A fault prediction method, device, equipment and computer readable storage medium

Abstract

The application discloses a kind of fault prediction method, device, equipment and computer readable storage medium, applied to fault detection technical field, the method includes: obtaining the first concentration sequence of methane in transformer, the second concentration sequence of ethylene and the third concentration sequence of acetylene;According to the first concentration sequence, the second concentration sequence, concentration sequence, respectively, concentration prediction is carried out, determines the first predicted concentration value of methane, the second predicted concentration value of ethylene and the third predicted concentration value of acetylene at future time point;In the case where at least one of first predicted concentration value, second predicted concentration value and third predicted concentration value exceeds corresponding preset threshold value, using three-ratio method, according to first predicted concentration value, second predicted concentration value and third predicted concentration value, fault diagnosis is carried out on transformer.And current model input multiple parameters for fault diagnosis, resulting in high fault prediction complexity, compared with the input parameter of the present application is single, so it can reduce the prediction complexity.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault detection, in particular to a fault prediction method and device, electronic equipment and a storage medium. BACKGROUND

[0002] Oil-immersed transformers are important equipment in power grids, and online monitoring and fault diagnosis of the same are crucial for ensuring stable operation of the power grid. In order to achieve fault prediction, a deep learning model with prediction capability, such as a recurrent neural network (RNN), can be used. However, there are multiple components in the oil-dissolved gas, and a complex prediction model with a large number of parameters needs to be constructed, which not only requires a large amount of training time, but also consumes a large amount of computing resources when performing fault prediction based on the prediction model.

[0003] Therefore, how to reduce the consumption of computing resources during fault prediction is a technical problem that needs to be solved by those skilled in the art. SUMMARY

[0004] To solve the problem of complex structure of the transformer fault prediction model, the present application provides a fault prediction method, device, electronic equipment and storage medium.

[0005] In a first aspect, the present application provides a fault prediction method, comprising:

[0006] obtaining a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points;

[0007] performing concentration prediction according to the first concentration sequence, the second concentration sequence and the third concentration sequence, respectively, to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene;

[0008] when at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value, performing fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a three-ratio method.

[0009] Optionally, performing concentration prediction according to the first concentration sequence, the second concentration sequence and the third concentration sequence, respectively, to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene, comprises:

[0010] performing data fitting according to a target concentration sequence corresponding to the target gas to determine a target function relationship between the concentration value of the target gas and time;

[0011] determine a target predicted concentration value corresponding to the target gas according to the target function relationship;

[0012] The target gas is at least one of methane, ethylene, and acetylene.

[0013] When the target gas is methane, the target concentration sequence is a first concentration sequence, and the target predicted concentration value is a first predicted concentration value.

[0014] When the target gas is ethylene, the target concentration sequence is a second concentration sequence, and the target predicted concentration value is a second predicted concentration value.

[0015] When the target gas is acetylene, the target concentration sequence is a third concentration sequence, and the target predicted concentration value is a third predicted concentration value.

[0016] Optionally, the data fitting according to the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value and the time of the target gas comprises:

[0017] A first function relationship in a polynomial form is established in advance for the target gas, and the first function relationship is: ; wherein, represents the ith time point, represents a predicted concentration value at the ith time point, , , , all are undetermined polynomial fitting coefficients.

[0018] The first function relationship is data fitted according to the target concentration sequence corresponding to the target gas, each polynomial fitting coefficient is determined, and a target function relationship corresponding to the target gas is generated; wherein, the target function relationship is a polynomial function relationship formula for predicting the concentration value.

[0019] Optionally, the polynomial fitting coefficient is: ; wherein, represents a concentration value corresponding to the ith time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

[0020] Optionally, the data fitting according to the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value and the time of the target gas comprises:

[0021] A second function relationship in an exponential form is established in advance for the target gas, and the second function relationship is: ; wherein, represents the ith time point, represents a predicted concentration value at the i-th time point, a is a first fitting coefficient to be determined, b is a second fitting coefficient to be determined, and c is a third fitting coefficient to be determined;

[0022] The second function relationship is data fitted according to a target concentration sequence corresponding to the target gas, to determine the first fitting coefficient, the second fitting coefficient, and the third fitting coefficient, and to generate the target function relationship; wherein the target function relationship is an exponential function relationship for predicting a concentration value.

[0023] Optionally, the data fitting of the second function relationship according to the target concentration sequence corresponding to the target gas, to determine the first fitting coefficient, the second fitting coefficient, and the third fitting coefficient, and to generate the target function relationship, comprises:

[0024] The error square sum between the real concentration value and the predicted concentration value is simplified to obtain a simplified loss function : ;

[0025] The loss function is minimized to obtain the corresponding second fitting coefficient b;

[0026] The first fitting coefficient a and the third fitting coefficient c are determined according to the target concentration sequence and the second fitting coefficient: ;

[0027] wherein, represents a concentration value corresponding to the i-th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

[0028] Optionally, the first concentration sequence of methane, the second concentration sequence of ethylene, and the third concentration sequence of acetylene are obtained by:

[0029] A sliding time window is set, and an end time point of the sliding time window is a current time point;

[0030] At each time point in the sliding time window, a concentration value of methane, a concentration value of ethylene, and a concentration value of acetylene are collected respectively, to generate the first concentration sequence containing a plurality of concentration values of methane, the second concentration sequence containing a plurality of concentration values of ethylene, and the third concentration sequence containing a plurality of concentration values of acetylene.

[0031] In a second aspect, the present application further provides a fault prediction device, comprising:

[0032] The acquisition module is configured to acquire a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points.

