Gas leakage diagnosis method for solid oxide fuel cell stack

Abstract

The invention belongs to the technical field of SOFC (solid oxide fuel cell) stack air leakage diagnosis, and discloses a solid oxide fuel cell stack air leakage diagnosis method, which comprises the following steps: constructing an SOFC stack air leakage fault state quantity model, carrying out real-time estimation on the stack air leakage fault state quantity model, comparing the fault state quantity estimated in real time with a warning threshold value, and carrying out stack air leakage diagnosis. According to the method, the gas leakage fault state quantity model is established based on the gas change at the inlet and the outlet of the electric pile instead of constructing a complex SOFC system model, so that the calculation complexity is low. And the fault state quantity is estimated by using particle filtering, and the gas leakage of the electric pile is diagnosed online, so that the accuracy, the universality and the stability of the real-time diagnosis of the gas leakage of the electric pile are improved. Meanwhile, the influence of noise on the state quantity is considered, an online alarm threshold value is designed, the gas leakage fault of the SOFC electric pile is jointly monitored by combining the fault state quantity and the alarm threshold value, and the gas leakage fault of the electric pile can be more sensitive under the condition of keeping noise interference resistance.

Description

Technical Field

[0001] The present invention belongs to the technical field of leakage diagnosis of SOFC stacks, and more specifically, relates to a method for diagnosing leakage of a solid oxide fuel cell stack. Background Art

[0002] A solid oxide fuel cell (SOFC for short) is an energy device that uses solid oxides as electrolytes. It directly realizes an electrochemical reaction of fuel at high temperature, and efficiently converts chemical energy into electrical energy. However, the stack needs to be sealed to ensure that fuel and air do not mix during operation in a high-temperature environment. However, after long-term operation of the SOFC, multiple thermal cycles, mechanical stresses, and defects and aging of the sealing material itself will cause the stack to leak air, eventually causing the stack to burn out, paralyzing the entire SOFC system and posing serious safety hazards. Therefore, online state detection and leakage fault diagnosis of the SOFC stack are essential for the SOFC system.

[0003] Leakage fault diagnosis of the SOFC stack includes data-based fault diagnosis and model-based fault diagnosis.

[0004] Data-based fault diagnosis: It depends on a large amount of historical and real-time SOFC operation data. Through data mining and feature learning techniques, such as neural networks, support vector machines, etc., implicit knowledge of the dependence relationship between internal variables of the system is extracted, and potential patterns and rules between system variables are captured. On this basis, the extracted implicit knowledge is compared with the characteristics shown in the actual operation process of the SOFC system to check the consistency or deviation between the two, so as to realize the monitoring and fault diagnosis of the leakage of the SOFC stack. This method depends on a large amount of historical data and prior knowledge, so it is difficult to achieve online real-time prediction. In addition, its generalization ability is limited and it is often only applicable to a specific system or stack; once applied to different systems, its prediction effect and adaptability will decrease significantly, thus affecting the universality and stability in practical applications.

[0005] Model-based Fault Diagnosis: A process model of a SOFC is constructed through physical laws or system identification, which specifically describes the relationships between various variables of the SOFC. On this basis, by comparing the consistency between the variables in the built SOFC system and the actual test data, the monitoring and fault diagnosis of the SOFC air leakage fault are realized. For example, in the prior art, a method for diagnosing the air leakage of a SOFC stack is disclosed, which uses the difference between the predicted values of the gas concentrations at the anode and cathode outlets of the stack and the measured values of the gas concentrations at the anode and cathode outlets of the stack for air leakage diagnosis. The main problem with this method is that, in addition to air leakage, other situations (such as temperature changes, etc.) may also cause changes in the difference between the predicted values and the measured values of the gas concentrations at the anode and cathode outlets of the stack. Therefore, this method will reduce the accuracy of the air leakage diagnosis of the stack. Moreover, the SOFC itself is a multi-input multi-output, non-linear system, and its internal coupling relationship is very complex, which makes the model construction and solution full of challenges. In addition, in practical applications, the system is often subject to large noise interference, which will have a significant impact on the model prediction results, thereby reducing the sensitivity of the fault diagnosis. Summary of the Invention

[0006] In view of the above defects or improvement requirements of the prior art, the present invention provides a method for diagnosing the air leakage of a solid oxide fuel cell stack, aiming to improve the accuracy, universality, and stability of the real-time diagnosis of the air leakage of the stack.

