correction offset; Match and verify the dynamic threshold range with the historical change rate of the key operating parameters, so that the change rate of the dynamic threshold range does not exceed the preset rate limit; Among them, the dynamically correcting the initial threshold range according to the sliding average and the sliding variance to generate a dynamic threshold range including a correction offset amount includes: Calculate the difference between the sliding average and the median of the initial threshold range as the mean offset amount; Determine the variance correction coefficient according to the ratio of the sliding variance to the historical variance; Multiply the mean offset amount by the variance correction coefficient to generate a dynamic correction amount; Add the dynamic correction amount to the upper limit value and the lower limit value of the initial threshold range respectively to generate a corrected dynamic threshold range; Perform boundary constraint processing on the corrected dynamic threshold range, so that the upper limit value of the dynamic threshold range does not exceed a preset percentage of the historical maximum value, and the lower limit value is not lower than a preset percentage of the historical minimum value.

7. The method according to claim 1, characterized in that The method further includes: Add a third analysis level to the multi-level state analysis model, and the third analysis level is after the second analysis level; Through the state prediction strategy in the third analysis level, predict the future state evolution path of the target device according to the historical change trend of the state evaluation index; Extract the key turning points in the future state evolution path, and generate preventive maintenance suggestions in combination with the real-time measurement values of the key operating parameters; According to the matching result between the preventive maintenance suggestion and the real-time operating state level, adjust the response priorities of the parameter screening strategy and the state evaluation strategy in the multi-level state analysis model.

8. The method according to claim 7, wherein The predicting the future state evolution path of the target device according to the historical change trend of the state evaluation index through the state prediction strategy in the third analysis level includes: Extract time series data from the state evaluation index, and the time series data includes the state evaluation index values of the target device at different time points and the corresponding real-time measurement values of the key operating parameters; Perform multi-scale decomposition on the time series data through the state prediction strategy in the third analysis level to generate decomposed time series data including a long-term trend component, a periodic fluctuation component, and a short-term noise component; Identify the overall degradation direction of the target device based on the long-term trend component, and extract the operating state change period of the target device in combination with the periodic fluctuation component; Generate a set of prediction sub-models according to the overall degradation direction and the operating state change period, and the set of prediction sub-models includes multiple prediction sub-models matching different degradation stages and periodic fluctuation characteristics; Input the decomposed time series data into each prediction sub-model in the set of prediction sub-models to generate future state prediction results under different confidence levels, and perform cross-validation on the future state prediction results; Select the target prediction sub-model with the smallest deviation from the latest state evaluation index value in the real-time operation data stream of the target device according to the cross-validation result; The short-term noise components in the time series data are corrected for trends by the target prediction sub-model, and a future state evolution path including the corrected noise impact is generated; The future state evolution path is associated and matched with the dynamic threshold range of the key operating parameters in the real-time operation data stream, and the abnormal prediction intervals exceeding the dynamic threshold range in the future state evolution path are identified; Based on the start time point, end time point of the abnormal prediction interval and the corresponding type of key operating parameter, a preventive maintenance recommendation including maintenance trigger conditions and maintenance parameter ranges is generated.

9. The method according to claim 1, wherein The method further includes: A feedback optimization mechanism is set in the multi-level state analysis model, and the feedback optimization mechanism iteratively updates the operation parameter screening strategy and the state evaluation strategy according to the output result of the real-time operation state level; The matching error between the real-time operation state level and the actual device maintenance record is obtained, and the model optimization weight is calculated according to the matching error; Based on the model optimization weight, the parameter correlation screening conditions in the first analysis level and the dynamic threshold range generation rules in the second analysis level are adjusted; The adjusted parameter correlation screening conditions and dynamic threshold range generation rules are applied to the newly received real-time operation data stream to generate an updated multi-level state analysis model; The historical operation data set is retrospectively verified through the updated multi-level state analysis model, so that the prediction accuracy of the updated multi-level state analysis model is not lower than the preset accuracy threshold.

10. A data analysis system, comprising a memory and a processor, the memory storing a computer program that can run on the processor, characterized in that, When the processor executes the program, the steps in the method according to any one of claims 1 to 9 are implemented.

Images