Intelligent root cause analysis and alarm method based on unsupervised decision model

Abstract

The invention discloses an intelligent root cause analysis and alarm method based on an unsupervised decision model, relates to the technical field of data processing, and solves the problems of low data processing speed, inaccurate processing result and low pertinence in the prior art. According to the method, the power grid operation and maintenance data, the weather data and the fault data are collected and subjected to data preprocessing, then the Pearson's correlation coefficients of the power grid operation and maintenance data and the weather data and the fault are obtained at the first data collection fault occurrence time, and then the fault occurrence time is calculated according to the calculated Pearson's correlation coefficients; the data types with the correlation coefficient absolute values larger than a set threshold value are screened out, and finally factors related to fault occurrence are mined out and sorted according to correlation strength. According to the method, preliminary information screening is carried out before fault related factors are mined by the unsupervised decision model, so that the data volume needing to be processed by the algorithm is reduced, the processing efficiency is ensured, the pertinence of the algorithm is higher, and the output result of the algorithm is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent root cause analysis and alarm method based on an unsupervised decision model. Background Art

[0002] With the rapid development of information technology, the management of power systems is increasingly relying on the digital transformation of power grid operation and maintenance data. Thanks to the popularization of smart meters and the application of big data technology, the distribution network has accumulated massive amounts of data during operation. These data are not only used in many fields such as predicting power load, optimizing the location selection of substations, and improving operation and management efficiency, but also greatly enhance the working efficiency of the power grid. The advancement of digital technology enables the efficient collection and preservation of power grid operation and maintenance data, laying a solid foundation for subsequent data analysis and application. In terms of the digitalization and intelligent construction of power grids, for example, the measures taken by the State Grid Chongqing Electric Power Company are to use advanced technologies such as cloud computing, big data, the Internet of Things, mobile Internet, artificial intelligence and blockchain to integrate multi-dimensional information such as power grid and equipment operation monitoring data, thereby improving the intelligent operation and maintenance capabilities of the power grid.

[0003] However, due to the large amount of operation and maintenance data, even if intelligent technologies such as big data are used to avoid the trouble of manual screening, the large amount of data still makes the processing speed slow, the processing results inaccurate and not targeted.

[0004] In view of this, an intelligent root cause analysis and alarm method based on an unsupervised decision model is needed. Summary of the invention

[0005] In order to solve the problems of slow data processing speed, inaccurate processing results and lack of pertinence in the prior art, the present invention provides an intelligent root cause analysis and alarm method based on an unsupervised decision model, which can calculate the Pearson correlation coefficient between the acquired power grid operation and maintenance data and weather data and the fault data, and then screen out the data with a greater correlation with the fault data, and then input the screened data into the unsupervised decision model for processing, and further derive the correlation ranking between these data and the fault data. Finally, with respect to the data screened for the second time, further screen out data with stronger correlation for prediction, and then perform fault warning based on the predicted data. The specific technical scheme is as follows:

[0006] An intelligent root cause analysis and alarm method based on an unsupervised decision model comprises the following steps:

[0007] Collecting power grid operation and maintenance data, weather data, and fault data and performing data preprocessing, wherein the data after the data preprocessing is the first data;

[0008] Obtain the Pearson correlation coefficient between each power grid operation and maintenance data and each weather data and the fault at the time when the first data acquisition fault occurs;

[0009] According to the calculated Pearson correlation coefficient, data types with absolute values ​​of correlation coefficients greater than a set threshold are screened out, and data types that do not meet the conditions are removed from the first data to form second data;

[0010] Based on the second data, the FP-Growth algorithm is used to mine out factors related to the occurrence of the fault and sort them according to the strength of association.

[0011] Preferably, the power grid operation and maintenance data include voltage, current, power, equipment temperature, and equipment environment humidity at different sampling times; the weather data include weather temperature, weather humidity, wind speed, light intensity, and season at different sampling times; and the fault data include faults and the time when the faults occurred.

[0012] Preferably, data preprocessing includes:

[0013] Process the sampling time of power grid operation and maintenance data and weather data into a unified time interval;

[0014] The data of different dimensions are standardized to eliminate the dimension effect so that the data can be compared at the same scale. In this embodiment, the normalization process is adopted to uniformly map the data into the interval [0,1].

[0015] Preferably, the calculation formula of the Pearson correlation coefficient is as follows:

[0016]

[0017] In the formula, x i is the i-th observation of feature x, y i is the ith observation of a failure occurrence, x is the mean of feature x, y is the mean of the failure occurrences, and n is the number of observations.