[0007] To achieve the above object, the present invention provides a method for diagnosing the air leakage of a solid oxide fuel cell stack, including:

[0008] Real-time estimate the fault state variables [θ fuel , θ air of the solid oxide fuel cell stack, and compare the real-time estimated fault state variables [θ fuel , θ air with the warning threshold to perform the air leakage diagnosis of the stack; wherein, the models of θ fuel and θ air are respectively:

[0009]

[0010] In the formula, θ fuel and θ air respectively represent the fault state variables of the fuel channel air leakage and the air channel air leakage of the stack, and the fuel gas in the fuel channel of the stack is hydrogen; λ fuel and λ air are respectively the gas leakage amounts of the fuel channel and the air channel, and are respectively the gas molar flow rates at the inlets of the fuel channel and the air channel of the stack; is the hydrogen concentration at the outlet of the fuel channel in the fuel cell stack, is the oxygen concentration at the outlet of the air channel in the fuel cell stack; is the hydrogen concentration at the leakage point of the fuel channel, is the oxygen concentration at the leakage point of the air channel; F is the Faraday constant, and I stk is the current of the fuel cell stack.

[0011] Furthermore, real-time estimation is performed on the leakage fault state variables [θ fuel , θ air of the solid oxide fuel cell stack, including:

[0012] Construct a state observation model:

[0013]

[0014] where x, u, and y represent the state variables, inputs, and outputs of the state observation model, respectively; and represent the hydrogen concentration in the fuel channel and the oxygen concentration in the air channel at the inlet of the fuel cell stack, respectively;

[0015] Discretize the state observation model, and use particle filtering to perform real-time estimation on the state variables of the discretized state observation model to obtain the real-time estimated values of the leakage fault state variables [θ fuel , θ air of the solid oxide fuel cell stack.

[0016] Furthermore, the warning threshold is the online alarm threshold. The online alarm threshold θ fuel-t for fuel channel leakage and the online alarm threshold θ air-t for air channel leakage are respectively:

[0017]

[0018] where respectively represent the variances of; respectively represent the variances of.

[0019] Furthermore, compare the real-time estimated fault state variables [θ fuel , θ air with the warning threshold to perform leakage diagnosis on the fuel cell stack, including:

[0020] |θ fuel | ≤ |θ fuel-t |, and when |θ air | ≤ |θ air-t |, no leakage fault occurs;

[0021] |θ fuel |>|θ fuel-t |, and |θ air |≤|θ air-t |, fuel leakage occurs in the fuel channel;

[0022] |θ fuel |≤|θ fuel-t |, and |θ air |>|θ air-t |, air leakage occurs in the air channel;

[0023] |θ fuel |>|θ fuel-t |, and |θ air |>|θ air-t |, fuel leakage and air leakage occur simultaneously in the fuel channel and the air channel.

[0024] Furthermore, the model construction method of the θ fuel and θ air includes:

[0025] Construct a pneumatic balance equation representing whether the stack leaks:

[0026]

[0027] where, represents the molar flow rate of the gas at the pipeline j of the stack i, represents the gas concentration at the pipeline j of the stack i, air represents the air channel, fuel represents the fuel channel, in represents the stack inlet, out represents the stack outlet, A represents the gas reaction coefficient, r represents the gas reaction rate, an, ca represent the anode and the cathode respectively;

[0028] Based on the flow rate changes at the stack inlet and outlet, rewrite the pneumatic balance equation into the following transformed pneumatic balance equation:

[0029]

[0030] where, n fuel and n air represent the total amount of substances input into the fuel channel and the air channel respectively, and represent respectively and the change rates of, and represent the hydrogen concentration in the fuel channel and the oxygen concentration in the air channel at the stack inlet respectively;

[0031] Extract the fault state variables related to stack leakage in the transformed pneumatic balance equation to obtain the θ fueland θ air model.

