Control method for efficient range extension of hydrogen fuel cell

Abstract

The invention relates to the technical field of fuel cells, and discloses a control method for efficient range extension of a hydrogen fuel cell, which comprises the following steps of: establishing a multi-dimensional probability density function of random interference data in a hydrogen fuel cell system model in a manner of establishing the hydrogen fuel cell system model and collecting operation data of the hydrogen fuel cell system model; a multi-dimensional probability density function is constructed, a chaotic polynomial orthogonal basis function is constructed by using the constructed multi-dimensional probability density function, the randomness in the hydrogen fuel cell system is processed by using chaotic expansion of the chaotic basis function, random uncertain parameters of the hydrogen fuel cell system are converted to obtain a deterministic model, and the hydrogen fuel cell system is controlled through the deterministic model. And the control of the hydrogen fuel cell system is more efficient.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cells, and in particular to a control method for efficiently extending the range of a hydrogen fuel cell. Background Art

[0002] With the increasing scarcity of fossil fuels and the continuous increase in human demand for energy, global energy is facing huge challenges. In recent years, my country has vigorously advocated the development of clean energy, striving to ensure the energy needs of the people on the basis of solving environmental pollution problems. Hydrogen fuel cells, as an efficient and clean form of energy, have attracted the research of many scholars and are considered to be the most promising clean alternative energy. They have been widely used in various fields. In the automotive field, fuel cells can be widely used in plug-in hybrid vehicles, pure electric vehicles and fuel cell vehicles. Among them, hydrogen fuel cells are an important direction for new energy vehicles.

[0003] However, the hydrogen fuel cell itself is a constrained system with uncertainty, nonlinearity, large hysteresis, time-varying, strong coupling, multiple inputs and multiple outputs. At the same time, hydrogen fuel cells can be used in various fields, which makes their working environment very complex and uncertain. This brings huge challenges to the research of control technology. Summary of the invention

[0004] The present invention aims to provide a control method for efficiently extending the range of a hydrogen fuel cell, so as to transform the random uncertainty optimization control problem of the hydrogen fuel cell into a deterministic optimization control problem.

[0005] To achieve the above object, the present invention adopts the following technical solution: a control method for high-efficiency range extension of a hydrogen fuel cell, comprising the following steps:

[0006] S1. Establish a hydrogen fuel cell system model;

[0007] S2, collecting the operating data of the hydrogen fuel cell system model established in S1, wherein the operating data includes external random interference data and internal random interference data;

[0008] S3, processing data, sorting and classifying the data collected in S2 according to the source of the collected data, obtaining the data of each parameter, and processing the data of each parameter separately to establish a multidimensional probability density function;

[0009] S4, using the multidimensional probability density function obtained in S3, construct the chaotic polynomial orthogonal basis function specific to hydrogen fuel cells, and obtain the basis function set {Φ i (r)};

[0010] S5, using the chaotic polynomial orthogonal basis function obtained in S4 to propagate uncertainty in the hydrogen fuel cell system, and using a non-invasive calculation method and the chaotic polynomial orthogonal basis function to perform chaotic expansion on the hydrogen fuel cell system to handle randomness and obtain a deterministic model;

[0011] S6, constructing a multi-objective operation optimization control task for a hydrogen fuel cell system, and transforming the random control problem of the hydrogen fuel cell system into a deterministic control problem through the deterministic model obtained in S5;

[0012] S7. Formulate multi-objective operation optimization control tasks for the hydrogen fuel cell system according to demand to achieve efficient range-extending control of the hydrogen fuel cell system.

[0013] The beneficial effects of this scheme are as follows: this scheme constructs a multidimensional probability density function of random interference data in the hydrogen fuel cell system model by establishing a hydrogen fuel cell system model and collecting operating data of the hydrogen fuel cell system model, and uses the constructed multidimensional probability density function to construct a chaotic polynomial orthogonal basis function, and uses the chaotic expansion of the chaotic basis function to process the randomness in the hydrogen fuel cell system, convert the random uncertain parameters of the hydrogen fuel cell system to obtain a deterministic model, and control the hydrogen fuel cell system through the deterministic model, making the control of the hydrogen fuel cell system more efficient.