[0033] The prediction module is configured to perform concentration prediction according to the first concentration sequence, the second concentration sequence and the concentration sequence respectively, to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene at a future time point.

[0034] The processing module is configured to perform fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a three-ratio method when at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value.

[0035] In a third aspect, the present application provides an electronic device, comprising a memory and a processor, the memory and the processor are connected to each other in communication, the memory stores computer instructions, and the processor executes the computer instructions to perform the fault prediction method of the first aspect or any of the corresponding embodiments thereof.

[0036] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions for making a computer execute the fault prediction method of the first aspect or any of the corresponding embodiments thereof.

[0037] The technical scheme provided by the embodiments of the present application comprises: acquiring a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points; performing concentration prediction according to the first concentration sequence, the second concentration sequence and the third concentration sequence respectively to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene; and when at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value, performing fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a three-ratio method.

[0038] The application has the beneficial effects that: compared with the current model with multiple input parameters, resulting in high calculation complexity and more consumed calculation resources, when predicting, the application respectively constructs the concentration sequences of multiple gases related to transformer failure, only needs to respectively predict based on each concentration sequence to determine the future predicted concentration value, and finally combines the three-ratio method to diagnose the transformer failure by using the future predicted concentration value of various gases. This way does not need to predict based on all data, does not need to construct a complex model with numerous parameters, and can simplify the prediction process; and by predicting the concentration value at multiple time points and combining the actual monitoring data, the accuracy and reliability of the fault diagnosis can be improved; by predicting the gas concentration value at the future time point, the abnormality can be found before the transformer actually fails, so that the time window of the fault warning is greatly advanced, the power outage accidents and losses caused by sudden transformer failure can be effectively avoided or reduced, and the stability and reliability of the power system can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the specific embodiments or related art, the following will briefly introduce the drawings needed to be used in the specific embodiments or related art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0040] Figure 1 A flowchart of a fault prediction method provided by an embodiment of the present application;

[0041] Figure 2 A principle schematic diagram of the Davis triangle method provided by an embodiment of the present application;

[0042] Figure 3 A flowchart of a fault prediction method provided by an embodiment of the present application;

[0043] Figure 4 A curve diagram of an exponential function relationship provided by an embodiment of the present application;

[0044] Figure 5 A structural schematic diagram of a fault prediction device provided by an embodiment of the present application;

[0045] Figure 6 A structural schematic diagram of a fault prediction device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] This invention provides a fault prediction method that can diagnose faults in transformers. Figure 1 A flowchart of a fault prediction method provided in an embodiment of the present invention is shown below. Figure 1 As shown, the process includes the following steps S101 to S103.

[0048] S101: Obtain the first concentration sequence of methane, the second concentration sequence of ethylene, and the third concentration sequence of acetylene; the first, second, and third concentration sequences all include concentration values ​​at multiple time points.

[0049] The execution subject of this embodiment is an electronic device. This electronic device can be a computer, mobile phone, etc. This embodiment is not limited to the subject acquiring the concentration sequence; any device capable of acquiring gas concentrations is acceptable. For example, multiple oil-immersed transformers can be managed uniformly by a single host computer, with the host computer acting as the subject. In this embodiment, when diagnosing faults in the transformer, to predict the transformer's faults, the concentration values ​​of various gases inside the transformer can be collected in real time. The gases involved include at least CH4 (methane), C2H4 (ethylene), and C2H2 (acetylene). In this embodiment, the concentration sequence is a sequence including concentration values ​​at multiple time points; wherein, the multiple time points include at least two time points.

[0050] It should be further explained that, in order to avoid collecting too much data, the above-mentioned obtaining the first concentration sequence of methane, the second concentration sequence of ethylene and the third concentration sequence of acetylene can include: setting a sliding time window, the end time point of the sliding time window being the current time point; collecting the concentration value of methane, the concentration value of ethylene and the concentration value of acetylene at each time point in the sliding time window, respectively, to generate the first concentration sequence containing a plurality of concentration values of methane, the second concentration sequence containing a plurality of concentration values of ethylene, and the third concentration sequence containing a plurality of concentration values of acetylene. The definition of the sliding time window in this embodiment is: assuming that the size of the window is T (for example, T can be 1 minute, 1 hour, 1 day, etc.), then the end time point of the window is always the current time point. The start time point of the window is the current time point minus T. This embodiment sets a sliding time window, and the concentration values collected in the sliding time window form the corresponding concentration sequence. Specifically, the concentration values of CH4, C2H4 and C2H2 can be determined at the corresponding sampling time points according to a certain sampling interval, and the concentration values collected at a plurality of time points can form a corresponding sequence. For example, the concentration values of CH4 collected at a plurality of time points can form a concentration sequence of methane, i.e., the first concentration sequence; similarly, the second concentration sequence of ethylene and the third concentration sequence of acetylene can be formed. Moreover, the end time point of the sliding time window is the current time point, so as to ensure that the collected concentration values are the latest. If the sliding time window contains n sampling time points in total, n concentration values can be collected in total, and the formed concentration sequence is also a one-dimensional sequence containing n elements. For example, if the sliding time window is 1 hour and the sampling frequency is 1 min, i.e., the concentration values of various gases are collected once every 1 min, then 60 concentration values can be collected in the sliding time window in total, i.e., n = 60. By setting the sliding time window, it is ensured that the collected concentration sequence is relatively new and has high timeliness, and it is also possible to avoid the influence of too old data on the accuracy of subsequent fault prediction, and to reduce the amount of data and improve the calculation efficiency.