[0032] The present invention also provides a leakage diagnosis system for a solid oxide fuel cell stack, including a computer-readable storage medium and a processor;

[0033] The computer-readable storage medium is used for storing executable instructions;

[0034] The processor is used for reading the executable instructions stored in the computer-readable storage medium and executing the leakage diagnosis method for the solid oxide fuel cell stack described in any one of the above.

[0035] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the leakage diagnosis method for the solid oxide fuel cell stack described in any one of the above is implemented.

[0036] The present invention also provides a computer program product, including a computer program, and when the computer program runs on a computer, the computer is enabled to execute the leakage diagnosis method for the solid oxide fuel cell stack described in any one of the above.

[0037] Generally speaking, through the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0038] (1) The leakage diagnosis method for the solid oxide fuel cell stack provided by the present invention directly establishes a fault state quantity model related to the leakage of the fuel channel and the air channel of the SOFC, and estimates the fault state quantity in real time to determine whether the stack leaks. The present invention establishes a leakage fault diagnosis model (fault state quantity model) based on the gas changes at the inlet and outlet of the stack, rather than constructing a complex SOFC system model, and the calculation complexity is greatly reduced. The present invention does not rely on a large amount of historical data and prior knowledge, and uses an estimation method to estimate the fault state quantity in real time, which is convenient for real-time online fault diagnosis of the SOFC system. Compared with diagnosing leakage by using the difference between the predicted value and the measured value of the gas concentration at the anode and cathode outlets of the stack, the present invention directly estimates the state quantity related to the leakage fault in real time, and the accuracy of the stack leakage diagnosis is relatively high. The diagnosis method of the present invention does not depend on a specific solid oxide fuel cell, and has high universality and stability.

[0039] (2) Preferably, the present invention uses a particle filter fault observer to estimate the fault state quantity of the leakage model, and can complete high-precision fault diagnosis of the SOFC system in real time online.

[0040] (3) Further, a dynamic alarm threshold is designed, and the leakage fault is diagnosed by combining the estimated fault state quantity and the dynamic threshold, ensuring that the leakage fault is not misreported due to system noise, and improving the sensitivity of the fault diagnosis. Description of the Drawings

[0041] Figure 1 This is the flowchart of the method for diagnosing air leakage in a solid oxide fuel cell stack according to an embodiment of the present invention;

[0042] Figure 2 This is the detection diagram of the fault state when air leakage faults occur in the fuel channel and the air channel according to an embodiment of the present invention; Figure 2 In it, a and b respectively represent the detection of the corresponding fault states when air leakage faults occur in the fuel channel and the air channel;

[0043] Figure 3 This is the detection of the fault state under different noise variances when air leakage faults occur in the fuel channel and the air channel according to an embodiment of the present invention; Figure 3 In it, a and b respectively represent the detection of the corresponding fault states when air leakage faults occur in the fuel channel and the air channel. Detailed Embodiments

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0045] Embodiment 1

[0046] As Figure 1 shown, an embodiment of the present invention provides a method for diagnosing air leakage in a solid oxide fuel cell stack, including:

[0047] Real-time estimation is performed on the air leakage fault state variables [θ fuel , θ air of the solid oxide fuel cell stack, and the real-time estimated fault state variables [θ fuel , θ air are compared with the warning threshold to perform air leakage diagnosis of the stack; wherein, θ fuel and θ air respectively represent the fault state variables of air leakage in the SOFC fuel channel and the air channel, and the fuel gas in the fuel channel of the stack is hydrogen; θ fuel and θ air are respectively:

[0048]

[0049] In the formula, λ fuel and λ air are respectively the gas leakage amounts of the fuel channel and the air channel, and are the molar flow rates of the fuel channel and the air channel at the inlet of the stack, respectively; is the hydrogen concentration at the outlet of the fuel channel of the stack, is the oxygen concentration at the outlet of the air channel of the stack; is the hydrogen concentration at the leakage point of the fuel channel, is the oxygen concentration at the leakage point of the air channel. When there is no leakage, is equal to ; is equal to . F is the Faraday constant, and I stk is the stack current.

