Meter fault detection method and system

Abstract

The invention discloses a meter fault detection method and system, and the method comprises the steps: collecting meter data, and extracting feature data; constructing a parameter threshold library to perform anomaly recognition on the feature data, and generating an anomaly monitoring index; constructing an expert knowledge base to classify the abnormal monitoring indexes, and determining categories of the abnormal monitoring indexes, the categories including an instant fault index type and a predictive maintenance index type; based on the expert knowledge base, performing index reconstruction on abnormal monitoring indexes and / or combinations of the abnormal monitoring indexes in the predictive maintenance index types to generate fault prediction indexes, the fault prediction indexes including predicted fault types and fault prediction probabilities; and providing corresponding maintenance suggestions based on the abnormal monitoring indexes and / or fault prediction indexes of the instant fault index types. According to the method, the abnormal monitoring indexes can be distinguished, the fault type and the fault occurrence time can be predicted, the utilization efficiency of maintenance resources is improved, and the probability of non-planned shutdown of the meter and the maintenance cost are reduced.

Description

Technical Field

[0001] The present application relates to the field of meter detection technology, and in particular to a meter fault detection method and system. Background Art

[0002] With the accelerating pace of urbanization in my country and the growing demand for natural gas energy, meter anomaly detection and fault diagnosis have become crucial components of intelligent management. However, while remote meters can report anomalies, the reporting parameters are relatively limited, making them prone to misjudgments and false alarms. This wastes maintenance resources and hinders timely response to meter anomalies. Therefore, accurately analyzing meter anomalies and addressing them has become a pressing technical challenge and a key research focus for those skilled in the art. Summary of the Invention

[0003] In order to solve the deficiencies of the prior art, this application adopts the following technical solutions: In a first aspect, the present application provides a meter fault detection method, the meter fault detection method comprising the following steps: Collect multi-dimensional data from meters and extract multiple feature data reflecting the meter's operating status from the meter data through feature engineering; Constructing a parameter threshold library, performing abnormality identification on each of the characteristic data based on the parameter threshold library, and generating abnormality monitoring indicators; Building an expert knowledge base, classifying the abnormal monitoring indicators based on the expert knowledge base, and determining the categories to which the abnormal monitoring indicators belong, the categories including: immediate failure indicator type and predictive maintenance indicator type; Based on the expert knowledge base, reconstruct the abnormality monitoring indicator and / or the combination of the abnormality monitoring indicators belonging to the predictive maintenance indicator type to generate a fault prediction indicator, wherein the fault prediction indicator is configured to identify a predicted fault type and a fault prediction probability; Based on the abnormal monitoring indicator and / or the fault prediction indicator of the immediate fault indicator type, corresponding maintenance suggestions are provided.

[0004] To sum up, the present application provides a meter fault detection method, which classifies abnormal monitoring indicators based on the expert knowledge base, distinguishes between immediate fault indicator types and predictive maintenance indicator types, and reconstructs the abnormal monitoring indicators belonging to the predictive maintenance indicator type, and predicts the fault type and fault occurrence time corresponding to the abnormal monitoring indicators belonging to the predictive maintenance indicator type, thereby providing sufficient windows for planned maintenance, avoiding treating all meter abnormalities as immediate faults, improving the utilization efficiency of maintenance resources, and reducing the probability of unplanned meter shutdown and maintenance costs.

[0005] Further, the abnormal monitoring indicators are classified, including: Based on the expert knowledge base, historical abnormal monitoring indicator data with classified categories is obtained; A binary classification model is trained using the historical abnormal monitoring indicator data; The unclassified abnormal monitoring indicators are classified by the trained binary classification model to determine the category to which the abnormal monitoring indicators belong.

[0006] Further, the instrument fault detection method further includes: Based on the principle of a large language model, an abnormal combination model and a fault prediction model are constructed respectively; The association strength between the abnormal monitoring indicators in the predictive maintenance indicator type and the fault types is obtained by the abnormal combination model, and / or the association strength between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault types is obtained; Based on the association strength, the abnormal monitoring indicators in the predictive maintenance indicator type and / or the combination of the abnormal monitoring indicators are predicted by the fault prediction model to obtain the predicted fault type and the fault prediction probability.

[0007] Further, the instrument fault detection method further includes: The abnormal combination model, based on the expert classification rules, instrument fault types, abnormal performance combinations, instrument degradation processes and handling records of faults that have occurred stored in the expert knowledge base, The association strength between the abnormal monitoring indicators in the predictive maintenance indicator type and the fault types is established and output, and / or the association strength between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault types is established and output.

[0008] Further, the instrument fault detection method further includes: The fault prediction model, based on the expert classification rules, instrument fault types, abnormal performance combinations, instrument degradation processes and handling records of faults that have occurred stored in the expert knowledge base, in combination with the association strength, The abnormal monitoring indicators in the predictive maintenance indicator type and / or the combination of the abnormal monitoring indicators are predicted, and the predicted fault type and the fault prediction probability are output.

[0009] Further, the expert classification rule is generated based on expert experience and data analysis summary, and is used to provide a basis for judgment when classifying the abnormal monitoring indicators, and assist in determining whether the abnormal monitoring indicators belong to the immediate failure indicator type or the predictive maintenance indicator type.

[0010] Further, the meter failure detection method further comprises: Based on the expert knowledge base, maintenance experience about failure types, maintenance processes, and component replacement periods is obtained. According to the abnormal monitoring indicators of the immediate failure indicator type and / or the failure types of the failure prediction indicators, corresponding maintenance experience is searched in the expert knowledge base to generate the maintenance suggestions.

[0011] Further, the meter failure detection method further comprises: Based on the fault-free operation period of the meter, statistical values of meter parameters are calculated, including one or more of the following types of statistical values: mean, variance, quantile, and standard deviation of the meter parameters. Based on at least one statistical value, abnormal judgment thresholds of different kinds of meter parameters are constructed, and the abnormal judgment thresholds are stored in the parameter threshold library, which dynamically updates the abnormal judgment thresholds. The abnormal judgment thresholds are obtained from the parameter threshold library, and by comparing the feature data with the abnormal judgment thresholds, it is determined whether the feature data is abnormal, and according to the determination result, if there is an abnormality, the abnormal monitoring indicators are generated.

