A method for detecting root cause of oscillation in a solid oxide fuel cell system

Abstract

A method for detecting the root cause of oscillation of a solid oxide fuel cell system, the method comprising the steps of: obtaining a preset number of solid oxide fuel cell system oscillation process variables; screening an oscillation analysis subset according to the preset number of solid oxide fuel cell system oscillation process variables; obtaining the optimal parameters of an oscillation root cause detection algorithm according to the oscillation analysis subset; determining the causal relationship between each process variable in the oscillation analysis subset according to the optimal parameters; and determining the oscillation root cause of the solid oxide fuel cell system according to the causal relationship. The method for detecting the root cause of oscillation of a solid oxide fuel cell system provided by the present application can accurately and efficiently detect the root cause of system oscillation, solving the technical problem that the efficiency of locating the root cause of oscillation of a solid oxide fuel cell system is low and affects the operation of the system.

Description

Technical Field

[0001] The present invention belongs to the technical field of solid oxide fuel cell systems, and in particular relates to a method for detecting oscillation root causes of solid oxide fuel cell systems. Background Art

[0002] Solid oxide fuel cell (SOFC) is a power generation technology that directly converts the chemical energy in fuel into electrical energy through high-temperature electrochemical reactions. Compared with traditional fuel power generation, SOFC technology greatly reduces the energy loss of fossil fuels in energy conversion, avoids the emission of sulfur oxides and nitrogen oxides, and significantly reduces carbon dioxide emissions. The power generation process of fuel cells is not limited by the Carnot cycle and has the advantages of high-efficiency power generation, low carbon and environmental protection. In the actual power generation process, due to the instability of factors such as temperature, pressure, and flow, the corresponding electrochemical reaction will be affected, thereby reducing the performance of the SOFC system.

[0003] The SOFC system has a strong coupling characteristic. A SOFC system consists of multiple subsystems, including a SOFC stack, a reforming subsystem, a gas supply unit, and an exhaust gas combustion subsystem. Each subsystem contains many variables, and there is also a coupling relationship between the variables. Therefore, when an oscillation occurs somewhere, the oscillation signal can easily propagate in multiple control loops, which will affect other parts or even the entire system, causing system failures and thus affecting the performance and life of the entire system. For example, a sudden change in the properties of the SOFC stack outlet gas (temperature, flow rate, composition, water-carbon ratio, etc.) will cause the temperature of the exhaust combustion chamber to vibrate, causing changes in the combustion performance of the combustion chamber. The change in combustion performance further causes a sharp rise and fall in the temperature inside the combustion chamber. As the high-temperature heat source peak component of the entire SOFC system, its temperature jitter will indirectly affect the temperature of the heat exchanger and the stack along the gas circuit, thereby affecting the operating characteristics of the system.

[0004] Therefore, in order to reduce the performance degradation caused by SOFC system oscillation and extend the service life of the SOFC system, it is crucial to detect the root cause of SOFC system oscillation. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for detecting the root cause of oscillation in a solid oxide fuel cell system, which overcomes the above problems or at least partially solves the above problems.

[0006] In order to solve the above technical problems, the present invention provides a method for detecting the root cause of oscillation of a solid oxide fuel cell system, the method comprising the steps of:

[0007] obtaining a preset number of solid oxide fuel cell system oscillation process variables;

[0008] Filtering an oscillation analysis subset according to a preset number of oscillation process variables of the solid oxide fuel cell system;

[0009] Obtaining optimal parameters of an oscillation root cause detection algorithm according to the oscillation analysis subset;

[0010] determining a cause-and-effect relationship between each process variable in the oscillation analysis subset according to the optimal parameters;

[0011] A root cause of oscillation of the solid oxide fuel cell system is determined based on the cause-effect relationship.

[0012] Preferably, the step of obtaining a preset number of solid oxide fuel cell system oscillation process variables comprises the steps of:

[0013] Disposing a measured variable detection sensor in the solid oxide fuel cell system;

[0014] Disposing an operating variable detection sensor in the solid oxide fuel cell system;

[0015] Acquiring a first number of measured variables when the solid oxide fuel cell system oscillates by means of the measured variable detection sensor;

[0016] Acquiring, by means of the operating variable detection sensor, a second number of operating variables when the solid oxide fuel cell system oscillates;

[0017] A first number of the measured variables and a second number of the manipulated variables are combined as the oscillating process variable.