[0050] In the embodiments of the present invention, the construction method of the leakage fault state quantity model of the solid oxide fuel cell stack includes:

[0051] The SOFC stack uses an ultimate current type high temperature gas sensor to directly detect the gases at the inlet and outlet of the stack. In the embodiments of the present invention, the solid oxide gas sensor in the fuel channel of the stack is a hydrogen sensor; the solid oxide gas sensor in the air channel of the stack is an oxygen sensor. The present invention constructs a stack leakage model based on the pneumatic balance of the gases at the inlet and outlet of the stack.

[0052] In the case of no leakage, the pneumatic balance of the stack can be expressed as:

[0053]

[0054] where represents the molar flow rate of the gas at the j-th position of the i-th pipeline of the SOFC, represents the molar fraction (gas concentration) of the gas at the j-th position of the i-th pipeline of the SOFC. air represents the air channel, fuel represents the fuel channel, in represents the inlet of the stack, out represents the outlet of the stack, A represents the gas reaction coefficient, r represents the gas reaction rate, and an and ca represent the anode and cathode, respectively.

[0055] Analyze the leakage fault. If leakage occurs, the estimated value of the gas concentration at the outlet of the stack will jump. The leakage in the internal flow channels of the stack is respectively fuel channel leakage and air channel leakage. Assume that the gas leakage amounts of the fuel channel and the air channel are λ fuel and λ air , respectively. When there is no gas leakage, it can be considered that [λ air , λ fuel = 0. Based on this, rewrite the above pneumatic balance expression of the stack to obtain the rewritten pneumatic balance expression of the stack:

[0056]

[0057] Considering that the gas flow rate and concentration at the inlet of the stack are determined by the operating parameters of the stack (fuel utilization rate FU, air excess ratio AR, bypass valve opening BP) and the stack current I stk and can be regarded as constant. In the fuel channel (the fuel is H 2 ), for every mole of H 2 losing electrons and O in the air channel 2- combining to generate one mole of H 2 O, so the molar flow rates at the inlet and outlet of the fuel channel are the same. In the air channel, O 2 obtains the electrons lost by H 2 in the fuel channel to generate O 2- which passes through the solid oxide electrolyte (YSZ electrolyte in the embodiment of the present invention). The molar flow rate at the outlet of the air channel is less than that at the inlet. Therefore, based on the flow rate changes at the inlet and outlet of the stack, the above rewritten stack pneumatic balance expression can be transformed into:

[0058]

[0059] where n fuel represents the total amount of substance input into the fuel channel, and n air represents the total amount of substance input into the air channel, represents the change rate of represents the change rate of represents the hydrogen concentration at the inlet of the fuel channel in the stack, represents the oxygen concentration at the inlet of the air channel in the stack.

[0060] Thus, the fault state variables [θ fuel , θ air related to the stack air leakage fault can be obtained:

[0061]

[0062] Thus, it is possible to determine whether the SOFC stack has air leakage by estimating the fault state variables. θ fuel and θ air respectively reflect the fuel channel gas leakage fault and the air channel gas leakage fault. In the ideal state, when there is no fault, θ fuel and θ air are both 0, otherwise it indicates an air leakage fault. In practical applications, affected by interference such as noise, a warning threshold can be set, and the real-time estimated fault state variables [θ fuel , θ air are compared with the warning threshold for stack air leakage diagnosis.

[0063] As a specific implementation method, the fault state variables [θ fuel , θ air can be estimated in real time by using unscented Kalman filter, extended Kalman filter, particle filter or the like. The warning threshold can be set according to experience or can be designed specifically. In the embodiments of the present invention, it is designed specifically based on the monitored noise level.

[0064] Preferably, in the embodiments of the present invention, particle filter is used to estimate the fault state variables [θ fuel , θ air in real time, including:

[0065] Taking the hydrogen concentration at the outlet of the stack oxygen concentration the fault state variable θ of fuel channel leakage fuel and the fault state variable θ of air channel leakage air as the system state; taking the hydrogen concentration at the inlet of the stack oxygen concentration molar flow rate of fuel molar flow rate of air and the stack current I stk as the input, and finally taking the hydrogen concentration and oxygen concentration at the outlet of the stack as the observation value. Thus, the state variables, input, and output can be obtained as follows:

[0066]

[0067] The essence of particle filter is to perform Bayesian recursion on discrete time series, and it is necessary to discretize the recursion process of the state. In the embodiments of the present invention, Euler approximation is used to convert the continuous-time state equation into a discrete-time state equation, where k is the discrete time step and T s is the sampling period. After discretizing the above continuous leakage fault model, the obtained discrete leakage fault model is:

[0068] x k+1 = f(x k , u k , w k )

[0069] y k = h(x k , u k , v k )

[0070] Among them, x k represents the state variable at time step k, u k represents the input at time step k, and yk denotes the output at time step k, w k and v k are the process noise and measurement noise at time step k, respectively.