[0051] S102: According to the first concentration sequence, the second concentration sequence and the third concentration sequence, concentration prediction is performed respectively to determine the first predicted concentration value of methane, the second predicted concentration value of ethylene and the third predicted concentration value of acetylene.

[0052] The embodiment is not limited to a specific prediction method, as long as the input is a concentration sequence and the output is a predicted concentration sequence. For example, the prediction method in the embodiment can be a time series model; or the prediction method in the embodiment can be a data fitting method; or the prediction method in the embodiment is a mathematical model. In this embodiment, the method of predicting each gas separately is used to realize the fault prediction of the transformer. Specifically, the concentration of methane is predicted according to the first concentration sequence, and the concentration value of methane at a future time point, i.e., the first predicted concentration value, can be predicted. Moreover, the concentration of ethylene is predicted according to the second concentration sequence, and the concentration value of ethylene at a future time point, i.e., the second predicted concentration value, can be predicted. The concentration of acetylene is predicted according to the third concentration sequence, and the concentration value of acetylene at a future time point, i.e., the third predicted concentration value, can be predicted. Since each gas is predicted separately, the mutual influence between gases can be reduced, and the problem of low prediction accuracy caused by too many types of data can be avoided. For example, for each gas, the concentration value can be predicted based on a time series model.

[0053] It should be further explained that based on any of the above embodiments, in order to reduce the complexity of calculation, the concentration prediction according to the first concentration sequence, the second concentration sequence, and the third concentration sequence, respectively, to determine the first predicted concentration value of methane, the second predicted concentration value of ethylene, and the third predicted concentration value of acetylene can include:

[0054] S1021, data fitting is performed according to a target concentration sequence corresponding to a target gas to determine a target function relationship between the concentration value of the target gas and time.

[0055] S1022, a target predicted concentration value corresponding to the target gas is determined according to the target function relationship.

[0056] The target gas in this embodiment is at least one of methane, ethylene, and acetylene; when the target gas is methane, the target concentration sequence is the first concentration sequence, and the target predicted concentration value is the first predicted concentration value; when the target gas is ethylene, the target concentration sequence is the second concentration sequence, and the target predicted concentration value is the second predicted concentration value; when the target gas is acetylene, the target concentration sequence is the third concentration sequence, and the target predicted concentration value is the third predicted concentration value. In this embodiment, data fitting can be used to establish a mathematical model to describe the relationship between data and predict future or unknown events. Since the mathematical model in this embodiment uses fewer parameters, the efficiency of prediction can be improved.

[0057] It needs to be further explained that the above-mentioned data fitting according to the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value of the target gas and the time can include: a first function relationship in the form of a polynomial is established for the target gas in advance, and the first function relationship is: ; wherein, represents the i-th time point, represents the concentration value predicted at the i-th time point, , , , are undetermined polynomial fitting coefficients, m is the index of the polynomial fitting coefficient, and can represent the number of polynomial fitting coefficients, wherein there are m+1 polynomial fitting coefficients; the first function relationship is data fitted according to the target concentration sequence corresponding to the target gas, and each polynomial fitting coefficient is determined to generate the target function relationship corresponding to the target gas; wherein the target function relationship is a polynomial function relationship formula for predicting the concentration value. The embodiment gives a specific way to determine the target function relationship, which improves the accuracy of the prediction method.

[0058] It needs to be further explained that the above-mentioned polynomial fitting coefficient is: ; wherein, represents the concentration value corresponding to the i-th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence. The embodiment gives a specific way to perform fitting, which improves the accuracy of fitting.

[0059] It needs to be further explained that the present application also provides another way to determine the target function relationship, and the above-mentioned data fitting according to the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value of the target gas and the time can include: a second function relationship in the form of an exponential function is established for the target gas in advance, and the second function relationship is: ; wherein, represents the i-th time point, represents the concentration value predicted at the i-th time point, a is the first undetermined fitting coefficient, b is the second undetermined fitting coefficient, and c is the third undetermined fitting coefficient;

[0060] According to the target concentration sequence corresponding to the target gas, the second function relationship is data fitted to determine the first fitting coefficient, the second fitting coefficient and the third fitting coefficient, and a target function relationship is generated; wherein the target function relationship is an exponential function relationship formula for predicting the concentration value. For each gas, the function relationship between the concentration value and the time can be determined by data fitting. After the function relationship is constructed, the concentration value corresponding to a future time point can be calculated, which is the predicted target concentration value. The embodiment gives a specific method for determining the exponential function relationship formula, and provides another prediction method.

[0061] It needs to be further explained that, based on any of the above embodiments, the second function relationship is data fitted according to the target concentration sequence corresponding to the target gas to determine the first fitting coefficient, the second fitting coefficient and the third fitting coefficient, and a target function relationship is generated, which can include: simplifying the error sum of squares between the real concentration value and the predicted concentration value to obtain a simplified loss function : ; minimizing the loss function to obtain the corresponding second fitting coefficient b; according to the target concentration sequence and the second fitting coefficient, the first fitting coefficient a and the third fitting coefficient c are determined: ; wherein represents the concentration value corresponding to the i-th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence. The embodiment gives a specific method for determining the fitting coefficient, which improves the accuracy of the fitting relationship determination.

[0062] S103: When at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds the corresponding preset threshold value, the transformer is diagnosed for fault by using the three-ratio method according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value.