[0014] Further, the hydrogen fuel cell includes a single stack battery. In S1, the hydrogen fuel cell system model includes a battery monomer subsystem model that simulates the working conditions of the single stack battery. The battery monomer subsystem model is x p (k+1)=f(x p ,u p ,r p ). In the battery cell subsystem model, x p represents the measurable state variable of subsystem p, u p represents the input acting on subsystem p, r p Represents the unknown model parameters of the subsystem p. Due to the uncertainty of the parameters, the established model is an incompletely determined model, that is, the values ​​of the parameters are different under different working conditions and environments, and the model is also different. This makes the established battery cell subsystem model match the actual single stack battery and reduce the error of the simulation value.

[0015] Furthermore, the external random interference data in S2 includes one or more of the following: ambient temperature, ambient humidity, ambient pressure changes, etc.; the internal random interference data includes one or more of the following: core temperature, gas flow changes, load changes, mode switching, etc. Taking multiple interference factors into consideration makes the interference factors controllable and reduces the interference of uncontrollable factors on battery control.

[0016] Furthermore, the processing method of each data in S3 is: the distribution of the collected data is analyzed by polynomial fitting, and the polynomial is f=a n r n +a n-1 r n-1 +…+a2r 2 +a1r+a0.

[0017] Furthermore, in S4, the basis function set {Φ i The basis functions in (r)} need to satisfy:

[0018]

[0019] In the formula, δ mn represents the Kronecker function. If m = n, δ mn =1; otherwise 0; <Φ m (r),Φ n (r)> defines the inner product:

[0020]

[0021] Where f is the multidimensional probability density function obtained by S4.

[0022] Furthermore, it also includes a starting base, which is a set of linearly independent polynomials, and the linearly independent polynomials are: a univariate polynomial A collection of .

[0023] Furthermore, according to the Gram-Schmidt method, the arbitrary chaotic polynomial basis function is constructed as:

[0024] In the formula, the coefficient c jk Defined as:

[0025] Further, in S5, the non-invasive method is: least squares estimation method. There are two methods for calculating the random model predictive control method based on chaotic polynomials, one is invasive and the other is non-invasive. Since the model of hydrogen fuel cells has strong nonlinearity and strong coupling, the non-invasive method is chosen to effectively reduce the huge amount of calculation caused by the strong nonlinearity and strong coupling of the model. The most common non-invasive method is the least squares method. Using the least squares method, the hydrogen fuel cell system is chaotically expanded using the basis function obtained above, which effectively handles the strong randomness of the hydrogen fuel cell, thereby obtaining a deterministic model. In this process, the influence of various random variables on the model is taken into account, the randomness is solved, and the control efficiency is improved. At the same time, the amount of calculation is reduced, thereby reducing the calculation time and improving the timeliness of control.

[0026] Furthermore, the hydrogen fuel cell includes a number of single-stack batteries, which constitute a multi-stack battery, and a number of battery cell subsystem models are arranged in a distributed manner to establish a multi-stack hydrogen fuel cell system model that simulates the working conditions of the multi-stack batteries. The battery cell subsystem models are arranged in a distributed manner, so that the computing tasks are dispersed to multiple nodes for parallel processing, which effectively improves the computing efficiency and the response speed of the system; at the same time, distributed control can also effectively solve the problem of computing load balancing in large-scale systems.

[0027] Furthermore, the method for constructing chaotic polynomial orthogonal basis functions is the Gram-Schmit method.

[0028] Combined with the above technical features, this solution also has the following technical effects:

[0029] 1. This technology takes into account the influence of uncertain factors and can more accurately reflect the operating status of the actual system by optimizing the objective function as desired.