[0012] In a second aspect, the application also provides a meter failure detection system, which comprises: A data acquisition module configured to acquire multi-dimensional data of the meter and pre-process the acquired data; A feature engineering module configured to extract features from the acquired data to obtain feature data reflecting the running state of the meter; An abnormal monitoring module configured to include a parameter threshold library, and based on the parameter threshold library, to generate abnormal monitoring indicators by identifying abnormalities in the feature data; An expert knowledge base module configured to store meter failure types, meter degradation processes, abnormal performance combinations, expert classification rules, handling records of occurred failures, and maintenance experience; An abnormal classification module configured to classify the abnormal monitoring indicators based on the expert knowledge base module to determine the categories of the abnormal monitoring indicators, including immediate failure indicator type and predictive maintenance indicator type; a fault prediction module configured to perform index reconstruction on the abnormal monitoring indicators belonging to the predictive maintenance index type and / or the combination of abnormal monitoring indicators based on the expert knowledge base, to generate a fault prediction index configured to identify a predicted fault type and a fault prediction probability; a result output module configured to output a corresponding maintenance suggestion based on the abnormal monitoring indicators of the instant fault index type and / or the fault prediction index.

[0013] Further, the fault prediction module is further configured to: train a large language model based on the expert knowledge base module; construct an abnormal combination model and a fault prediction model based on the trained large language model, respectively; input the abnormal monitoring indicators of the predictive maintenance index type and / or the combination of abnormal monitoring indicators into the abnormal combination model to generate an association strength between the abnormal monitoring indicators and the fault type, and / or to generate an association strength between the combination of abnormal monitoring indicators and the fault type; input the abnormal monitoring indicators of the predictive maintenance index type and / or the combination of abnormal monitoring indicators, the association strength into the fault prediction model, to perform fault prediction on the abnormal monitoring indicators of the predictive maintenance index type and / or the combination of abnormal monitoring indicators, to obtain the predicted fault type and the fault prediction probability. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 a step flow chart of a meter fault detection method provided by an embodiment of the present application; Figure 2 a step flow chart of abnormal identification on feature data in a meter fault detection method provided by an embodiment of the present application; Figure 3 a step flow chart of classification on abnormal monitoring indicators in a meter fault detection method provided by an embodiment of the present application; Figure 4 a step flow chart of index reconstruction on abnormal monitoring indicators belonging to the predictive maintenance index type and / or the combination of abnormal monitoring indicators in a meter fault detection method provided by an embodiment of the present application; Figure 5 a step flow chart of generating a maintenance suggestion in a meter fault detection method provided by an embodiment of the present application; Figure 6 a structural schematic diagram of a meter fault detection system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0015] The application will be described in detail below with reference to specific embodiments shown in the drawings, but these embodiments do not limit the application, and the changes made by those of ordinary skill in the art in structure, method or function based on these embodiments are included in the protection scope of the application.

[0016] In order to solve the problems of the prior art, in a first aspect, the embodiments of the application provide a meter fault detection method, which can be applied to a gas meter, and monitors and analyzes the temperature, pressure, flow and gas consumption data of the gas meter in the running process to detect whether the meter has abnormal faults. As shown in the figure, Figure 1 The meter fault detection method comprises the following steps: Step S101, multi-dimensional data of the meter is collected, and a plurality of feature data reflecting the running state of the meter is extracted from the meter data through feature engineering.

[0017] Step S102, a parameter threshold library is constructed, and each feature data is identified based on the parameter threshold library to generate an abnormal monitoring index.

[0018] Step S103, an expert knowledge base is constructed, and the abnormal monitoring index is classified based on the expert knowledge base to determine the category of the abnormal monitoring index, including: immediate fault index type and predictive maintenance index type.

[0019] Step S104, based on the expert knowledge base, the abnormal monitoring index and / or the combination of the abnormal monitoring index belonging to the predictive maintenance index type are reconstructed to generate a fault prediction index, and the fault prediction index is configured to identify the predicted fault type and the fault prediction probability.

[0020] Step S105, based on the abnormal monitoring index of the immediate fault index type and / or the fault prediction index, a corresponding maintenance suggestion is provided.

[0021] Specifically, data collection is carried out for the gas meter, and the collected data covers the original data of the meter running (such as flow, pressure, temperature, cumulative reading, battery status, alarm code, etc.) and external data related to the meter (such as local ambient temperature, user water consumption history data, user electricity consumption history data, etc.). The collected meter data is preprocessed, including data cleaning, denoising, format conversion, time synchronization, missing value filling, etc., to generate structured preprocessed data. Further, feature extraction is performed on the preprocessed data of the meter to obtain feature data reflecting the running state of the meter, including gas consumption, temperature, pressure, flow and other meter parameters.

[0022] Set the abnormality judgment threshold of the meter parameter, store the abnormality judgment threshold in the parameter threshold library, and the parameter threshold library is configured to dynamically update the abnormality judgment threshold based on the collected meter data. Perform abnormality identification on each feature data based on the abnormality judgment threshold in the parameter threshold library, judge the size of the feature data and the abnormality judgment threshold, and generate an abnormal monitoring indicator if an abnormal phenomenon exists based on the judgment result. In order to further illustrate the abnormality monitoring indicator provided by the embodiment of the present application, an example of an abnormality monitoring indicator is provided below: Based on the collected meter data, the temperature standard deviation corresponding to each meter is calculated respectively, and according to the normal distribution In principle, the abnormality threshold is three times the standard deviation of the meter temperature. Each meter's temperature fluctuation is independently assessed. If the meter temperature is within the abnormality threshold, the temperature fluctuation is considered normal. If the temperature exceeds the abnormality threshold, the temperature fluctuation is considered abnormal, and an abnormality monitoring indicator is generated. Examples of abnormality monitoring indicators include the meter ID, abnormality description, and the time the abnormality occurred.

[0023] In step S103, an expert knowledge base is constructed based on domain expert knowledge, historical data analysis, and actual maintenance experience. This expert knowledge base stores domain knowledge related to meter fault diagnosis and prediction, and is used to classify abnormal monitoring indicators and predict meter failures. Based on the expert knowledge base, the generated abnormal monitoring indicators are classified to determine whether they are immediate failure indicators or predictive maintenance indicators. Immediate failure indicators correspond to abnormal monitoring indicators that indicate an immediate failure in the meter, requiring priority treatment for the target meter. Predictive maintenance indicators correspond to early warning signals of failures, indicating that the meter may fail in the future, providing a basis for preventive maintenance of the meter.

[0024] For abnormal monitoring indicators belonging to the predictive maintenance indicator type, in step S104, based on the expert knowledge base, the strength of association between the abnormal monitoring indicators in the predictive maintenance indicator type and the meter fault type is analyzed, and / or the strength of association between combinations of abnormal monitoring indicators in the predictive maintenance indicator type and the meter fault type is analyzed. Fault prediction is performed on the abnormal monitoring indicators in the predictive maintenance indicator type, and the abnormal monitoring indicators are converted into specific fault predictions to generate fault prediction indicators. The predicted fault type and fault prediction probability for the abnormal monitoring indicators in the predictive maintenance indicator type are obtained. For example, if the abnormal monitoring indicator in the predictive maintenance indicator type is "abnormal pressure fluctuation," the fault prediction indicator obtained through predictive analysis is "a valve sticking fault is predicted to occur within a day, with a probability of b%."