[0018] Preferably, the steps of using the FP-Growth algorithm to mine factors related to the occurrence of a fault and sorting them according to the strength of association are as follows:

[0019] Numerically encode all features and fault labels in the second data so that the algorithm can process them;

[0020] Format the data into transaction datasets, where each transaction contains a set of feature items;

[0021] Calculate the support of the item, and then use the minimum support threshold Support(X) min , select the support greater than or equal to Support(X) minThe calculation formula of the support of the item is as follows:

[0022]

[0023] Where X is an item set, N(X) is the number of transactions containing item set X, and N total is the total number of transactions;

[0024] For each frequent item set, a conditional pattern base and a conditional FP tree are constructed to mine deeper frequent item sets;

[0025] Use FP-tree to recursively mine frequent itemsets until no more frequent itemsets can be found;

[0026] Based on the frequent item sets, association rules are generated and the confidence of each rule is calculated. The calculation formula of the confidence is as follows:

[0027]

[0028] In the formula, X is the antecedent of the rule, Y is the consequent of the rule, Support(X∪Y) is the frequency of item sets X and Y appearing at the same time, and Support(X) is the frequency of item set X appearing;

[0029] Select a confidence level greater than or equal to the minimum confidence threshold min The rules are then sorted according to their confidence.

[0030] Preferably, the method further comprises the following steps:

[0031] Setting a sorting threshold, obtaining the data type before the sorting threshold, and further filtering the second data to obtain the third data;

[0032] Acquire prediction data of third data, wherein the prediction data of power grid operation and maintenance data is acquired by a trained data prediction model;

[0033] An alarm threshold is set, and the predicted data of the third data is compared with the alarm threshold. If the predicted data exceeds the alarm threshold, an alarm operation is performed.

[0034] Preferably, the construction process of the data prediction model is as follows:

[0035] Determine the ARIMA model parameters: Draw the autocorrelation function and partial autocorrelation function graphs to determine the parameters of the ARIMA model, namely the autoregressive term p, the difference order d, and the moving average term q; use the grid search method to traverse different combinations of p, d, and q to find the optimal model parameters;

[0036] Model building and training: Define the ARIMA model based on the determined parameters; fit the model using the training data set;

[0037] Model evaluation: Check whether the residual is white noise to evaluate the model fitting effect and calculate the prediction error;

[0038] Model prediction: Use the fitted model to predict future data.

[0039] Preferably, the forecast data of the weather data in the third data is directly obtained from the meteorological platform.

[0040] A computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the intelligent root cause analysis and alarm method based on an unsupervised decision model as described above.

[0041] A processor is used to run a program, wherein when the program is run, the intelligent root cause analysis and alarm method based on an unsupervised decision model as described above is executed.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The present invention collects power grid operation and maintenance data, weather data and fault data and performs data preprocessing, and the data after data preprocessing is the first data; then obtain the Pearson correlation coefficient between each power grid operation and maintenance data and each weather data and the fault at the time when the first data is collected; then, according to the calculated Pearson correlation coefficient, screen out the data types whose absolute value of the correlation coefficient is greater than the set threshold, remove the data types that do not meet the conditions in the first data, and form the second data; finally, based on the second data, use the FP-Growth algorithm to mine out the factors related to the occurrence of the fault, and sort them according to the strength of association. Before using the FP-Growth algorithm to mine out the fault-related factors, the present invention also uses the Pearson correlation coefficient to perform preliminary information screening, thereby reducing the amount of data that the FP-Growth algorithm needs to process, while ensuring the processing efficiency, it also makes the FP-Growth algorithm more targeted and the output results of the algorithm more accurate.

[0044] 2. The present invention also sets a sorting threshold, obtains the data type before the sorting threshold, further filters from the second data, and obtains the third data; then obtains the predicted data of the third data, wherein the predicted data of the power grid operation and maintenance data is obtained by the trained data prediction model; and then sets an alarm threshold, compares the predicted data of the third data with the alarm threshold, and performs an alarm operation if it exceeds the alarm threshold. Through the above steps, the present invention further processes the output results of the FP-Growth algorithm, and pays attention to the first few factors of the output results of the FP-Growth algorithm by setting a threshold, that is, the factors that have a great impact on the fault, and then based on the predicted data of these factors that have a great impact on the fault, sets the corresponding alarm advance and then performs an alarm operation, further prompting the management personnel to pay attention to the possible faults that may be caused by data changes, and then facilitates the adjustment of the fault influencing factors before the fault occurs, thereby preventing the occurrence of the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0046] Figure 1 is a flow chart of the overall method of the present invention;