[0071] Based on the above discrete air leakage fault model, the process of particle filtering is as follows:

[0072] (1) Initialization: By independently sampling the prior distribution of the initialized state variables, an initial set of N particles is obtained and the same initial weight is assigned to the N initial particles N is determined according to the complexity of the problem, the state space dimension, and the computing resources.

[0073] (2) Prediction step: According to the state transition equation of the system each particle at the previous time step k - 1 is sampled to obtain a set of N particles predicted at the current time step k The set of N particles predicted at the current time step k is used to represent the prior distribution p(x k |y 1:k-1 ) of the state variable at the current time step k:

[0074]

[0075] where, denotes the Dirac function, and y 1:k-1 represents the output of the previous k - 1 time steps.

[0076] (3) Update step: Based on the output y k observed at the current time step k, and the prior distribution p(x k |y 1:k-1 ) of the state variable at the current time step k, the weights of the particles are updated to make the particle distribution approximate the target posterior distribution p(x k |y 1:k ):

[0077] p(x k |y 1:k ) ∝ p(y k |x k ) p(x k |y 1:k-1 )

[0078] where, p(y k |x k ) represents the probability distribution of y k when x k is known, and ∝ represents a proportional relationship.

[0079] Thus, the weight update formula is obtained:

[0080]

[0081] wherein, the importance density function of sampling in the embodiments of the present invention is equal to the state transition probability.

[0082] Normalize the weights:

[0083]

[0084] (4) Resampling:

[0085] As the filtering process progresses, due to the cumulative effect of the observed data, the weights will concentrate on a very small number of particles, resulting in filtering estimation using only a very small number of particles. Resampling is to extract a new set of particles from the current particles, such that the extraction probability of each particle is proportional to its weight. In the embodiments of the present invention, systematic resampling is adopted, and the effective sample size (ESS) is used to measure the degree of particle degradation:

[0086]

[0087] When the ESS is lower than a certain threshold, resampling is initiated, that is, for the new particles The cumulative weights are evenly divided, and then sampling points are determined within each division interval to obtain updated particles and the corresponding weights

[0088] The updated particles and the corresponding weights are weighted to obtain the state variables at the current time step k, realizing the estimation of θ at the current time step k fuel and θ air and let k = k + 1, enter the estimation of θ at the next time step fuel and θ air to realize the real-time estimation of θ fuel and θ air .

[0089] As Figure 2 shown, Figure 2 in a, it shows that the fuel channel has an air leakage fault at 15000 s. Before the occurrence of Fault Ⅰ, the fault state value θ of the fuel channel estimated by the particle filter state observer fuel fluctuates around 0, but is always less than the dynamic threshold θ of the fuel channel fuel-t , and the alarm is not triggered. When a step signal is injected into the SOFC stack to simulate the air leakage of the fuel channel, θ fuel rapidly rises and breaks through the dynamic threshold θ fuel-t, an alarm is triggered, and the system prompts that a fuel channel air leakage fault has occurred. In Figure 2 A similar phenomenon can be seen in b of Figure 2 It shows that when the air leakage fault occurs, the system can quickly and accurately identify the air leakage accidents in the fuel channel and the air channel, proving that the stack air leakage fault state quantity model and the fault observer based on particle filter constructed by the present invention are effective.

[0090] As a further design of the present invention, the above warning threshold is an online alarm threshold designed specifically based on the monitored noise level.