[0063] The first predicted concentration value, the second predicted concentration value and the third predicted concentration value in the embodiment are the concentration values of each gas at a future time point, and the corresponding preset threshold value can be set for each gas in advance. If the concentration value of the gas exceeds the preset threshold value, it indicates that the transformer may have a fault and needs to be diagnosed for fault, thereby improving the efficiency of fault diagnosis. The embodiment does not limit the specific preset threshold value of each predicted concentration threshold value, which can be set according to the actual situation. For example, the preset threshold value in the embodiment can be 80, 100, etc. The three predicted threshold values can be the same or different. Therefore, if at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds the corresponding preset threshold value, it indicates that the transformer may have a fault at a future time point, so fault diagnosis can be performed in advance.

[0064] Among them, in order to realize fault detection of oil-immersed transformer, the characteristic gas method can be adopted; and in order to further improve the diagnosis speed and accuracy, the three-ratio method can be adopted as the fault diagnosis strategy. The three-ratio method, that is, based on the ratio of CH4 (methane), C2H4 (ethylene) and C2H2 (acetylene) three gas concentrations in the transformer, realizes fault diagnosis, and specifically can adopt the David triangle method.

[0065] Figure 2 A principle schematic diagram of a David triangle method is shown, the David triangle method judges the fault type according to CH4, C2H4 and C2H2 three gas concentrations, and three sides of the triangle respectively represent the relative proportions of CH4, C2H4 and C2H2 concentrations. When the gas concentration ratio point falls in which area when using the David triangle for diagnosis, the fault type corresponding to the area is the fault type corresponding to the transformer. For example, if at a future time, the first predicted concentration value is 100 ppm, the second predicted concentration value is 60 ppm, and the third predicted concentration value is 40 ppm, the proportion of methane is 50%, the proportion of ethylene is 30%, and the proportion of acetylene is 20%. Based on the proportion, the corresponding fault position can be determined, and the possible fault of the transformer at the future time point can be predicted in advance, realizing early warning.

[0066] The fault prediction method provided by the embodiment of the present application can include: S101, obtaining a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points; S102, respectively performing concentration prediction according to the first concentration sequence, the second concentration sequence and the third concentration sequence to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene; S103, when at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value, performing fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using the three-ratio method. The embodiment respectively constructs concentration sequences of multiple gases related to transformer faults, respectively performs prediction based on each concentration sequence to determine predicted concentration values in the future, and finally combines the three-ratio method to perform fault diagnosis on the transformer by using the predicted concentration values of various gases in the future. This method does not need to perform prediction based on all data and does not need to construct a complex model with numerous parameters, so that the prediction process can be simplified. Moreover, the concentration values at multiple time points are used for prediction, and the actual monitoring data are combined, so that the accuracy and reliability of fault diagnosis can be improved. By predicting the gas concentration values at future time points, abnormalities can be found before the transformer actually fails, so that the time window of fault early warning is greatly advanced, power outage accidents and losses caused by sudden transformer faults can be effectively avoided or reduced, and the stability and reliability of the power system can be improved.

[0067] Another fault prediction method is provided in the embodiment of the present application, which can realize fault diagnosis of the oil-immersed transformer. Figure 3 As shown in the flowchart of the fault prediction method according to the embodiment of the present application, the flowchart includes the following steps S301 to S303. Figure 3

[0068] Step S301: Obtain a first concentration sequence of methane, a second concentration sequence of ethylene and a concentration sequence of acetylene in the transformer; the first concentration sequence, the second concentration sequence and the concentration sequence each include concentration values at multiple time points.

[0069] Step S302: According to the first concentration sequence, the second concentration sequence and the concentration sequence, respectively perform concentration prediction to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene at a future time point.

[0070] Optionally, since the relationship between the concentration value and the time is generally nonlinear, polynomial fitting can be performed. The above step of "performing data fitting according to the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value and the time of the target gas" can include steps b1 to b2.

[0071] Step b1: A first function relationship in the form of a polynomial is pre-established for the target gas, and the first function relationship is: wherein, represents the i th time point, represents the concentration value predicted at the i th time point, , , , are to-be-determined polynomial fitting coefficients. m is the index of the polynomial fitting coefficient, which can represent the number of polynomial fitting coefficients, wherein there are m+1 polynomial fitting coefficients in total.

[0072] Step b2: Perform data fitting on the first function relationship according to the target concentration sequence to determine each polynomial fitting coefficient and generate a polynomial function relationship formula for predicting the concentration value.

[0073] In this embodiment, based on the to-be-determined polynomial fitting coefficients , , , ​Thus, a polynomial functional relationship, namely the first functional relationship, is constructed. Furthermore, using the concentration values ​​corresponding to each time point in the target concentration sequence, polynomial fitting can be performed to obtain the values ​​of the polynomial fitting coefficients. Therefore, the undetermined first functional relationship can be determined as a polynomial functional relationship with known coefficients, which represents the time points... With concentration value The relationship between them, therefore for a certain future point in time Based on this polynomial function relationship, the corresponding concentration value can be calculated. This concentration value This is the target predicted concentration value.

[0074] Alternatively, the values ​​of the fitting coefficients of each polynomial can be directly calculated based on matrix operations.