[0025] When the abnormal monitoring indicators of the instant fault indicator type are acquired, it indicates that the meter has appeared instant fault, at this time, the maintenance experience corresponding to the fault type is searched in the expert knowledge base, including maintenance process, component replacement suggestion, etc. For the fault prediction indicators, combined with the prediction fault type, prediction probability and other information, the corresponding maintenance experience and maintenance strategy are searched. Then, the information corresponding to the instant fault indicators and the fault prediction indicators is integrated, and specific maintenance suggestions are generated for different fault conditions or potential risks. The maintenance suggestions include real-time fault alarm, fault diagnosis report, predictive maintenance suggestion, maintenance work order or dispatching instruction and other information, which ensures that the operation and maintenance personnel can timely and accurately maintain the meter according to the suggestions, improves the reliability and operation efficiency of the meter, and reduces the maintenance cost and safety risk.

[0026] According to the above description, the meter fault detection method provided by the application classifies the abnormal monitoring indicators based on the expert knowledge base, distinguishes the instant fault indicator type and the predictive maintenance indicator type, and reconstructs the abnormal monitoring indicators belonging to the predictive maintenance indicator type. The fault type and fault occurrence time corresponding to the abnormal monitoring indicators belonging to the predictive maintenance indicator type are predicted, so as to provide sufficient window for planned maintenance, avoid treating all meter abnormalities as instant faults, improve the utilization efficiency of maintenance resources, and reduce the probability of non-planned shutdown of the meter and the maintenance cost.

[0027] As an optional implementation manner, in step S101, feature calculation is performed on the collected meter data, including one or more of the following type feature calculation: for the pressure data during gas use, the data with zero flow in a day is searched, the proportion of the number of pressure values less than the minimum threshold value set by the meter to the number of data is judged, the interference data during non-gas use period and non-gas use period of more than half a day is eliminated, the pressure data of the effective gas use period is screened, the daily average value of the pressure during gas use is calculated, and the daily standard deviation of the pressure during gas use is calculated to reflect the overall level and fluctuation amplitude of the pressure during daily gas use. At the same time, for the all-weather pressure data, the average value and standard deviation of the pressure are counted by day or month without distinguishing whether gas is used, which is used for monitoring the pressure stability of the meter in the whole period.

[0028] For the gas consumption and flow data, the average value and standard deviation are calculated by day, month, etc. to obtain the daily average value and daily standard deviation of the gas consumption, the daily average value and daily standard deviation of the flow, the monthly average value and monthly standard deviation of the gas consumption, and the monthly average value and monthly standard deviation of the flow. The daily average value of the gas consumption and the flow can reflect the use scale of the daily gas consumption and the flow, and the standard deviation of the gas consumption and the flow can reflect the daily fluctuation degree; the monthly dimension statistics can capture the long-term trend.

[0029] Furthermore, based on the collected meter data, the user's time-to-hour unevenness coefficient can be calculated. The time-to-hour unevenness coefficient is the ratio of hourly gas consumption to the average daily gas consumption. Specifically, the time-to-hour unevenness coefficient can be calculated using the following formula: ; By using the above formula, the unevenness coefficient can be calculated to eliminate the influence of dimension, reflect the regular changes in user gas consumption data, and facilitate the diagnosis of user gas consumption data.

[0030] As an optional implementation, such as Figure 2 As shown, in step S102, the abnormality identification of each feature data further includes the following steps: Step S201 : calculating statistical values ​​of meter parameters based on the fault-free operation period of the meter, where the statistical values ​​include one or more of the following types of statistical values: mean, variance, quantile and standard deviation of the meter parameters.

[0031] Step S202: constructing abnormality judgment thresholds for different types of meter parameters based on at least one statistical value. The abnormality judgment thresholds are stored in a parameter threshold library, and the parameter threshold library dynamically updates the abnormality judgment thresholds.

[0032] Step S203: Obtain an abnormality determination threshold from the parameter threshold library, and determine whether the feature data has an abnormality by comparing the feature data with the abnormality determination threshold. If an abnormality exists, generate an abnormality monitoring indicator based on the determination result.

[0033] Specifically, based on the meter's historical maintenance records and known fault periods, reverse deduction is performed to determine the meter's normal operating period without faults. Data for various meter parameters (such as temperature, pressure, flow rate, and gas usage) during normal operating periods for a specific meter, a group of meters of the same type, or a group of meters in the same region are analyzed. Statistics for each meter parameter are calculated. These statistics include one or more of the following: mean, variance, quantile, and standard deviation. Anomaly thresholds are set based on the statistical values ​​of each meter parameter. Different statistical values ​​can be used to construct anomaly thresholds for different meter parameters. For example, for meter temperature, the variance value can be used to construct an anomaly threshold; for meter pressure, the mean value can be used to construct an anomaly threshold. Alternatively, in an optional embodiment, three times the standard deviation can be used as the anomaly threshold. If the meter characteristic data exceeds three times the standard deviation, an anomaly is determined; if the meter characteristic data is within three times the standard deviation, the condition is considered normal. Furthermore, one or more statistical values ​​may be used to construct an abnormality judgment threshold for a meter parameter. For example, for meter flow, an abnormality judgment threshold may be constructed by combining a variance value with a standard deviation value.

[0034] The abnormality judgment threshold is stored in the parameter threshold library, and the parameter threshold library can update the abnormality judgment threshold according to a preset period or irregularly to improve the accuracy of abnormality judgment.

[0035] Furthermore, based on the abnormality judgment threshold, the characteristic data is compared with each other to determine whether the characteristic data is abnormal, and an abnormality monitoring indicator is generated based on the judgment result. In the embodiment of the present application, the abnormality identification of the characteristic data of the meter includes at least temperature abnormality identification, pressure abnormality identification, flow abnormality identification, and gas usage abnormality identification. The above-mentioned abnormality identification will be further described below.