[0047] Figure 2 is a flow chart of the present invention using the FP-Growth algorithm;

[0048] Figure 3 It is a flow chart of the training prediction model of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0051] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0052] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0053] In one embodiment of the present invention, an intelligent root cause analysis and alarm method based on an unsupervised decision model is provided, such as Figure 1 As shown, it includes the following steps:

[0054] Step 1: Collect power grid operation and maintenance data, weather data and fault data and perform data preprocessing, the data after data preprocessing is the first data;

[0055] The power grid operation and maintenance data includes voltage, current, power, equipment temperature, and equipment environment humidity at different sampling times; the weather data includes weather temperature, weather humidity, wind speed, light intensity, and season at different sampling times; and the fault data includes faults and the time when the faults occurred;

[0056] Data preprocessing includes:

[0057] Process the sampling time of power grid operation and maintenance data and weather data into a unified time interval;

[0058] The data of different dimensions are standardized to eliminate the dimension effect so that the data can be compared at the same scale. In this embodiment, the normalization process is adopted to uniformly map the data into the interval [0,1].

[0059] Step 2: Obtain the Pearson correlation coefficient between each power grid operation and maintenance data and each weather data and the fault at the time when the first data acquisition fault occurs.

[0060] The calculation formula of the Pearson correlation coefficient is as follows:

[0061]

[0062] In the formula, x i is the i-th observation of feature x, y i is the ith observation of a failure occurrence, x is the mean of feature x, y is the mean of the failure occurrences, and n is the number of observations.

[0063] The calculated Pearson correlation coefficient can reflect which data in each power grid operation and maintenance data and each weather data have a strong linear relationship with the occurrence of faults.

[0064] Step 3: According to the calculated Pearson correlation coefficient, the data types whose absolute value of the correlation coefficient is greater than a set threshold (such as 0.5) are screened out, and the data types that do not meet the conditions are removed from the first data to form the second data;

[0065] Step 4: Based on the second data, use the FP-Growth algorithm to mine the factors related to the occurrence of the fault and sort them according to the strength of association; Figure 2 As shown, the details are as follows:

[0066] S401: Numerically encode all features (such as voltage, current, power, equipment temperature, weather temperature, wind speed, season, time) and fault labels in the second data so that the algorithm can process them;

[0067] S402: Format the data into a transaction data set, where each transaction contains a set of feature items. For example, a transaction may be “voltage is high, current is normal, power is high, equipment temperature is high, weather temperature is low, wind speed is medium, season: winter, time: 14:00”. If a fault occurs during this time period, a “fault” label is added to the end of the transaction.

[0068] S403: Construct an FP-tree. The FP-tree is a compressed data structure for storing frequent item sets. The construction steps are: calculate the support of the item, and then according to the minimum support threshold Support(X) min , select the support greater than or equal to Support(X) min The calculation formula of the support of the item is as follows:

[0069]

[0070] Where X is an item set, N(X) is the number of transactions containing item set X, and N total is the total number of transactions;

[0071] S404: For each frequent item set, a conditional pattern base (CM) and a conditional FP (CFPT) tree are constructed to mine deeper frequent item sets.

[0072] S405: recursively mine frequent item sets using the FP-tree until no more frequent item sets can be found;

[0073] S406: Generate association rules based on frequent item sets and calculate the confidence of each rule. The calculation formula of the confidence is as follows:

[0074]

[0075] In the formula, X is the antecedent of the rule, Y is the consequent of the rule, Support(X∪Y) is the frequency of item sets X and Y appearing at the same time, and Support(X) is the frequency of item set X appearing;

[0076] S407: Select a confidence level greater than or equal to the minimum confidence threshold Confidence min Then sort the rules according to the confidence. Set a sorting threshold (such as 3, only get the first three data), get the data type before the sorting threshold, further filter from the second data, and get the third data.

[0077] After the above steps, we can determine the data items that affect the fault and the order of these data items based on the degree of influence. In other words, we can determine which data among the power grid operation and maintenance data and weather data has the greatest impact on the fault, which data has the second greatest impact on the fault, and so on. At this point, the root cause analysis of the fault is completed.