[0075] Among them, with Represents the i-th time point in the target concentration sequence. The corresponding concentration value means that any concentration value in the target concentration sequence can be represented as an array. To accurately represent the concentration values ​​in the target concentration sequence. With time point The functional relationship between them needs to be such that the actual concentration values ​​are... The concentration values ​​predicted based on functional relationships The sum of squared errors between them is minimized, and this sum of squared errors R 2 for: Minimize the sum of squared errors R 2 This allows us to obtain the optimal polynomial fitting coefficients. Furthermore, by taking the partial derivatives of any polynomial fitting coefficients, the sum of squared errors R... 2 The corresponding partial derivatives should all be 0. For the polynomial fitting coefficients... Find the partial derivatives:

[0076] The fitting coefficients for the polynomial Find the partial derivatives: Similarly, the final polynomial fitting coefficients... Find the partial derivatives: The above m+1 equations can be represented in matrix form as follows: (1) (1)

[0077] Therefore, these polynomial fitting coefficients can be represented in matrix form as follows (2):

[0078] (2)

[0079] Here, -1 represents the inverse operation of the matrix.

[0080] In the embodiment of the present application, the polynomial form function relationship is fitted based on the target concentration sequence, which can accurately fit the relationship between the concentration value and the time, so as to predict the concentration value at the future time point in advance.

[0081] It can be understood that when the polynomial fitting is adopted, a large number of fitting coefficients are generally required for fitting, and the matrix operation has a relatively large amount of calculation. The three function relationships corresponding to methane, ethylene and acetylene are fitted respectively, which is time-consuming and is not conducive to online real-time monitoring. In the embodiment of the present application, the function relationship between the concentration value and the time is established by using an exponential function, which can be represented based on fewer coefficients. Specifically, the above step of "performing data fitting on the target concentration sequence corresponding to the target gas to determine the target function relationship between the concentration value and the time of the target gas" can include steps c1 to c2.

[0082] Step c1, a second function relationship in exponential form is established for the target gas, and the second function relationship is: ; wherein, represents the i-th time point, represents the concentration value predicted at the i-th time point, a is a first fitting coefficient to be determined, b is a second fitting coefficient to be determined, and c is a third fitting coefficient to be determined. Step c2, data fitting is performed on the second function relationship according to the target concentration sequence to determine the first fitting coefficient, the second fitting coefficient and the third fitting coefficient, and an exponential function relationship formula for predicting the concentration value is generated. In the embodiment of the present application, the second function relationship in exponential form is fitted, only three fitting coefficients, i.e. the first fitting coefficient a, the second fitting coefficient b and the third fitting coefficient c, need to be determined. After the three fitting coefficients are fitted, the function relationship in exponential form, i.e. the exponential function relationship formula, is obtained, and the prediction of the concentration value is realized.

[0083] It can be understood that since methane, ethylene and acetylene need to be fitted, the second function relationship in exponential form can be used for all three gases, or a part of the second function relationship in exponential form and a part of the second function relationship in polynomial form can be used, which is not limited in the embodiment.

[0084] Optionally, since the second fitting coefficient b in the second function relationship in exponential form is located in the exponential position, this coefficient cannot be directly calculated, so it is generally necessary to calculate the optimal fitting coefficient through multiple rounds of iterative calculation based on the least square method or other iterative methods, which also affects the calculation efficiency. In the embodiment of the present application, the optimal fitting coefficient, including the optimal first fitting coefficient a, the optimal second fitting coefficient b and the optimal third fitting coefficient c, can be quickly calculated by optimizing the loss function of the concentration value.

[0085] Specifically, as shown above, In the embodiments of the present application, the total loss between the actual concentration value and the predicted concentration value is determined by the sum of squares of errors, and a corresponding loss function is established, which is as follows (3):

[0086] (3)

[0087] The above formula (3) is decomposed to obtain:

[0088] The partial derivatives of the decomposed formula with respect to the first fitting coefficient a and the third fitting coefficient c are obtained as follows (4) and (5):

[0089] (4)

[0090] (5)

[0091] Using the above formulas (4) and (5), the decomposed formula (3) is represented to obtain:

[0092] (6)

[0093] Similar to the above polynomial fitting process, to achieve the optimal solution, the loss function L needs to be minimized, and correspondingly, the partial derivatives of the first fitting coefficient a and the third fitting coefficient c are both 0, that is, the above formulas (4) and (5) are both 0, so the loss function L represented by the above formula (6) can be simplified as:

[0094] (7)

[0095] And since the corresponding values of the above formulas (4) and (5) are 0, based on this, the first fitting coefficient a and the third fitting coefficient c can be calculated as follows: (8)

[0096] Substituting the third fitting coefficient c calculated by the above formula (8) into the simplified loss function Ls, that is, the above formula (7), the simplified loss function Ls can be calculated as follows: (9)

[0097] Then, substituting the first fitting coefficient a in the above formula (8) into the above formula (9), the simplified loss function Ls is as follows: (10)

[0098] ​​The loss function Ls shown in the above formula (10) is only related to the second fitting coefficient b, and thus the second fitting coefficient b that can minimize the loss function Ls shown in the above formula (10) needs to be found. Since the latter half of the above formula (10) is a fixed value (each is known), which does not affect the minimization process of the above formula (10), the loss function related to the second fitting coefficient b can finally be simplified as: (11)

[0099] In summary, the solution of the second fitting coefficient b is equivalent to the minimization solution of the above formula (11), that is, to determine the second fitting coefficient b that minimizes the above formula (11). For example, the golden section method can be used to determine the second fitting coefficient, and the specific way of solving the above formula (11) in this embodiment is not limited.

[0100] After the second fitting coefficient b is calculated, the first fitting coefficient a and the third fitting coefficient c can be calculated based on the above formula (8).