[0018] Preferably, the screening of the oscillation analysis subset according to a preset number of the solid oxide fuel cell system oscillation process variables comprises the steps of:

[0019] Acquire a preset number of the oscillation process variables;

[0020] Calculate the principal component coefficient and variance of each of the oscillatory process variables using principal component analysis;

[0021] Arrange the variances in descending order and calculate the cumulative variance contribution rate;

[0022] Deleting the oscillation process variable corresponding to the cumulative variance contribution rate greater than a preset value from all the oscillation process variables and obtaining a first residual oscillation process variable;

[0023] Calculating the oscillation significance index of each of the remaining oscillation process variables according to the principal component coefficient and variance corresponding to each of the oscillation process variables in the first remaining oscillation process variables;

[0024] Deleting the oscillation process variable corresponding to the oscillation significance index less than a preset value from the first residual oscillation process variable and obtaining a second residual oscillation process variable;

[0025] All of the oscillating process variables in the second residual oscillating process variables are grouped into the oscillation analysis subset.

[0026] Preferably, the expression of the cumulative variance contribution rate is:

[0027]

[0028] Among them, λ i represents the variance, m represents the number of oscillation process variables in the first residual oscillation process variables, n represents the number of oscillation process variables in the preset number of oscillation process variables, and i represents the sequence number of the oscillation process variables.

[0029] Preferably, the step of obtaining the optimal parameters of the oscillation root cause detection algorithm according to the oscillation analysis subset comprises the steps of:

[0030] Obtaining the oscillation analysis subset;

[0031] Obtaining an oscillation analysis subset reconstructed phase space corresponding to the oscillation analysis subset;

[0032] Calculating an optimal time delay for reconstructing phase space of the oscillation analysis subset;

[0033] Calculating the optimal embedding dimension of the oscillation analysis subset reconstructed phase space according to the optimal time delay;

[0034] The optimal time delay and the optimal embedding dimension are used as the optimal parameters.

[0035] Preferably, determining the causal relationship between the process variables in the oscillation analysis subset according to the optimal parameters comprises the steps of:

[0036] Obtaining an optimal time delay and an optimal embedding dimension among the optimal parameters;

[0037] obtaining the oscillation process variable;

[0038] Reconstructing a reconstructed phase space that is topologically equivalent to the oscillating process variable according to the optimal time delay and the optimal embedding dimension;

[0039] Calculating predicted values ​​of the oscillating process variables obtained by cross-mapping according to the reconstructed phase space;

[0040] Obtaining observation values ​​of the reconstructed phase space;

[0041] Calculating the correlation coefficient between the predicted value and the observed value;

[0042] quantifying the strength of the causal relationship between the oscillating process variables according to the correlation coefficient and obtaining a convergent cross-mapping causal relationship diagram;

[0043] The causal relationship between the oscillation process variables in the oscillation analysis subset is determined according to the converged cross-map causal relationship diagram.

[0044] Preferably, the expression of the reconstructed phase space is:

[0045]

[0046] Wherein, x(t) represents the coordinate of the reconstructed phase space, X() represents the time series, t represents the time,

[0047] τ represents the sampling interval, and E represents the embedding dimension of the reconstructed phase space.

[0048] Preferably, the expression of the predicted value is:

[0049]

[0050] in, represents the predicted value, w i Represents the reconstructed phase space M X superior

[0051] The distance weight between the coordinate x(t) and its i-th neighboring point, E represents the embedding dimension of the reconstructed phase space.

[0052] Preferably, the expression of the correlation coefficient is:

[0053]

[0054] in, ρ represents the correlation coefficient, Y(i) represents the time series, represents the mean value of the time series, represents the predicted value, represents the average value of the predicted value, L represents the length of the time series, and i represents the sequence number of the adjacent point in the reconstructed phase space.

[0055] Preferably, the determining the root cause of the oscillation of the solid oxide fuel cell system according to the cause-effect relationship comprises the steps of:

[0056] Obtain a convergent cross-mapping causal relationship graph corresponding to the causal relationship;

[0057] acquiring process knowledge of the solid oxide fuel cell system;

[0058] The oscillation root cause is detected based on the converged cross-map causal relationship graph and the process knowledge.