[0036] The specific temperature anomaly identification includes: setting the upper and lower temperature thresholds according to the installation position of the meter, making an anomaly judgment based on the collected meter data, and generating an abnormal monitoring indicator representing that the temperature exceeds the threshold if the set upper and lower temperature thresholds are exceeded; setting the number of data with unchanged temperature for each meter based on the frequency of collection and different time granularities, and making an independent judgment on each meter respectively, and generating an abnormal monitoring indicator representing that the temperature of the meter remains unchanged if the amount of data with unchanged temperature in the meter data is greater than the set number of data; calculating the temperature standard deviation corresponding to each meter based on the collected meter data, and In principle, the abnormality threshold is three times the standard deviation of the meter temperature. Each meter's temperature fluctuation is independently assessed. If the meter temperature is within the abnormality threshold, the temperature fluctuation is considered normal. If the meter temperature exceeds the abnormality threshold, the temperature fluctuation is considered abnormal, and an abnormality monitoring indicator representing the abnormal temperature fluctuation is generated.

[0037] Flow anomaly identification specifically includes: based on the collected meter data, setting the number of data below the minimum metering threshold of the meter; if the amount of data below the minimum metering threshold of the meter is greater than the set number of data, then generating an abnormal monitoring indicator representing a small flow of flow. It should be noted that different types of meters correspond to different minimum metering thresholds; based on the collection frequency and different time granularities, setting the number of data with unchanged flow for each meter, and making independent judgments on each meter; if the amount of data with unchanged flow in the meter data is greater than the set number of data, then generating an abnormal monitoring indicator representing that the flow of the meter remains unchanged.

[0038] The pressure anomaly identification comprises: setting upper and lower thresholds of pressure according to installation positions of the meters, performing anomaly determination according to collected pressure data, and generating an abnormal monitoring index representing pressure exceeding the threshold if the pressure exceeds the set upper and lower thresholds; setting the number of data in which pressure of each meter is unchanged based on collection frequency and different time granularities, respectively, performing independent determination on each meter, and generating an abnormal monitoring index representing pressure being unchanged of the meter if the number of data in which pressure is unchanged in meter data is greater than the set number of data; calculating abnormal upper and lower limits in the month by using a box plot method based on pressure daily dimension data in the gas using period, performing anomaly determination according to the number of abnormal value data, and generating an abnormal monitoring index representing pressure fluctuation anomaly if the number of pressure fluctuation anomalies is greater than the set number.

[0039] Further, the pressure anomaly identification further comprises determining whether there is pressure jump, and specifically: Based on pressure daily dimension data in the gas using period, a correlation coefficient expression of original sequence and jump sequence is used to construct the relationship of mean value difference degree before and after variation, and the specific calculation formula is as follows: The original sequence pressure daily mean value data is marked as , and it is assumed that there is a variation point in the original sequence, the original sequence is divided into two ends, the number of front data is , and the length of rear data is , then the overall mean value of the original sequence is: ; In the formula, , the overall mean value of the original sequence is represented.

[0040] The mean value of the original sequence before jump is marked as: ; In the formula, , the mean value of the original sequence before jump is represented.

[0041] The mean value of the original sequence after jump is marked as: ; In the formula, , the mean value of the original sequence after jump is represented.

[0042] Another new sequence is constructed, which is a sequence composed of the sequence before and after the jump point of the original sequence, and is called a jump sequence , and: ; ; The overall mean value of the jump sequence is: ; In the formula, The original sequence mean before the jump, The original sequence mean after the jump, The overall mean of the jump sequence.

[0043] The original sequence X and the jump sequence Y have a correlation, and the correlation coefficient calculation formula is: ; In the formula, The overall mean of the original sequence, The overall mean of the jump sequence.

[0044] Let , The variable is brought into the above formula, and the following can be obtained: ; Where, The correlation coefficient is positive or negative, and when it is positive, the original sequence is proportional to the jump sequence, indicating that the sequence has upward jump variation; otherwise, the sequence jumps downward, and the absolute value of the correlation coefficient is controlled to determine the amplitude of the jump.

[0045] The gas anomaly identification includes: based on the collected meter data, if the daily gas consumption exceeds three times the standard deviation of any month gas consumption in the past year, an abnormal monitoring index representing a sudden increase in gas consumption is generated; based on the collected meter data, if the monthly gas consumption is zero, an abnormal monitoring index representing no gas use in the month is generated; the cumulative reading data of the user meter is obtained, the first data of today is selected, and it is judged whether they are all less than the first cumulative reading data of the previous day, and an abnormal monitoring index representing abnormal small reading is generated according to the judgment result.

[0046] Further, the gas anomaly identification also includes: based on the collected meter data, the daily gas consumption data is obtained, the regression equation of the daily gas consumption data is calculated, and the slope value of the daily gas consumption data is obtained, and the calculation formula is as follows: ; In the formula, The daily gas consumption data is represented by x, and x is the time, The equation coefficient. The regression equation of the user gas and other indicators (water consumption data or electricity consumption data) is calculated, and if the two slope values are negatively correlated, it indicates that there is an abnormality in gas consumption, and an abnormal monitoring index representing abnormal gas consumption trend is generated.

[0047] Further, the gas anomaly identification further includes: based on the collected meter data, calculating characteristic values such as hourly unevenness coefficients, and based on a least square algorithm similarity measure, calculating the similarity of the daily hourly usage pattern (or hourly unevenness coefficient sequence) and the historical typical pattern. In addition, the similarity calculation of the local daily temperature data and the daily gas usage data is combined, and compared with the knowledge about the gas usage characteristics of heating / non-heating users, to determine whether the user is a heating user, and to divide the heating month and the non-heating month for separate regularity analysis or similarity calculation according to the user type (heating / non-heating), to accurately determine whether the gas usage regularity is abnormal, and to generate an abnormal monitoring index representing the abnormality of the gas usage regularity according to the determination result.

[0048] By dynamically constructing an abnormality judgment threshold for anomaly identification of the meter data, the accuracy and adaptability of the meter fault detection are improved. The abnormality judgment threshold can be updated according to the running state of the meter and environmental changes, avoiding the misjudgment problem caused by fixed threshold due to scene differences, and especially adapting to the running characteristics of meters at different pipe network positions and altitudes. Moreover, the abnormal monitoring index generated after comparing the feature data with the abnormality judgment threshold can capture meter abnormal phenomena from multiple dimensions, covering immediate fault signals and gradual abnormal trends, which can reduce the risk of misjudgment caused by reliance on human experience, and through systematic threshold management and abnormal classification, provides a standardized data basis for subsequent fault type positioning and predictive maintenance index generation, effectively improving the efficiency and reliability of gas meter intelligent operation and maintenance.

[0049] As an optional implementation manner, as shown in FIG. 10, in step S103, after the abnormal monitoring index is generated, the classification of the abnormal monitoring index further includes the following steps: Figure 3 As shown in FIG. 10, in step S103, after the abnormal monitoring index is generated, the classification of the abnormal monitoring index further includes the following steps: Step S301: Based on the expert knowledge base, historical abnormal monitoring index data that has been classified is obtained.