[0091] In the embodiment of the present invention, an online alarm dynamic threshold generator is designed. Its core principle is that when the system is in stable operation and no air leakage fault occurs, the noise level is monitored in real time, and the alarm threshold is dynamically estimated accordingly, so that the generated threshold always remains higher than the maximum absolute value of the current noise, thus effectively preventing false alarms caused by noise interference. When the system is in a stable working state and there is no air leakage, the flow rate at the leakage point can be expressed as:

[0092]

[0093] Under the stable state of the stack, it can be considered that the input flow rate is constant. Assuming that the measurement noise is independent zero-mean Gaussian white noise, the variance of the error can be obtained as:

[0094]

[0095] Among them, and are the variances of θ fuel and θ air respectively; respectively represent the variances of; respectively represent the variances of.

[0096] It can be inferred that under normal working conditions, θ fuel and θ air follow a normal distribution, so 99.7% of the data falls within 3 standard deviations, and the warning threshold is obtained as follows:

[0097]

[0098] Comparing the fault estimation value with the warning threshold can determine the actual fault type as follows:

[0099] a) No air leakage fault, |θ fuel | ≤ |θ fuel-t | and |θ air | ≤ |θ air-t |

[0100] b) Leakage in the fuel channel, |θ fuel | > |θ fuel-t | and |θ air | ≤ |θ air-t |

[0101] c) Leakage in the air channel, |θ fuel | ≤ |θ fuel-t | and |θ air | > |θ air-t |

[0102] d) Leakage in the dual channels, |θ fuel | > |θ fuel-t | and |θ air | > |θ air-t |

[0103] As Figure 3 shown, Figure 3 in a of [] shows the fuel channel fault status quantity and the dynamic threshold when the noise variance is 1E - 8. In addition, the dynamic threshold with a noise variance of 1E - 6 is also shown. As can be seen from Figure 3 a in [], before the leakage fault in the fuel channel occurs, the fuel channel fault status quantity θ fuel is simultaneously less than the dynamic threshold with a noise variance of 1E - 6 and the dynamic threshold with a noise variance of 1E - 8. And after the fault occurs, θ fuel can exceed both dynamic thresholds to trigger an alarm. However, the fluctuation range of the fault status value will increase with the increase of noise. The dynamic threshold with a noise variance of 1E - 6 is significantly farther from the fuel channel fault status quantity θ fuel than the dynamic threshold with a noise variance of 1E - 8. This will lead to an increase in the airspace between the dynamic threshold and the fault status quantity, reducing the sensitivity of the alarm system to the leakage fault. In extreme cases, when the dynamic threshold is too large, although the fault status quantity θ fuel rises rapidly, it may not trigger an alarm because it does not exceed the dynamic threshold. On the contrary, when the dynamic threshold is too small, the airspace between the dynamic threshold and the fault status quantity is too small, and the alarm system will be over - sensitive. The noise generated by the system itself can easily exceed the threshold, resulting in false alarms of the system. The same phenomenon can also be verified in Figure 3 b of [].

[0104] In the embodiment of the present invention, a fault model is constructed through pneumatic balance instead of constructing a complex SOFC system model. This fault model gives the fault state variables when gas leakage occurs, and particle filtering is used to estimate the fault state variables for online diagnosis of stack air leakage. At the same time, considering the influence of noise on the state variables, an online alarm threshold is designed, and the fault state variables and the alarm threshold are combined to jointly monitor the gas leakage fault of the SOFC stack. It can be more sensitive to the stack air leakage fault while maintaining anti-noise interference.

[0105] Embodiment 2

[0106] The embodiment of the present invention provides a system for diagnosing air leakage in a solid oxide fuel cell stack, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method for diagnosing air leakage in the solid oxide fuel cell stack in Embodiment 1 above are implemented.

[0107] The related technical solutions are the same as above and will not be elaborated here.

[0108] Embodiment 3

[0109] The embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for diagnosing air leakage in the solid oxide fuel cell stack in Embodiment 1 above are implemented.

[0110] Specifically, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash device, or other volatile solid-state storage devices.

[0111] The related technical solutions are the same as above and will not be elaborated here.

[0112] Embodiment 4

[0113] The embodiment of the present application provides a computer program product, including a computer program. When the computer program runs on a computer, the computer is caused to execute the steps of the method for diagnosing air leakage in the solid oxide fuel cell stack in Embodiment 1 above.

[0114] The related technical solutions are the same as above and will not be elaborated here.