[0059] One or more technical solutions in the embodiments of the present invention have at least the following technical effects or advantages: a solid oxide fuel cell system oscillation root cause detection method provided by the present application, in order to evaluate the performance of the SOFC system and find the root cause of the system oscillation, each process variable in the system is derived from the complete power generation process in the experiment, so that the fault analysis result is true and reliable; in order to narrow the scope of causal analysis and improve the accuracy of fault source identification, the fault variables in the original data are screened and an oscillation analysis subset is obtained before performing fault root cause detection; the variables in the oscillation analysis subset are causally analyzed using a convergent cross-mapping algorithm, and the root cause of the system oscillation is detected from the causal analysis results, which provides an accurate and effective technical solution for solving the root cause detection of the oscillation root cause of the solid oxide fuel cell system, can accurately and efficiently detect the root cause of the system oscillation, and solves the technical problem of low efficiency in locating the root cause of the oscillation of the solid oxide fuel cell system and affecting the operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 It is a flow chart of a method for detecting root cause of oscillation of a solid oxide fuel cell system provided by an embodiment of the present invention;

[0062] Figure 2 It is a schematic diagram of the effect of an oscillation analysis subset in a method for detecting oscillation root causes of a solid oxide fuel cell system provided by an embodiment of the present invention;

[0063] Figure 3 is a schematic diagram of an optimal time delay in a method for detecting root causes of oscillations in a solid oxide fuel cell system provided by an embodiment of the present invention;

[0064] Figure 4 is a schematic diagram of an optimal embedding dimension in a method for detecting root causes of oscillations in a solid oxide fuel cell system provided by an embodiment of the present invention;

[0065] Figure 5 It is a schematic diagram of a convergent cross-mapping causal relationship diagram in a solid oxide fuel cell system oscillation root cause detection method provided by an embodiment of the present invention;

[0066] Figure 6It is a structural schematic diagram of a solid oxide fuel cell system in a solid oxide fuel cell system oscillation root cause detection method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The present invention will be described in detail below in conjunction with specific implementations and examples, and the advantages and various effects of the present invention will be more clearly presented. It should be understood by those skilled in the art that these specific implementations and examples are used to illustrate the present invention, rather than to limit the present invention.

[0068] Throughout the specification, unless otherwise specifically stated, the terms used herein should be understood as meanings commonly used in the art. Therefore, unless otherwise defined, all technical and scientific terms used herein have the same meanings as those generally understood by those skilled in the art to which the present invention belongs. In the event of a conflict, the present specification takes precedence.

[0069] Unless otherwise specified, various raw materials, reagents, instruments and equipment used in the present invention can be purchased from the market or prepared by existing methods.

[0070] like Figure 1 In an embodiment of the present application, the present invention provides a method for detecting the root cause of oscillation of a solid oxide fuel cell system, the method comprising the steps of:

[0071] S1: obtaining a preset number of solid oxide fuel cell system oscillation process variables;

[0072] In an embodiment of the present application, the step of obtaining a preset number of solid oxide fuel cell system oscillation process variables comprises the following steps:

[0073] Disposing a measured variable detection sensor in the solid oxide fuel cell system;

[0074] Disposing an operating variable detection sensor in the solid oxide fuel cell system;

[0075] Acquiring a first number of measured variables when the solid oxide fuel cell system oscillates by means of the measured variable detection sensor;

[0076] Acquiring, by means of the operating variable detection sensor, a second number of operating variables when the solid oxide fuel cell system oscillates;

[0077] A first number of the measured variables and a second number of the manipulated variables are combined as the oscillating process variable.

[0078] In the embodiments of the present application, Figure 6For the solid oxide fuel cell system in this application, it should be noted that there are a large number of process variables in the solid oxide fuel cell (SOFC) system, and these process variables are detected and obtained by various sensors set inside the solid oxide fuel cell system. There are many process variables involved in the entire power generation process of the SOFC system, including 36 measurement variables and 8 operation variables, each process variable corresponds to a sensor for acquisition, that is, 36 measurement variable detection sensors are arranged in the solid oxide fuel cell system for corresponding detection of 36 measurement variables, and 8 operation variable detection sensors are also arranged in the solid oxide fuel cell system for corresponding detection of 8 operation variables. In this application, when a preset number of solid oxide fuel cell system oscillation process variables are obtained, 24 variables (the number of measurement variables and operation variables can be selected as needed) when the solid oxide fuel cell system oscillates are selected to constitute the oscillation process variables of the main control loop of the SOFC system, such as gas flow, pressure and internal temperature measurement data of each component and electrical characteristic data.