[0050] Step S302: The historical abnormal monitoring index data is used to train a binary classification model.

[0051] Step S303: The unclassified abnormal monitoring index is classified by using the trained binary classification model to determine the category to which the abnormal monitoring index belongs.

[0052] Specifically, the expert knowledge base stores domain knowledge related to meter fault diagnosis, including historical abnormal monitoring indicator data that has been categorized by domain experts or based on historical maintenance records. This data has been accumulated over a long period of time and covers abnormal monitoring indicators for different meter models, operating environments, and fault scenarios. Each abnormal monitoring indicator is clearly labeled as either an "immediate fault indicator type" or a "predictive maintenance indicator type." This categorized historical abnormal monitoring indicator data is obtained from the expert knowledge base to form a targeted training dataset, providing reliable labeled data support for subsequent model training.

[0053] The above-classified historical abnormal monitoring indicator data is input into the binary classification model as a training data set. The binary classification model is trained with the immediate fault indicator type and the predictive maintenance indicator type as the category labels, and the association rules between different abnormal monitoring indicators and category labels in the historical abnormal monitoring indicator data are learned to improve the classification accuracy. Optionally, the binary classification model can be a support vector machine (SVM) classification model. Based on the trained binary classification model, the unclassified abnormal monitoring indicators are classified, and the unclassified abnormal monitoring indicators and related meter models, maintenance records and other information are input into the binary classification model. The binary classification model outputs the probability that the abnormal monitoring indicator belongs to the "immediate fault indicator type" or "predictive maintenance indicator type" based on the classification rules learned in the training phase. Alternatively, the binary classification model can also be configured to directly output the category judgment result.

[0054] Relying on the historical abnormal monitoring indicator data accumulated in the expert knowledge base, the binary classification model is learned and trained, and the abnormal monitoring indicators are classified using the binary classification model, which realizes the automation of the classification process, avoids the classification bias caused by relying solely on human experience, reduces misjudgments and missed judgments in complex abnormal scenarios, and improves the accuracy and efficiency of the classification of abnormal monitoring indicators; and, by dividing the abnormal monitoring indicators into two categories: immediate fault indicators and predictive maintenance indicators, it avoids treating all abnormalities as immediate faults, thereby improving the utilization rate of maintenance resources.

[0055] As an optional implementation, such as Figure 4 As shown, in step S104, the abnormality monitoring indicator and / or the combination of abnormality monitoring indicators belonging to the predictive maintenance indicator type is reconstructed, and the following steps are also included: Step S401: Based on the principle of the large language model, an abnormal combination model and a fault prediction model are constructed respectively.

[0056] Step S402: obtaining the correlation strength between the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type through the abnormal combination model, and / or obtaining the correlation strength between the combination of abnormal monitoring indicators in the predictive maintenance indicator type and the fault type.

[0057] In step S403, based on the correlation strength, the abnormal monitoring indicators in the predictive maintenance indicator type and / or the combination of abnormal monitoring indicators are predicted by the fault prediction model to obtain the predicted fault type and the fault prediction probability.

[0058] Specifically, the learning ability of the large language model for meter failure related knowledge is used to construct an abnormal combination model and a fault prediction model. The abnormal combination model focuses on analyzing the correlation strength between abnormal monitoring indicators (or combinations of abnormal monitoring indicators) and fault types. The fault prediction model focuses on further predicting the specific fault type and the probability of occurrence in combination with the correlation strength.

[0059] Through the abnormal combination model, the correlation strength between abnormal monitoring indicators and fault types is quantified. Single abnormal monitoring indicators (such as "pressure fluctuation anomaly") or combinations of abnormal monitoring indicators (such as "pressure fluctuation anomaly + flow invariance") belonging to the predictive maintenance indicator type are input into the abnormal combination model. The abnormal combination model analyzes the frequency of occurrence of abnormal monitoring indicators or combinations of abnormal monitoring indicators in historical fault cases to obtain the correlation strength between abnormal monitoring indicators in the predictive maintenance indicator type and fault types, and / or the correlation strength between combinations of abnormal monitoring indicators in the predictive maintenance indicator type and fault types. For example, there are 5 cases of stolen gas in the historical fault case library, and 4 of them are discovered to be "stolen gas" through gas usage trend and gas usage regularity. Therefore, the two indicator abnormalities of "gas usage trend" and "gas usage regularity" are associated with the fault prediction indicator "stolen gas", and the correlation strength is calculated.

[0060] Further, the correlation strength can be calculated as follows: all cases with the specified target fault type are filtered from the historical fault case library, and the total number of cases is counted. For each case of the target fault type, the abnormal records within a certain time window before the case occurs are checked. The number of related cases in which the target abnormal monitoring indicator appears in the time window is counted, and the correlation strength between the target abnormal monitoring indicator and the target fault type is calculated by the following formula: ; After obtaining the correlation strength, the abnormal monitoring indicators in the predictive maintenance indicator type and / or the combination of the abnormal monitoring indicators, and the obtained correlation strength are input into a fault prediction model. The fault prediction model performs fault type prediction on the abnormal monitoring indicators in the predictive maintenance indicator type and / or the combination of the abnormal monitoring indicators, and the fault prediction model can also predict a time window and a probability of fault occurrence in combination with the degradation process knowledge in the expert knowledge base. For example, according to the duration of the same abnormality and the interval of fault occurrence, it is concluded that the probability of sensor fault occurring in the next 10-15 days is 65%, thereby generating a fault prediction indicator.

[0061] Optionally, the fault prediction indicator can also include structured information such as a predicted fault type, a predicted fault probability, an abnormal monitoring indicator name, a correlation strength, and a possible root cause analysis.

[0062] Through the double-model mechanism of the abnormal combination model and the fault prediction model, the correlation strength generated by the abnormal combination model can quantify the correlation degree between the multi-dimensional abnormal monitoring indicators and the fault type, solve the problem of high false positive rate of a single indicator, the fault prediction model combines the correlation strength and the expert knowledge, realizes the conversion of the abnormal monitoring indicators into a specific fault and a fault occurrence time, provides a sufficient window for planned maintenance, and reduces the probability of non-planned shutdown of the gauge and the maintenance cost.

[0063] As an optional implementation, in step S402, the abnormal combination model establishes and outputs the correlation strength between the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type based on the expert classification rules, the gauge fault types, the abnormal performance combinations, the gauge degradation processes, and the processing records of the faults that have occurred stored in the expert knowledge base in the form of prompt words, and / or establishes and outputs the correlation strength between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type.