[0101] Taking the methane gas of a transformer as an example, Figure 4 is a curve diagram of the exponential function relationship provided by an embodiment of the present application, with the horizontal coordinate being time (min) and the vertical coordinate being concentration value (ppm). Based on the method provided by this embodiment, a=0.9866, b=0.0378, and c=3.2608 are fitted. If the preset threshold of methane is 80 ppm, based on the exponential function relationship, it can be determined that when the predicted concentration of methane at the future time point is 115, the first predicted concentration value of methane exceeds 80 ppm, and thus at this time, the first predicted concentration value and the second predicted concentration value and the third predicted concentration value of other gases can be used for fault diagnosis based on the prediction.

[0102] Step S303: In the case where at least one of the first predicted concentration value, the second predicted concentration value, and the third predicted concentration value exceeds the corresponding preset threshold, the transformer is diagnosed for fault according to the first predicted concentration value, the second predicted concentration value, and the third predicted concentration value by using the three-ratio method.

[0103] The detailed description of this step can be found in the step S103 of the embodiment shown in Figure 1 , which will not be repeated here.

[0104] The fault prediction method provided by the embodiment of the present application does not need to be based on all data for prediction, does not need to construct a complex model with numerous parameters, can simplify the prediction process, and guarantees the stability and reliability of the power system. In a polynomial or exponential fitting manner, the functional relationship between the concentration value and the time is determined, so that the gas concentration value at a future time point can be predicted in real time based on the functional relationship, without training a prediction model, and the implementation is simple; when expressed in an exponential form, the loss function with multiple coefficients is simplified into a function with only a second fitting coefficient by simplifying the loss function, so that the optimal second fitting coefficient can be quickly calculated based on the golden section method, and then the first fitting coefficient and the third fitting coefficient are calculated, the fitting of the exponential function can be quickly realized, the calculation efficiency is guaranteed, and the online real-time monitoring scene of the transformer is suitable.

[0105] In the embodiment, a fault prediction device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. The terms "module" and "unit" can be a combination of software and / or hardware that realizes a predetermined function. Although the device described in the following embodiments is preferably realized in software, the realization of hardware or a combination of software and hardware is also possible and conceived.

[0106] The embodiment provides a fault prediction device, as shown in Figure 5 , and Figure 5 The structure schematic diagram of the fault prediction device provided by the embodiment of the present application comprises:

[0107] The acquisition module 100 is used to acquire a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence all include concentration values at multiple time points;

[0108] The prediction module 200 is used to respectively perform concentration prediction according to the first concentration sequence, the second concentration sequence and the concentration sequence, to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene at a future time point.

[0109] The processing module 300 is used to perform fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using the three-ratio method in a case that at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value.

[0110] Further, based on any of the above-mentioned embodiments, the prediction module 200 can comprise:

[0111] The target function relationship determining unit is configured to perform data fitting according to the target concentration sequence corresponding to the target gas, and determine a target function relationship between the concentration value of the target gas and time.

[0112] The target predicted concentration value determining unit is configured to determine a target predicted concentration value corresponding to the target gas according to the target function relationship.

[0113] The target gas is at least one of methane, ethylene, and acetylene.

[0114] When the target gas is methane, the target concentration sequence is a first concentration sequence, and the target predicted concentration value is a first predicted concentration value.

[0115] When the target gas is ethylene, the target concentration sequence is a second concentration sequence, and the target predicted concentration value is a second predicted concentration value.

[0116] When the target gas is acetylene, the target concentration sequence is a third concentration sequence, and the target predicted concentration value is a third predicted concentration value.

[0117] Further, based on any of the above embodiments, the target function relationship determining unit can include:

[0118] The first function relationship determining subunit is configured to pre-establish a first function relationship in a polynomial form for the target gas, and the first function relationship is: ; wherein, represents the i-th time point, represents a predicted concentration value at the i-th time point, , , , are undetermined polynomial fitting coefficients.

[0119] The data fitting subunit is configured to perform data fitting on the first function relationship according to the target concentration sequence corresponding to the target gas, determine each polynomial fitting coefficient, and generate a target function relationship corresponding to the target gas; wherein the target function relationship is a polynomial function relationship formula for predicting the concentration value.

[0120] Further, based on the above embodiment, the polynomial fitting coefficient is: ; wherein, represents a concentration value corresponding to the i-th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

[0121] Further, based on any of the above embodiments, the target function relationship determining unit can include:

[0122] The second function relationship determining subunit is configured to pre-establish a second function relationship in an exponential form for the target gas, and the second function relationship is: ; wherein, represents the ith time point, represents a predicted concentration value at the ith time point, a is a first fitting coefficient to be determined, b is a second fitting coefficient to be determined, and c is a third fitting coefficient to be determined.

[0123] The second target function relationship determining subunit is configured to perform data fitting on the second function relationship according to a target concentration sequence corresponding to the target gas, to determine the first fitting coefficient, the second fitting coefficient, and the third fitting coefficient, and to generate the target function relationship; wherein the target function relationship is an exponential function relationship formula for predicting a concentration value.

[0124] Further, based on any of the above embodiments, the second target function relationship determining subunit can include:

[0125] The loss function determining subunit is configured to simplify an error sum between a real concentration value and a predicted concentration value to obtain a simplified loss function :

[0126] ;

[0127] The first fitting coefficient determining subunit is configured to minimize the loss function to obtain a corresponding second fitting coefficient b.

[0128] The second fitting coefficient determining subunit is configured to determine the first fitting coefficient a and the third fitting coefficient c according to the target concentration sequence and the second fitting coefficient:

[0129] ;

[0130] wherein, represents a concentration value corresponding to the ith time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

[0131] Further, based on any of the above embodiments, the acquisition module 100 can include:

[0132] The sliding time window determining unit is configured to set a sliding time window, and an ending time point of the sliding time window is a current time point.