[0079] S2: screening an oscillation analysis subset according to a preset number of oscillation process variables of the solid oxide fuel cell system;

[0080] In an embodiment of the present application, the screening of the oscillation analysis subset according to a preset number of oscillation process variables of the solid oxide fuel cell system comprises the steps of:

[0081] Acquire a preset number of the oscillation process variables;

[0082] Calculate the principal component coefficient and variance of each of the oscillatory process variables using principal component analysis;

[0083] Arrange the variances in descending order and calculate the cumulative variance contribution rate;

[0084] Deleting the oscillation process variable corresponding to the cumulative variance contribution rate greater than a preset value from all the oscillation process variables and obtaining a first residual oscillation process variable;

[0085] Calculating the oscillation significance index of each of the remaining oscillation process variables according to the principal component coefficient and variance corresponding to each of the oscillation process variables in the first remaining oscillation process variables;

[0086] Deleting the oscillation process variable corresponding to the oscillation significance index less than a preset value from the first residual oscillation process variable and obtaining a second residual oscillation process variable;

[0087] All of the oscillating process variables in the second residual oscillating process variables are grouped into the oscillation analysis subset.

[0088] In the embodiment of the present application, considering that the control loop of the SOFC system is complex and involves many process variables, the oscillation signal is easy to propagate in multiple control loops, so it is difficult to directly determine the root cause of the oscillation of the SOFC system. Therefore, it is necessary to apply the principal component analysis method to calculate the oscillation significance index and filter out the oscillation analysis subset from the original oscillation process variable data to improve the accuracy of oscillation source identification.

[0089] Specifically, when the oscillation analysis subset is selected according to a preset number of oscillation process variables of the solid oxide fuel cell system, n (n=24 in this embodiment) oscillation process variables are first obtained through step S1, and then the principal component analysis method is used to calculate the principal component coefficient of each oscillation process variable (recorded as α1, α2, ..., α n ) and variance (respectively denoted as λ1, λ2…λ n ); then the variances are arranged in descending order and the cumulative variance contribution rate is calculated. In the embodiment of the present application, the expression of the cumulative variance contribution rate is:

[0090]

[0091] Among them, λ i Represents the variance, m represents the number of oscillation process variables in the first residual oscillation process variables, n represents the number of oscillation process variables in the preset number of oscillation process variables, and i represents the sequence number of the oscillation process variables. Then, the oscillation process variables whose corresponding cumulative variance contribution rate is greater than the preset value (95%) are deleted from the 24 oscillation process variables to obtain the first residual oscillation process variables (m oscillation process variables); then, the oscillation significance index of each of the remaining oscillation process variables is calculated based on the m principal component coefficients and variances. At this time, the variance vector LAMBDA and the principal component coefficient matrix COEFF can be obtained, and the oscillation significance index OSI = COEFF 2 *LAMBDA; then, the oscillation process variables whose corresponding oscillation significance index is less than a preset value (10%) are deleted from the remaining m oscillation process variables to obtain a second remaining oscillation process variable, and the second remaining oscillation process variable is used as an oscillation analysis subset.

[0092] For details, see Figure 2 In this embodiment, after step S2, 5 process variables are finally screened out from the 24 oscillation process variables. These 5 process variables constitute the oscillation analysis subset, which includes the variables of combustion chamber inlet temperature, methane pressure, stack voltage, air bypass flow and reformer fuel inlet temperature.

[0093] S3: Obtaining optimal parameters of an oscillation root cause detection algorithm according to the oscillation analysis subset;

[0094] In an embodiment of the present application, the step of obtaining the optimal parameters of the oscillation root cause detection algorithm according to the oscillation analysis subset comprises the following steps:

[0095] Obtaining the oscillation analysis subset;

[0096] Obtaining an oscillation analysis subset reconstructed phase space corresponding to the oscillation analysis subset;

[0097] Calculating an optimal time delay for reconstructing phase space of the oscillation analysis subset;

[0098] Calculating the optimal embedding dimension of the oscillation analysis subset reconstructed phase space according to the optimal time delay;

[0099] The optimal time delay and the optimal embedding dimension are used as the optimal parameters.