[0064] Specifically, the expert knowledge base includes an expert rule base, a fault mode base, and a historical fault case base. The expert rule base stores expert classification rules summarized based on expert experience and data analysis, and the expert classification rules provide a basic criterion for the abnormal combination model to judge the correlation between the abnormal monitoring indicators and the fault type; the fault mode base stores the fault types that can occur for different gauge models, the typical gauge abnormal performance combinations, and the gauge degradation processes, and the gauge degradation process describes the change law of each indicator during the process from the normal state to the fault state of the gauge; the historical fault case base stores detailed records of past faults, including the abnormal monitoring indicator performance before the fault, the fault type, the occurrence time, the maintenance record, and the root cause, and provides the actual situation of the abnormal development into a fault in the historical cases.

[0065] The abnormal combination model is trained based on the expert rule base, the fault case base and the data in the historical fault case base, so that the abnormal combination model learns to understand the correlation logic between the abnormal monitoring indicators and the fault types, and masters the indicator evolution law in the instrument degradation process. For example, "the daily average value of pressure is continuously lower than the threshold value by 10% and the standard deviation increases" may indicate a slight pipeline leakage, and "the flow rate remains unchanged accompanied by abnormal decrease in the reading" may indicate a failure of the instrument metering module.

[0066] Further, the abnormal combination model is input with a prompt word, and the abnormal combination model performs deep thinking according to the content of the prompt word, and finally generates a result according to the content of the prompt word. In an embodiment, the prompt word includes a target fault type, a target abnormal monitoring indicator, an analysis time window, a correlation strength calculation rule, and a task requirement; the abnormal combination model performs case retrieval on the target fault type and the target abnormal monitoring indicator in the expert knowledge base according to the task requirement, and filters the retrieval results based on the analysis time window, and finally calculates and outputs the correlation strength between the target fault type and the target abnormal monitoring indicator through the correlation strength calculation rule.

[0067] When facing a single abnormal monitoring indicator in the predictive maintenance indicator type, the abnormal combination model analyzes which fault types the abnormality is commonly associated with in the historical cases and the closeness of the association. For the combination of abnormal monitoring indicators, the abnormal combination model also refers to the association of multiple abnormal monitoring indicators with a specific fault type in the expert knowledge base when they appear simultaneously, and then determines the correlation strength between the combined abnormal monitoring indicators and the fault type. The correlation strength can reflect the closeness of the correlation between the abnormal monitoring indicators (or the combination of abnormal monitoring indicators) and the fault type. Finally, the abnormal combination model outputs the analysis result in the form of correlation strength, which provides key support for subsequent fault prediction.

[0068] As an optional implementation, in step S403, the fault prediction model further includes: predicting the abnormal monitoring indicator and / or the combination of abnormal monitoring indicators in the predictive maintenance indicator type based on the expert classification rules, the instrument fault types, the abnormal performance combination, the instrument degradation process and the processing records of the occurred faults stored in the expert knowledge base, in combination with the correlation strength, through the prompt word, and outputting the predicted fault type and the fault prediction probability.

[0069] Specifically, the fault prediction model receives the abnormal monitoring indicator and / or the combination of abnormal monitoring indicators in the predictive maintenance indicator type, and guides the fault prediction model to accurately call the content related to the abnormal monitoring indicator in the expert knowledge base through the prompt word, including the expert classification rules, the instrument fault types, the abnormal performance combination, the instrument degradation process and the processing records of the occurred faults.

[0070] The fault prediction model combines the content in the expert knowledge base and the correlation strength obtained in step S402 for comprehensive analysis. For example, when a certain abnormal monitoring indicator has a high correlation strength with a certain fault type, the fault prediction model focuses on the information such as the performance of the fault type in historical cases and the degradation process to determine the possibility of the current abnormality developing into the fault and calculate the corresponding probability. Through the above analysis process, the fault prediction model finally determines the predicted fault type and the corresponding fault prediction probability and outputs them, providing a clear basis for the predictive maintenance of the meter.

[0071] Further, in an embodiment, a prompt word is input to the fault prediction model, and the fault prediction model performs deep thinking according to the content of the prompt word and finally outputs a result generated according to the content of the prompt word. For example, the prompt word includes the target abnormal monitoring indicator, the situation description, the correlation strength, and the task requirement, etc. The fault prediction model performs case retrieval in the expert knowledge base based on the content of the prompt word according to the task requirement, and performs deep learning on the retrieval result to output the predicted fault type, the fault prediction probability, the possible cause analysis, and the action suggestion, etc.

[0072] As an optional implementation, the expert classification rule is generated based on expert experience and data analysis, and the expert classification rule is used to provide a basis for judgment when classifying the abnormal monitoring indicator, assisting in determining whether the abnormal monitoring indicator belongs to the immediate fault indicator type or the predictive maintenance indicator type.

[0073] Specifically, the expert classification rule is formed by combining two aspects, one is the practical experience accumulated by experts who have long been engaged in fault diagnosis of gas meters, and the other is the rules summarized after data analysis on a large amount of meter operation data and historical fault cases. After repeated verification and refinement of the actual situation, the expert classification rule can accurately reflect the potential relationship between the abnormal monitoring indicator and the fault type.

[0074] In the process of classifying the abnormal monitoring indicator, the characteristics, the scene of occurrence, and the possible impact of the abnormal monitoring indicator can be analyzed according to the expert classification rule. For example, once some abnormal monitoring indicators appear, it usually means that the meter has already occurred an immediate fault, which needs to be processed first, so the abnormal monitoring indicator can be determined to belong to the immediate fault indicator type. While other abnormal monitoring indicators may only be an early signal of the fault of the meter, and the fault will not occur immediately in the short term, so the expert classification rule assists in determining that the abnormal monitoring indicator belongs to the predictive maintenance indicator type.

[0075] As an optional implementation, as shown in Figure 5 In step S105, the maintenance suggestion is generated based on the immediate fault indicator and / or the fault prediction indicator, and further includes the following steps: Step S501, based on the expert knowledge base, obtain maintenance experience about fault type, repair process and component replacement period.

[0076] Step S502, according to the fault type of the abnormal monitoring index type of the instant fault index and / or the fault prediction index, find the corresponding maintenance experience in the expert knowledge base to generate maintenance suggestions.

[0077] Specifically, the expert knowledge base further includes a maintenance experience library, which stores detailed experience about fault type, repair process and component replacement period accumulated by maintenance personnel in long-term practice. These experiences exist in structured and unstructured forms, including standard repair procedures for different fault types, root cause analysis of common faults, component aging rules and recommended replacement period, etc.