[0133] The concentration value acquisition unit is configured to acquire the concentration value of methane, the concentration value of ethylene and the concentration value of acetylene at each time point within the sliding time window respectively, and generate the first concentration sequence containing a plurality of concentration values of methane, the second concentration sequence containing a plurality of concentration values of ethylene, and the third concentration sequence containing a plurality of concentration values of acetylene.

[0134] Further function descriptions of the above-mentioned modules and units are the same as those of the above-mentioned corresponding embodiments, and will not be described here again.

[0135] The fault prediction device in the embodiment is in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, including a processor and a memory for executing one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0136] The fault prediction device provided by the embodiment of the application can comprise: an acquisition module 100 configured to acquire a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each comprise concentration values of a plurality of time points; a prediction module 200 configured to perform concentration prediction according to the first concentration sequence, the second concentration sequence and the concentration sequence respectively, and determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene at a future time point; and a processing module 300 configured to perform fault diagnosis on a transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a three-ratio method in a case where at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value. Compared with the current prediction method of using one model to input a plurality of parameters for prediction, the prediction method of the embodiment reduces the number of input parameters by performing prediction based on each concentration sequence respectively, so as to combine the three-ratio method and use the predicted concentration values of various gases in the future to perform fault diagnosis. This method does not need to perform prediction based on all data, and does not need to construct a complex model with a large number of parameters, so that the prediction process can be simplified.

[0137] Next, a fault prediction device provided by the embodiment of the application will be described. The fault prediction device described below can be referred to in combination with the fault prediction method described above.

[0138] Please refer to Figure 6 , Figure 6 A structural schematic diagram of the fault prediction device provided by the embodiment of the application can comprise:

[0139] a memory 10 for storing a computer program;

[0140] a processor 20 for executing the computer program to implement the failure prediction method described above.

[0141] The memory 10, the processor 20 and the communication interface 30 all communicate with each other through a communication bus 40.

[0142] In the embodiment of the present application, the memory 10 stores one or more programs, which can include program codes including computer operation instructions. In the embodiment of the present application, the memory 10 can store programs for implementing the following functions:

[0143] obtain a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points;

[0144] perform concentration prediction according to the first concentration sequence, the second concentration sequence and the third concentration sequence to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene;

[0145] when at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value, perform failure diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a three-ratio method.

[0146] In a possible implementation, the memory 10 can include a program storage area and a data storage area, where the program storage area can store an operating system and at least one application program required by a function, etc.; and the data storage area can store data created during use.

[0147] In addition, the memory 10 can include a read-only memory and a random access memory, and provide instructions and data for the processor. A part of the memory can also include an NVRAM. The memory stores an operating system and operation instructions, executable modules or data structures, or subsets thereof, or an extended set thereof, where the operation instructions can include various operation instructions for implementing various operations. The operating system can include various system programs for implementing various basic tasks and processing hardware-based tasks.

[0148] The processor 20 can be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic device. The processor 20 can be a microprocessor or any conventional processor, etc. The processor 20 can invoke programs stored in the memory 10.

[0149] The communication interface 30 can be an interface of a communication module, used for connecting with other devices or systems.

[0150] Of course, it should be noted that, Figure 6 The structure shown does not constitute a limitation on the fault prediction device in the embodiments of the present application, and in actual applications, the fault prediction device can include more or fewer components than Figure 6 those shown, or some components can be combined.

[0151] The computer readable storage medium provided by the embodiments of the present application is described below, and the computer readable storage medium described below can be referred to in conjunction with the fault prediction method described above.

[0152] The present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the fault prediction method described above.

[0153] The computer readable storage medium can include a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0154] Each embodiment in the specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0155] The skilled person can further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been described in the above description. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0156] Finally, it needs to be explained that, in this document, relationships such as first and second, and the like, are intended to distinguish one entity or action from another, without necessarily requiring or implying any actual such relationship or order between entities or actions. Also, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0157] The above describes in detail the fault prediction method, device, equipment and computer readable storage medium provided by the present application. The principles and implementation manners of the present application are described by applying specific examples in this document. The above example is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A failure prediction method characterized by, Applied to an oil-immersed transformer, comprising: obtaining a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values at multiple time points; According to the first concentration sequence, the second concentration sequence and the third concentration sequence, respectively, the first predicted concentration value of methane, the second predicted concentration value of ethylene and the third predicted concentration value of acetylene are determined by concentration prediction; When at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds the corresponding preset threshold value, the transformer is diagnosed for fault according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using the David triangle method; According to the first concentration sequence, the second concentration sequence and the third concentration sequence, respectively, the first predicted concentration value of methane, the second predicted concentration value of ethylene and the third predicted concentration value of acetylene are determined by concentration prediction, including: According to the target concentration sequence corresponding to the target gas, the target function relationship between the concentration value and the time of the target gas is determined by data fitting; According to the target function relationship, the target predicted concentration value corresponding to the target gas is determined; The target gas is at least one of methane, ethylene and acetylene; When the target gas is methane, the target concentration sequence is the first concentration sequence, and the target predicted concentration value is the first predicted concentration value; When the target gas is ethylene, the target concentration sequence is the second concentration sequence, and the target predicted concentration value is the second predicted concentration value; When the target gas is acetylene, the target concentration sequence is the third concentration sequence, and the target predicted concentration value is the third predicted concentration value; According to the target concentration sequence corresponding to the target gas, the target function relationship between the concentration value and the time of the target gas is determined by data fitting, including: a second functional relationship in the form of an exponential function is pre-established for the target gas, the second functional relationship being ; wherein denotes the i-th time point, denotes the predicted concentration value at the i-th time point, a is a first fitting coefficient to be determined, b is a second fitting coefficient to be determined, and c is a third fitting coefficient to be determined; According to the target concentration sequence corresponding to the target gas, the second function relationship is fitted by data fitting to determine the first fitting coefficient, the second fitting coefficient and the third fitting coefficient, and the target function relationship is generated; wherein the target function relationship is an exponential function relationship formula for predicting the concentration value; According to the target concentration sequence corresponding to the target gas, the second function relationship is fitted by data fitting to determine the first fitting coefficient, the second fitting coefficient and the third fitting coefficient, and the target function relationship is generated, including: sum of squared errors between the true concentration values and the predicted concentration values is simplified to a simplified loss function : ; minimizing the loss function , to obtain a corresponding second fitting coefficient b; determining the first fitting coefficient a and the third fitting coefficient c according to the target concentration sequence and the second fitting coefficient: ; wherein, denotes a concentration value corresponding to an i-th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