[0100] In the embodiment of the present application, when the optimal parameters of the oscillation root cause detection algorithm are obtained according to the oscillation analysis subset, the mutual information I(x(t), x(t+τ)) between the original time series x(t) and the time series x(t+τ) offset by τ is calculated for the five process variables selected in S2, respectively. The calculation formula is as follows:

[0101]

[0102] In the formula, p i , represents the probability of the bin i of the histogram constructed from the data points in x(t), p ij (τ) represents the probability that x(t) is in bin i and x(t+τ) is in bin j.

[0103] Using this method, the time delay τ is estimated by averaging the mutual information over all process variables in the oscillation analysis subset, and the position where the average mutual information function is below a threshold 1 / e is selected as the optimal time delay.

[0104] For details, see Figure 3, in this embodiment, the position where the average mutual information is lower than the threshold is at τ = 1, so the optimal time delay is 1. Based on the calculated optimal time delay, the original time series x(t) is reconstructed into a high-dimensional time series y(t) of D dimensions: y(t) = (x(t), x(t+τ), ... x(t+(D-1)τ)), and the time series of all process variables in the oscillation analysis subset are gradually embedded into higher dimensions, and the embedding dimension is estimated by checking the change in the distance between adjacent points in the phase space. Specifically, the pseudo nearest neighbor (FNN) method is used to select the optimal embedding dimension. The basic idea of ​​this method is to check whether the distance between data points changes significantly by increasing the embedding dimension D. If the embedding significantly changes the distance between neighbors, they are called false neighbors, which indicates that the data needs to be further embedded until the number of FNNs drops to 0, or the dimension corresponding to the subsequent embedding does not change the number of FNNs is the optimal embedding dimension.

[0105] For details, see Figure 4 In this embodiment, when the number of embeddings is 2, the percentage of pseudo nearest neighbors drops to 0, so the optimal embedding dimension is 2.

[0106] S4: determining the causal relationship between each process variable in the oscillation analysis subset according to the optimal parameters;

[0107] In an embodiment of the present application, determining the causal relationship between the process variables in the oscillation analysis subset according to the optimal parameters comprises the following steps:

[0108] Obtaining an optimal time delay and an optimal embedding dimension among the optimal parameters;

[0109] obtaining the oscillation process variable;

[0110] Reconstructing a reconstructed phase space that is topologically equivalent to the oscillating process variable according to the optimal time delay and the optimal embedding dimension;

[0111] Calculating predicted values ​​of the oscillating process variables obtained by cross-mapping according to the reconstructed phase space;

[0112] Obtaining observation values ​​of the reconstructed phase space;

[0113] Calculating the correlation coefficient between the predicted value and the observed value;

[0114] quantifying the strength of the causal relationship between the oscillating process variables according to the correlation coefficient and obtaining a convergent cross-mapping causal relationship diagram;

[0115] The causal relationship between the oscillation process variables in the oscillation analysis subset is determined according to the converged cross-map causal relationship diagram.

[0116] In the embodiment of the present application, specifically, according to the principle of reconstructing the phase space, for time series X and Y with length L, assuming that the embedding dimension of the reconstructed phase space is E and the sampling interval is τ, then at time t, the phase space M is reconstructed x 、M y The coordinates are:

[0117]

[0118]

[0119] Among them, x(t) and y(t) represent the coordinates of the reconstructed phase space, X() and Y() represent time series, t represents the moment, τ represents the sampling interval, and E represents the embedding dimension of the reconstructed phase space.

[0120] Calculate from M X The coordinates are cross-mapped to obtain the predicted value of Y(t) The formula is as follows:

[0121]

[0122] w i =u i / ∑u j j=1…E+1,

[0123] u i =exp{-d[x(t), x(t i )] / d[x(t),x(t1)]},

[0124] Among them, among them, represents the predicted value, w i Represents the reconstructed phase space M X The distance weight between the coordinate x(t) and its i-th neighboring point, E represents the embedding dimension of the reconstructed phase space, d[x(t), x(t i )] represents the Euclidean distance between two vectors.

[0125] Calculate predicted values The correlation coefficient ρ between the observed value Y(t) is as follows:

[0126]

[0127] in, ρ represents the correlation coefficient, Y(i) represents the time series, represents the mean value of the time series, represents the predicted value, represents the average value of the predicted value, L represents the length of the time series, and i represents the sequence number of the adjacent point in the reconstructed phase space.