[0115] Those skilled in the art can easily understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A solid oxide fuel cell stack leakage diagnosis method, characterized in that: include: The solid oxide fuel cell stack leakage fault state quantity [θ fuel ,θ air ] to estimate the real-time fault state quantity [θ fuel ,θ air ] is compared with the warning threshold to diagnose stack leakage; where θ fuel and θ air The models are: In the formula, θ fuel and θ air They represent the fault state quantity of gas leakage in the fuel channel of the stack and the fault state quantity of gas leakage in the air channel, respectively, and the fuel gas in the fuel channel of the stack is hydrogen; λ fuel and λ air are the gas leakage of the fuel channel and the air channel respectively, and are the gas molar flow rates of the fuel channel and the air channel at the inlet of the fuel stack, respectively; is the hydrogen concentration of the fuel channel at the stack outlet, is the oxygen concentration of the air channel at the outlet of the fuel cell stack; is the hydrogen concentration at the fuel channel leakage point, is the oxygen concentration at the air passage leakage point; F is the Faraday constant, I stk is the stack current.

2. The solid oxide fuel cell stack leakage diagnosis method according to claim 1, characterized in that: The solid oxide fuel cell stack leakage fault state quantity [θ fuel ,θ air ] for real-time estimates, including: Constructing a state observation model: Among them, x, u, and y represent the state variables, input, and output of the state observation model respectively; and They represent the hydrogen concentration in the fuel channel and the oxygen concentration in the air channel at the stack inlet respectively; The state observation model is discretized, and the state variables of the discretized state observation model are estimated in real time using particle filtering to obtain the solid oxide fuel cell stack leakage fault state quantity [θ fuel ,θ air ] real-time estimate.

3. The solid oxide fuel cell stack leakage diagnosis method according to claim 1 or 2, characterized in that: The warning threshold is the online alarm threshold, and the online alarm threshold θ of the fuel channel leakage fuel-t and the online alarm threshold θ of air passage leakage air-t They are: in, Respectively The variance of Respectively The variance of .

4. The solid oxide fuel cell stack leakage diagnosis method according to claim 3, characterized in that: The real-time estimated fault state quantity [θ fuel ,θ air ]Compare with the warning threshold to perform stack leakage diagnosis, including: |θ fuel |≤|θ fuel-t |, and |θ air |≤|θ air-t |, no air leakage occurs; |θ fuel |>|θ fuel-t |, and |θ air |≤|θ air-t |, there is leakage in the fuel channel; |θ fuel |≤|θ fuel-t |, and |θ air |>|θ air-t |, the air passage leaks; |θ fuel |>|θ fuel-t |, and |θ air |>|θ air-t |, leakage occurs simultaneously in the fuel passage and the air passage.

5. The solid oxide fuel cell stack leakage diagnosis method according to claim 1, characterized in that: The θ fuel and θ air The model building methods include: Construct the aerodynamic balance equation to characterize whether the stack is leaking: in, i∈{air,fuel},j∈{in,out} represents the gas molar flow rate at pipe j of stack i, represents the gas concentration at the pipe j of the stack i, air represents the air channel, fuel represents the fuel channel, in represents the stack inlet, out represents the stack outlet, A represents the gas reaction coefficient, r represents the gas reaction rate, an and ca represent the anode and cathode respectively; Based on the flow rate changes at the inlet and outlet of the fuel cell stack, the aerodynamic balance equation is rewritten as the following deformed aerodynamic balance equation: Among them, n fuel and n air Respectively represent the total amount of material input into the fuel channel and the air channel, and Respectively and The rate of change, and They represent the hydrogen concentration in the fuel channel and the oxygen concentration in the air channel at the stack inlet respectively; Extract the fault state quantity related to stack leakage in the deformed aerodynamic balance equation to obtain the θ fuel and θ air Model.

6. A solid oxide fuel cell stack leakage diagnosis system, characterized in that: comprising a computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the solid oxide fuel cell stack leakage diagnosis method according to any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the solid oxide fuel cell stack leakage diagnosis method as described in any one of claims 1 to 5 is implemented.

8. A computer program product, characterized in that It comprises a computer program, which, when running on a computer, enables the computer to execute the solid oxide fuel cell stack leakage diagnosis method as described in any one of claims 1 to 5.

Images