[0078] According to the fault type corresponding to the instant fault index or the fault prediction index currently to be processed, the fault type is taken as an index to accurately find in the expert knowledge base and obtain the corresponding maintenance experience. The found maintenance experience is arranged and structured to be converted into specific maintenance suggestions. The maintenance suggestions include instant fault alarm, fault diagnosis report, fault prediction report, predictive maintenance suggestion, repair work order or dispatch instruction, fault handling steps, safety precautions, etc., to ensure that the maintenance suggestions are operable and can guide the operation and maintenance personnel to accurately and efficiently handle faults or preventive maintenance.

[0079] Further, if the maintenance case corresponding to the fault type is not found in the expert knowledge base, similar cases can also be found in the expert knowledge base according to the abnormal situation of the meter and the meter model, etc. The maintenance suggestions are generated based on the similar cases to provide valuable guidance for the operation and maintenance personnel.

[0080] According to the above description, the meter fault detection method provided by the present application trains a binary classification model based on an expert knowledge base, classifies abnormal monitoring indexes through the trained binary classification model, distinguishes between instant fault index types and predictive maintenance index types, avoids treating all meter abnormalities as instant faults, and improves the utilization efficiency of maintenance resources.

[0081] Further, the application provides a meter fault detection method, which constructs an abnormal combination model and a fault prediction model through a large language model, reconstructs abnormal monitoring indicators belonging to a predictive maintenance indicator type through the abnormal combination model and the fault prediction model, obtains the correlation strength between the abnormal monitoring indicators and the fault types and the combination of the abnormal monitoring indicators and the fault types, and predicts the fault types and the fault occurrence time corresponding to the abnormal monitoring indicators belonging to the predictive maintenance indicator type through the obtained correlation strength, so as to provide a sufficient window for planned maintenance, and reduce the probability of non-planned shutdown of the meter and the maintenance cost.

[0082] In a second aspect, the application also provides a meter fault detection system, which is applied to a gas meter, and includes a data acquisition module, a feature engineering module, an abnormal monitoring module, an expert knowledge base module, an abnormal classification module, a fault prediction module, and a result output module.

[0083] As shown in Figure 6 The data acquisition module is configured to acquire multi-dimensional data of the meter and pre-process the acquired data. The feature engineering module is configured to extract features from the acquired data and obtain feature data reflecting the running state of the meter. The abnormal monitoring module is configured to include a parameter threshold library, and generate abnormal monitoring indicators by identifying the features based on the parameter threshold library. The expert knowledge base module is configured to store meter fault types, meter degradation processes, abnormal performance combinations, expert classification rules, processing records of faults that have occurred, and maintenance experience.

[0084] The abnormal classification module is configured to classify the abnormal monitoring indicators based on the expert knowledge base module, and determine the categories of the abnormal monitoring indicators, including an immediate fault indicator type and a predictive maintenance indicator type. The fault prediction module is configured to reconstruct the abnormal monitoring indicators and / or combinations of the abnormal monitoring indicators belonging to the predictive maintenance indicator type based on the expert knowledge base, and generate fault prediction indicators, which are configured to identify predicted fault types and fault prediction probabilities. The result output module is configured to output corresponding maintenance suggestions based on the abnormal monitoring indicators of the immediate fault indicator type and / or the fault prediction indicators.

[0085] According to the above description, the meter fault detection system provided by the application classifies abnormal monitoring indicators based on an expert knowledge base, distinguishes between immediate fault indicator types and predictive maintenance indicator types, improves the classification accuracy of abnormal monitoring indicators, and avoids treating all meter abnormalities as immediate faults, thereby improving the utilization efficiency of maintenance resources; and the abnormal monitoring indicators belonging to the predictive maintenance indicator types are reconstructed, the fault types and fault occurrence times corresponding to the abnormal monitoring indicators belonging to the predictive maintenance indicator types are predicted, thereby providing sufficient windows for planned maintenance, and reducing the probability of meter unplanned downtime and maintenance costs.

[0086] As an optional implementation manner, the fault prediction module is further configured to: based on the principle of a large language model, construct an abnormal combination model and a fault prediction model respectively.

[0087] The abnormal combination model is used to obtain the association strength between the abnormal monitoring indicators in the predictive maintenance indicator types and the fault types, and / or the association strength between the combination of the abnormal monitoring indicators in the predictive maintenance indicator types and the fault types.

[0088] Based on the association strength, the fault prediction model is used to predict the abnormal monitoring indicators and / or the combination of the abnormal monitoring indicators in the predictive maintenance indicator types, to obtain the predicted fault types and fault prediction probabilities.

[0089] Based on the principle of a large language model, the abnormal combination model and the fault prediction model are constructed respectively, the abnormal monitoring indicators belonging to the predictive maintenance indicator types are reconstructed by using the abnormal combination model and the fault prediction model, the association strength between the abnormal monitoring indicators and the fault types, and the association strength between the combination of the abnormal monitoring indicators and the fault types are obtained, the fault types and fault occurrence times corresponding to the abnormal monitoring indicators belonging to the predictive maintenance indicator types are predicted by using the obtained association strength, thereby providing sufficient windows for planned maintenance, and reducing the probability of meter unplanned downtime and maintenance costs.

[0090] As an optional implementation manner, the abnormal classification module is further configured to: based on the expert knowledge base module, obtain historical abnormal monitoring indicator data that has been classified; use the historical abnormal monitoring indicator data to train a binary classification model, classify unclassified abnormal monitoring indicators by using the trained binary classification model, and determine the category to which the abnormal monitoring indicators belong.

[0091] The binary classification model is trained by using historical abnormal monitoring index data accumulated in the expert knowledge base, and the binary classification model is used for classifying the abnormal monitoring index, so that the classification process is automated, the classification deviation caused by simply relying on manual experience is avoided, the misjudgment and omission in a complex abnormal scenario are reduced, and the accuracy and efficiency of classifying the abnormal monitoring index are improved. Moreover, by classifying the abnormal monitoring index into instant failure indicators and predictive maintenance indicators, all abnormalities are avoided from being treated as instant failures, and the utilization rate of maintenance resources is improved.

[0092] It can be understood that the word "exemplary" used herein means "serving as an example, instance, or illustration". Any embodiment described as "exemplary" is not necessarily preferred or superior over other embodiments and / or does not exclude features that are combined with other embodiments. It should be understood that certain features of the present application described in the context of separate embodiments can also be provided in combination, for clarity and for illustration, various features of the present application described in the context of a single embodiment can also be provided separately or in any appropriate combination. Conversely, various features of the present application described in the context of a single embodiment can also be provided separately or in any appropriate combination or as any other described embodiment of the present application.