[0128] If variable Y is the driving factor of variable X, then as the time series length L increases, the correlation coefficient ρ gradually converges to [0, 1], and the larger the ρ value, the stronger the causal relationship between the variables.

[0129] See the effect diagram of the above results for Figure 5 .

[0130] S5: Determine the root cause of the oscillation of the solid oxide fuel cell system according to the cause-effect relationship.

[0131] In an embodiment of the present application, determining the root cause of the oscillation of the solid oxide fuel cell system according to the cause-effect relationship comprises the steps of:

[0132] Obtain a convergent cross-mapping causal relationship graph corresponding to the causal relationship;

[0133] acquiring process knowledge of the solid oxide fuel cell system;

[0134] The oscillation root cause is detected based on the converged cross-map causal relationship graph and the process knowledge.

[0135] In the embodiments of the present application, Figure 5 The convergence cross-mapping causal relationship diagram between the five oscillating variables screened is shown. In the figure, the two lines in the (i, j)th sub-graph represent the oscillation influence of the i-th oscillating variable on the j-th variable and the oscillation influence of the j-th variable on the i-th variable, respectively. It can be seen from the figure that the causal influence of methane pressure on the stack voltage and the stack voltage on methane pressure are the most significant, and there is no significant difference between the two, which indicates that the methane pressure and the stack voltage affect each other. Combined with the process knowledge of the SOFC system, the discharge voltage in the gas circuit changes, and the amount of methane consumed is different when the methane supply remains unchanged. The methane fluctuation leads to the pressure fluctuation in the gas circuit. In addition, it can be seen from the positions (1, 5) and (5, 1) that the oscillation influence between the reformer fuel inlet temperature and the combustion chamber inlet temperature is more significant. As the high-temperature heat source peak component of the entire SOFC system, the temperature oscillation of the tail gas combustion chamber will indirectly affect the temperature of the heat exchanger and the reformer along the gas circuit. The temperature fluctuation of the exhaust combustion chamber will be transmitted to the cathode gas through the heat exchanger, causing the temperature of the cathode air to oscillate. When the cathode air enters the stack, it will also cause the temperature inside the stack to change, eventually causing the electrical characteristics of the SOFC stack to fluctuate, thus affecting the operating characteristics of the entire system. In addition, the causal influence between the air bypass flow and other variables is very small, close to 0, which indicates that the air bypass flow does not participate in the above oscillation propagation.

[0136] Based on the above analysis and SOFC system structure (see Figure 6 ), it is inferred that the unstable methane pressure is the root cause of the oscillation failure, which directly affects the reformer fuel inlet temperature fluctuation, and then causes the stack voltage to fluctuate, and finally causes the fuel chamber temperature to fluctuate.

[0137] The present application provides a method for detecting the root cause of oscillation of a solid oxide fuel cell system. In order to evaluate the performance of the SOFC system and find the root cause of the system oscillation, each process variable in the system is derived from the complete power generation process in the experiment, so that the fault analysis result is true and reliable; in order to narrow the scope of causal analysis and improve the accuracy of identifying the fault source, the fault variables in the original data are screened and an oscillation analysis subset is obtained before performing the root cause detection of the fault; the variables in the oscillation analysis subset are subjected to causal analysis using a convergent cross-mapping algorithm, and the root cause of the system oscillation is detected from the causal analysis results, which provides an accurate and effective technical solution for solving the root cause detection of the oscillation of the solid oxide fuel cell system, can accurately and efficiently detect the root cause of the system oscillation, and solves the technical problem that the efficiency of locating the root cause of the oscillation of the solid oxide fuel cell system is low and affects the operation of the system.

[0138] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements. The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. The various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features applied for herein.