[0078] Next, data prediction and alarm are performed. That is, starting from the root cause of the fault, focusing on several factors that have a high degree of impact on the fault (such as voltage, equipment temperature, and weather temperature), these factors are predicted, and then the fault is warned based on the prediction results. This process is implemented through the prediction model, which predicts the data at future times, and then determines whether an alarm is needed based on the set alarm threshold.

[0079] Step 5: Train a data prediction model to predict future data based on historical data. The data training model is used to predict the data value of the power grid operation and maintenance data in the third data at future times, and the predicted data of weather data is obtained from meteorological platforms such as weather forecasts.

[0080] This embodiment uses the ARIMA model to build a data prediction model, such as Figure 3 As shown, the following are the specific steps:

[0081] S01: Determine the ARIMA model parameters: Draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) graphs to determine the parameters p (autoregressive term), d (difference order), and q (moving average term) of the ARIMA model; use the grid search method to traverse different combinations of p, d, and q to find the optimal model parameters.

[0082] S02: Model building and training: Define the ARIMA model based on the determined parameters; fit the model using the training data set.

[0083] S03: Model evaluation: Check whether the residual is white noise to evaluate the fitting effect of the model, and calculate the prediction error, such as mean square error (MSE), to evaluate the prediction performance of the model.

[0084] S04: Model prediction: Use the fitted model to predict future data.

[0085] In short, the ARIMA model uses historical data (voltage, current, power, equipment temperature, etc.) as the input of the model, and the model analyzes the time series characteristics of these data, including trends, seasonality, and random fluctuations. By fitting these historical data, the ARIMA model can capture the law of data changes over time and predict future data points accordingly.

[0086] Specifically, the output of the model will include predicted values ​​for the same variables (voltage, current, power, device temperature) in the future. These predictions can be short-term, such as the next few hours or days, or long-term, such as the next few weeks or months, depending on the characteristics of the data and the prediction goals. The accuracy of the prediction will be affected by many factors, including the choice of model parameters, the stationarity of the data, the influence of external factors, etc. Therefore, in practical applications, it is usually necessary to carefully adjust and verify the model to ensure the reliability of the prediction results.

[0087] Step 6: Set an alarm threshold, compare the predicted data of the third data with the alarm threshold, and if it exceeds the alarm threshold, perform an alarm operation.

[0088] In summary, the present invention collects power grid operation and maintenance data, weather data and fault data and performs data preprocessing, and the data after data preprocessing is the first data; then the Pearson correlation coefficient between each power grid operation and maintenance data and each weather data and the fault at the time when the first data is collected and the fault occurs is obtained; then, according to the calculated Pearson correlation coefficient, the data types whose absolute value of the correlation coefficient is greater than the set threshold are screened out, and the data types that do not meet the conditions in the first data are removed to form the second data; finally, based on the second data, the FP-Growth algorithm is used to mine out the factors related to the occurrence of the fault, and the factors are sorted according to the strength of association. Before using the FP-Growth algorithm to mine out the fault-related factors, the present invention also uses the Pearson correlation coefficient to perform preliminary information screening, thereby reducing the amount of data that the FP-Growth algorithm needs to process, while ensuring the processing efficiency, it also makes the FP-Growth algorithm more targeted and the output results of the algorithm more accurate.

[0089] Moreover, the present invention also sets a sorting threshold, obtains the data type before the sorting threshold, further screens from the second data, and obtains the third data; then obtains the predicted data of the third data, wherein the predicted data of the power grid operation and maintenance data is obtained by the trained data prediction model; and then sets an alarm threshold, compares the predicted data of the third data with the alarm threshold, and performs an alarm operation if it exceeds the alarm threshold. Through the above steps, the present invention further processes the output results of the FP-Growth algorithm, and pays attention to the first few factors of the output results of the FP-Growth algorithm by setting a threshold, that is, the factors that have a great impact on the fault, and then based on the predicted data of these factors that have a great impact on the fault, sets the corresponding alarm advance and then performs an alarm operation, further prompting the management personnel to pay attention to the possible faults that may be caused by data changes, and then facilitates the adjustment of the fault influencing factors before the fault occurs, thereby preventing the occurrence of the fault.