[0093] In the description of the present application, unless otherwise specified, " / " means "or", for example, A / B can mean A or B. "And / or" herein is only a description of the relationship between the associated objects, which means that there can be three relationships, for example, A and / or B, which can mean that A exists alone, A and B exist together, and B exists alone. In addition, "at least one" means one or more, and "multiple" means two or more. "First", "second", and the like do not limit the quantity and execution order, and "first", "second", and the like do not necessarily mean different.

[0094] The above disclosure is only the preferred embodiment of the present application, and is not intended to limit the scope of the present application. Those skilled in the art can understand that changes, modifications, substitutions, combinations, simplifications, etc. without departing from the spirit and scope of the present application and the appended claims, are equivalent replacement methods and still belong to the scope of the present application.

Claims

1. A meter fault detection method, characterized in that: The meter fault detection method comprises the following steps: Collect multi-dimensional data from meters and extract multiple feature data reflecting the meter's operating status from the meter data through feature engineering; Constructing a parameter threshold library, performing abnormality identification on each of the characteristic data based on the parameter threshold library, and generating abnormality monitoring indicators; Building an expert knowledge base, classifying the abnormal monitoring indicators based on the expert knowledge base, and determining the categories to which the abnormal monitoring indicators belong, the categories including: immediate failure indicator type and predictive maintenance indicator type; Based on the expert knowledge base, reconstruct the abnormality monitoring indicator and / or the combination of the abnormality monitoring indicators belonging to the predictive maintenance indicator type to generate a fault prediction indicator, wherein the fault prediction indicator is configured to identify a predicted fault type and a fault prediction probability; Based on the abnormal monitoring indicator and / or the fault prediction indicator of the immediate fault indicator type, corresponding maintenance suggestions are provided.

2. The meter fault detection method according to claim 1, characterized in that: Classifying the abnormal monitoring indicators includes: Based on the expert knowledge base, obtaining historical abnormal monitoring indicator data that has been classified into categories; Using the historical anomaly monitoring indicator data to train a binary classification model; The unclassified abnormal monitoring indicators are classified by the trained binary classification model to determine the category to which the abnormal monitoring indicators belong.

3. The meter fault detection method according to claim 1, characterized in that: The meter fault detection method further includes: Based on the principle of large language model, an abnormal combination model and a fault prediction model are constructed respectively; Obtaining, by means of the abnormal combination model, the strength of association between the abnormal monitoring indicator in the predictive maintenance indicator type and the fault type, and / or obtaining the strength of association between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type; Based on the association strength, the abnormal monitoring indicator and / or the combination of abnormal monitoring indicators in the predictive maintenance indicator type is predicted by the fault prediction model to obtain a predicted fault type and a fault prediction probability.

4. The meter fault detection method according to claim 3, characterized in that: The meter fault detection method further includes: The abnormal combination model is based on the expert classification rules, meter fault types, abnormal performance combinations, meter degradation processes and fault handling records stored in the expert knowledge base in the form of prompt words. Establishing and outputting the strength of association between the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type, and / or establishing and outputting the strength of association between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type.

5. The meter fault detection method according to claim 3, characterized in that: The meter fault detection method further includes: The fault prediction model uses prompt words based on the expert classification rules, meter fault types, abnormal performance combinations, meter degradation processes, and handling records of existing faults stored in the expert knowledge base, combined with the association strength. The abnormality monitoring indicator and / or the combination of the abnormality monitoring indicators in the predictive maintenance indicator type are predicted, and the predicted fault type and the fault prediction probability are output.

6. The meter fault detection method according to claim 5, characterized in that: The expert classification rules are generated based on expert experience and data analysis summary. The expert classification rules are used to provide a judgment basis when classifying the abnormal monitoring indicators, and to assist in judging whether the abnormal monitoring indicators belong to the immediate fault indicator type or the predictive maintenance indicator type.

7. The meter fault detection method according to claim 1, characterized in that: The meter fault detection method further includes: Based on the expert knowledge base, obtain maintenance experience regarding fault types, maintenance processes, and component replacement cycles; According to the abnormal monitoring indicator of the immediate fault indicator type and / or the fault type of the fault prediction indicator, the corresponding maintenance experience is searched in the expert knowledge base to generate the maintenance suggestion.

8. The meter fault detection method according to claim 1, characterized in that: The meter fault detection method further includes: Calculate statistical values ​​of meter parameters based on the fault-free operation period of the meter, wherein the statistical values ​​include one or more of the following types of statistical values: mean, variance, quantile, and standard deviation of the meter parameters; Based on at least one statistical value, abnormality judgment thresholds for different types of instrument parameters are respectively constructed, the abnormality judgment thresholds are stored in the parameter threshold library, and the parameter threshold library dynamically updates the abnormality judgment thresholds; The abnormality determination threshold is obtained from the parameter threshold library, and the characteristic data is compared with the abnormality determination threshold to determine whether the characteristic data has an abnormality. According to the judgment result, if an abnormality exists, the abnormality monitoring indicator is generated.

9. A meter fault detection system, characterized in that: The meter fault detection system includes: The data collection module is configured to collect multi-dimensional data from the meter and pre-process the collected data; a feature engineering module configured to extract features from the collected data to obtain feature data that can reflect the operating status of the meter; an anomaly monitoring module configured to include a parameter threshold library, and to generate an anomaly monitoring indicator by performing anomaly identification on the feature data based on the parameter threshold library; An expert knowledge base module is configured to store meter fault types, meter degradation processes, abnormal performance combinations, expert classification rules, handling records of existing faults, and maintenance experience; an abnormality classification module, configured to classify the abnormality monitoring indicators based on the expert knowledge base module and determine the categories of the abnormality monitoring indicators, the categories including: immediate failure indicator type and predictive maintenance indicator type; a fault prediction module configured to reconstruct the abnormality monitoring indicator and / or the combination of abnormality monitoring indicators belonging to the predictive maintenance indicator type based on the expert knowledge base to generate a fault prediction indicator, wherein the fault prediction indicator is configured to identify a predicted fault type and a predicted fault probability; The result output module is configured to output corresponding maintenance suggestions based on the abnormal monitoring indicator and / or the fault prediction indicator of the immediate fault indicator type.

10. The meter fault detection system according to claim 9, characterized in that: The fault prediction module is further configured to: Based on the principle of large language model, an abnormal combination model and a fault prediction model are constructed respectively; Obtaining, by means of the abnormal combination model, the strength of association between the abnormal monitoring indicator in the predictive maintenance indicator type and the fault type, and / or obtaining the strength of association between the combination of the abnormal monitoring indicators in the predictive maintenance indicator type and the fault type; Based on the association strength, the abnormal monitoring indicator and / or the combination of abnormal monitoring indicators in the predictive maintenance indicator type is predicted by the fault prediction model to obtain a predicted fault type and a fault prediction probability.

Images