[0139] In short, the above is only a preferred embodiment of the technical solution of the present invention, and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting root causes of oscillation in a solid oxide fuel cell system, characterized in that: The method comprises the steps of: obtaining a preset number of solid oxide fuel cell system oscillation process variables; Filtering an oscillation analysis subset according to a preset number of oscillation process variables of the solid oxide fuel cell system; Obtaining optimal parameters of an oscillation root cause detection algorithm according to the oscillation analysis subset; determining a cause-and-effect relationship between each process variable in the oscillation analysis subset according to the optimal parameters; determining a root cause of oscillation of the solid oxide fuel cell system based on the cause-effect relationship; The step of obtaining the optimal parameters of the oscillation root cause detection algorithm according to the oscillation analysis subset comprises the following steps: Obtaining the oscillation analysis subset; Obtaining an oscillation analysis subset reconstructed phase space corresponding to the oscillation analysis subset; Calculating an optimal time delay for reconstructing phase space of the oscillation analysis subset; Calculating the optimal embedding dimension of the oscillation analysis subset reconstructed phase space according to the optimal time delay; Taking the optimal time delay and the optimal embedding dimension as the optimal parameters; Determining the causal relationship between the process variables in the oscillation analysis subset according to the optimal parameters comprises the following steps: Obtaining an optimal time delay and an optimal embedding dimension among the optimal parameters; obtaining the oscillation process variable; Reconstructing a reconstructed phase space that is topologically equivalent to the oscillating process variable according to the optimal time delay and the optimal embedding dimension; Calculating predicted values ​​of the oscillating process variables obtained by cross-mapping according to the reconstructed phase space; Obtaining observation values ​​of the reconstructed phase space; Calculating the correlation coefficient between the predicted value and the observed value; quantifying the strength of the causal relationship between the oscillating process variables according to the correlation coefficient and obtaining a convergent cross-mapping causal relationship diagram; Determining the causal relationship between each of the oscillation process variables in the oscillation analysis subset according to the converged cross-mapping causal relationship diagram; The expression of the reconstructed phase space is: x(t)=<X(t),X(t-T),X(t-2T),...,X(t-(E-1)T)> , Wherein, x(t) represents the coordinate of the reconstructed phase space, X() represents the time series, t represents the moment, τ represents the sampling interval, and E represents the embedding dimension of the reconstructed phase space; The expression of the predicted value is: in, Represents the predicted value, wi represents the distance weight between the coordinate x(t) and its i-th neighboring point on the reconstructed phase space MX, and E represents the embedding dimension of the reconstructed phase space.

2. The method for detecting root cause of oscillation of a solid oxide fuel cell system according to claim 1, characterized in that: The method of obtaining a preset number of solid oxide fuel cell system oscillation process variables comprises the steps of: Disposing a measured variable detection sensor in the solid oxide fuel cell system; Disposing an operating variable detection sensor in the solid oxide fuel cell system; Acquiring a first number of measured variables when the solid oxide fuel cell system oscillates by means of the measured variable detection sensor; Acquiring, by means of the operating variable detection sensor, a second number of operating variables when the solid oxide fuel cell system oscillates; A first number of the measured variables and a second number of the manipulated variables are combined as the oscillating process variable.

3. The method for detecting root cause of oscillation of a solid oxide fuel cell system according to claim 1, characterized in that: The step of screening the oscillation analysis subset according to a preset number of oscillation process variables of the solid oxide fuel cell system comprises the following steps: Acquire a preset number of the oscillation process variables; Calculate the principal component coefficient and variance of each of the oscillatory process variables using principal component analysis; Arrange the variances in descending order and calculate the cumulative variance contribution rate; Deleting the oscillation process variable corresponding to the cumulative variance contribution rate greater than a preset value from all the oscillation process variables and obtaining a first residual oscillation process variable; Calculating the oscillation significance index of each of the remaining oscillation process variables according to the principal component coefficient and variance corresponding to each of the oscillation process variables in the first remaining oscillation process variables; Deleting the oscillation process variable corresponding to the oscillation significance index less than a preset value from the first residual oscillation process variable and obtaining a second residual oscillation process variable; All of the oscillating process variables in the second residual oscillating process variables are grouped into the oscillation analysis subset.

4. The method for detecting root cause of oscillation of a solid oxide fuel cell system according to claim 3, characterized in that: The expression of the cumulative variance contribution rate is: Among them, λ i represents the variance, m represents the number of oscillation process variables in the first residual oscillation process variables, n represents the number of oscillation process variables in the preset number of oscillation process variables, and i represents the sequence number of the oscillation process variables.

5. The method for detecting root cause of oscillation of a solid oxide fuel cell system according to claim 1, characterized in that: Determining the root cause of oscillation of the solid oxide fuel cell system according to the cause-effect relationship comprises the steps of: Obtain a convergent cross-mapping causal relationship graph corresponding to the causal relationship; acquiring process knowledge of the solid oxide fuel cell system; The oscillation root cause is detected based on the converged cross-map causal relationship graph and the process knowledge.

Images