[0090] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0091] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0092] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0093] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., which can store program code.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. An intelligent root cause analysis and alarm method based on an unsupervised decision model, characterized in that: The following steps are involved: Collecting power grid operation and maintenance data, weather data, and fault data and performing data preprocessing, wherein the data after the data preprocessing is the first data; Obtain the Pearson correlation coefficient between each power grid operation and maintenance data and each weather data and the fault at the time when the first data acquisition fault occurs; According to the calculated Pearson correlation coefficient, data types with absolute values ​​of correlation coefficients greater than a set threshold are screened out, and data types that do not meet the conditions are removed from the first data to form second data; Based on the second data, the FP-Growth algorithm is used to mine out factors related to the occurrence of the fault and sort them according to the strength of association.

2. According to claim 1, an intelligent root cause analysis and alarm method based on an unsupervised decision model is characterized in that: The power grid operation and maintenance data includes voltage, current, power, equipment temperature, and equipment environment humidity at different sampling times; the weather data includes weather temperature, weather humidity, wind speed, light intensity, and season at different sampling times; and the fault data includes faults and the time when the faults occurred.

3. According to claim 1, the intelligent root cause analysis and alarm method based on an unsupervised decision model is characterized in that: Data preprocessing includes: Process the sampling time of power grid operation and maintenance data and weather data into a unified time interval; The data of different dimensions are standardized to eliminate the dimension effect so that the data can be compared at the same scale. In this embodiment, normalization is used to uniformly map the data into the interval [0, 1].

4. The intelligent root cause analysis and alarm method based on an unsupervised decision model according to claim 1 is characterized in that: The calculation formula of the Pearson correlation coefficient is as follows: In the formula, x i is the ith observation of feature x, y i is the ith observation value of the fault occurrence, is the mean of feature x, y is the mean number of failures, and n is the number of observations.

5. The intelligent root cause analysis and alarm method based on an unsupervised decision model according to claim 1 is characterized in that: The steps to use the FP-Growth algorithm to mine the factors related to the occurrence of the fault and sort them according to the strength of association are as follows: Numerically encode all features and fault labels in the second data so that the algorithm can process them; Format the data into transaction datasets, where each transaction contains a set of feature items; Calculate the support of the item, and then use the minimum support threshold Support(X) min , select the support greater than or equal to Support(X) min The calculation formula of the support of the item is as follows: Where X is an item set, N(X) is the number of transactions containing item set X, and N total is the total number of transactions; For each frequent item set, a conditional pattern base and a conditional FP tree are constructed to mine deeper frequent item sets; Use FP-tree to recursively mine frequent itemsets until no more frequent itemsets can be found; Based on the frequent item sets, association rules are generated and the confidence of each rule is calculated. The calculation formula of the confidence is as follows: In the formula, X is the antecedent of the rule, Y is the consequent of the rule, Support(X∪Y) is the frequency of item sets X and Y appearing at the same time, and Support(X) is the frequency of item set X appearing; Select a confidence level greater than or equal to the minimum confidence threshold min The rules are then sorted according to their confidence.

6. The intelligent root cause analysis and alarm method based on an unsupervised decision model according to claim 1 is characterized in that: The following steps are also included: Setting a sorting threshold, obtaining the data type before the sorting threshold, and further filtering the second data to obtain the third data; Acquire prediction data of third data, wherein the prediction data of power grid operation and maintenance data is acquired by a trained data prediction model; An alarm threshold is set, and the predicted data of the third data is compared with the alarm threshold. If the predicted data exceeds the alarm threshold, an alarm operation is performed.

7. The intelligent root cause analysis and alarm method based on an unsupervised decision model according to claim 6 is characterized in that: The construction process of the data prediction model is as follows: Determine the ARIMA model parameters: Draw the autocorrelation function and partial autocorrelation function graphs to determine the parameters of the ARIMA model, namely the autoregressive term p, the difference order d, and the moving average term q; use the grid search method to traverse different combinations of p, d, and q to find the optimal model parameters; Model building and training: Define the ARIMA model based on the determined parameters; fit the model using the training data set; Model evaluation: Check whether the residual is white noise to evaluate the model fitting effect and calculate the prediction error; Model prediction: Use the fitted model to predict future data.

8. The intelligent root cause analysis and alarm method based on an unsupervised decision model according to claim 6 is characterized in that: The forecast data of the weather data in the third data is directly obtained from the meteorological platform.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the intelligent root cause analysis and alarm method based on the unsupervised decision model as described in any one of claims 1 to 8.

10. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the intelligent root cause analysis and alarm method based on an unsupervised decision model as described in any one of claims 1 to 8.

Images