2. The failure prediction method according to claim 1, characterized by, According to the target concentration sequence corresponding to the target gas, the target function relationship between the concentration value and the time of the target gas is determined by data fitting, including: A first function relationship in polynomial form is established in advance for the target gas, and the first function relationship is ; wherein, represents the ith time point, represents the predicted concentration value at the ith time point, , , , are all undetermined polynomial fitting coefficients. According to the target concentration sequence corresponding to the target gas, the first function relationship is fitted by data fitting to determine each polynomial fitting coefficient, and the target function relationship corresponding to the target gas is generated; wherein the target function relationship is a polynomial function relationship formula for predicting the concentration value.

3. The failure prediction method according to claim 2, characterized by, The polynomial fitting coefficient is: ; wherein, represents the concentration value corresponding to the i th time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

4. The failure prediction method according to any one of claims 1 to 3, characterized by, The first concentration sequence of methane, the second concentration sequence of ethylene and the third concentration sequence of acetylene are obtained, including: A sliding time window is set, and an end time point of the sliding time window is a current time point; At each time point in the sliding time window, a concentration value of methane, a concentration value of ethylene and a concentration value of acetylene are collected respectively, and the first concentration sequence containing a plurality of concentration values of methane, the second concentration sequence containing a plurality of concentration values of ethylene, and the third concentration sequence containing a plurality of concentration values of acetylene are generated.

5. A failure prediction device characterized by comprising: Applied to an oil-immersed transformer, comprising: An acquisition module is configured to acquire a first concentration sequence of methane, a second concentration sequence of ethylene and a third concentration sequence of acetylene; the first concentration sequence, the second concentration sequence and the third concentration sequence each include concentration values of a plurality of time points; A prediction module is configured to perform concentration prediction according to the first concentration sequence, the second concentration sequence and the concentration sequence respectively, to determine a first predicted concentration value of methane, a second predicted concentration value of ethylene and a third predicted concentration value of acetylene at a future time point; A processing module is configured to perform fault diagnosis on the transformer according to the first predicted concentration value, the second predicted concentration value and the third predicted concentration value by using a David triangle method in a case where at least one of the first predicted concentration value, the second predicted concentration value and the third predicted concentration value exceeds a corresponding preset threshold value. The prediction module is configured to perform: data fitting according to a target concentration sequence corresponding to a target gas to determine a target function relationship between a concentration value of the target gas and time; and determining a target predicted concentration value corresponding to the target gas according to the target function relationship; wherein the target gas is at least one of methane, ethylene, and acetylene; when the target gas is methane, the target concentration sequence is the first concentration sequence, and the target predicted concentration value is the first predicted concentration value; when the target gas is ethylene, the target concentration sequence is the second concentration sequence, and the target predicted concentration value is the second predicted concentration value; and when the target gas is acetylene, the target concentration sequence is the third concentration sequence, and the target predicted concentration value is the third predicted concentration value; and data fitting according to a target concentration sequence corresponding to a target gas to determine a target function relationship between a concentration value of the target gas and time includes: pre-establishing a second function relationship in an exponential form for the target gas, the second function relationship is ; wherein, represents the ith time point, represents a predicted concentration value at the ith time point, a is a first fitting coefficient to be determined, b is a second fitting coefficient to be determined, and c is a third fitting coefficient to be determined; data fitting according to a target concentration sequence corresponding to a target gas to determine the first fitting coefficient, the second fitting coefficient, and the third fitting coefficient, and generating the target function relationship; wherein the target function relationship is an exponential function relationship for predicting a concentration value; wherein data fitting according to a target concentration sequence corresponding to a target gas to determine the first fitting coefficient, the second fitting coefficient, and the third fitting coefficient, and generating the target function relationship includes: simplifying an error sum of squares between a real concentration value and a predicted concentration value to obtain a simplified loss function : ; minimizing the loss function to obtain a corresponding second fitting coefficient b; and determining the first fitting coefficient a and the third fitting coefficient c according to the target concentration sequence and the second fitting coefficient: ; wherein, represents a concentration value corresponding to the ith time point in the target concentration sequence, and n is the number of concentration values in the target concentration sequence.

6. An electronic device, comprising: Comprising: A memory and a processor are communicatively connected between each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the fault prediction method in any one of claims 1 to 4.

7. A computer readable storage medium characterized in that, The computer readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the fault prediction method in any one of claims 1 to 4.

Images