[0030] 2. The present invention uses a distributed control method, so that each subsystem performs control tasks separately. At the same time, each subsystem coordinates and communicates with each other, so that each subsystem is organically combined into a whole, which greatly improves the computing efficiency and the response speed of the system. At the same time, distributed control can also effectively solve the problem of computing load balancing in large-scale systems.

[0031] 3. The present invention adopts a random model predictive control method, which can better cope with the uncertainty and changes of the system by establishing a random model of the system and making predictions. At the same time, the method can also continuously optimize the control strategy through online learning and model updating to adapt to the needs of dynamic changes in the system.

[0032] 4. The present invention can achieve real-time and robust control by using a distributed random model predictive control method. Even when there are random noise, interference or uncertainty factors in the system, it can still maintain a stable control effect and respond to system changes in a timely manner to achieve efficient range extension control of the hydrogen fuel cell system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of efficient range extension control according to an embodiment of the present invention;

[0034] Figure 2 A block diagram of efficient range-extending control according to an embodiment of the present invention;

[0035] Figure 3 This is a logic diagram of the efficient extended-range distributed control according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following is further described in detail through specific implementation methods:

[0037] Example

[0038] The embodiment is basically as shown in the attached Figure 1-3 As shown, Figure 1-3 A control method for efficiently extending the range of a hydrogen fuel cell is shown, comprising the following steps:

[0039] S1. Establish a hydrogen fuel cell system model;

[0040] S2, collecting the operating data of the hydrogen fuel cell system model established in S1, wherein the operating data includes external random interference data and internal random interference data;

[0041] S3, processing data, sorting and classifying the data collected in S2 according to the source of the collected data, obtaining the data of each parameter, and processing the data of each parameter separately to establish a multidimensional probability density function;

[0042] S4, using the multidimensional probability density function obtained in S3, using the Gram-Schmit method to construct a chaotic polynomial orthogonal basis function specific to hydrogen fuel cells, and obtaining the basis function set {Φ i (r)};

[0043] S5, using the chaotic polynomial orthogonal basis function obtained in S4 to propagate uncertainty in the hydrogen fuel cell system, and using a non-invasive calculation method and the chaotic polynomial orthogonal basis function to perform chaotic expansion on the hydrogen fuel cell system to handle randomness and obtain a deterministic model;

[0044] S6, constructing a multi-objective operation optimization control task for a hydrogen fuel cell system, and transforming the random control problem of the hydrogen fuel cell system into a deterministic control problem through the deterministic model obtained in S5;

[0045] S7. Formulate multi-objective operation optimization control tasks for the hydrogen fuel cell system according to demand to achieve efficient range-extending control of the hydrogen fuel cell system.

[0046] The specific implementation is as follows:

[0047] S1. Establishing a single-stack system model of a hydrogen fuel cell, and then establishing a multi-stack hydrogen fuel cell system model;

[0048] When establishing a hydrogen fuel cell system, it is necessary to consider the uncertainty of system parameters. In this embodiment, the hydrogen fuel cell monomer subsystem model is established as follows:

[0049] x p (k+1)=f(x p ,u p ,r p )

[0050] Among them, x p represents the measurable state variable of subsystem p, u p represents the input acting on subsystem p, r p represents the unknown model parameters of subsystem p. It is assumed that the unknown parameter vectors of each subsystem are independent of each other and there is a coupling relationship between the subsystems, such as Figure 2 , Figure 3 As shown, in this embodiment, the SMPC method is used to control each subsystem in the multi-stack battery system model to achieve mutual independence and coupling relationship between each subsystem.

[0051] S2. Collecting the operating data of the hydrogen fuel cell system through the sub-controller. The collected data includes the data of the normal operation of the hydrogen fuel cell system under various working conditions and environments. In this embodiment, the collected data includes the data of external random interference such as ambient temperature, humidity, pressure changes when the hydrogen fuel cell system is running, and the data of internal random interference such as the core temperature, gas flow change, load change, mode switching, etc. inside the hydrogen fuel cell.

[0052] S3. Sort and classify the collected data, process the data of each parameter separately, and then use polynomial fitting to analyze the distribution of the collected data. At the same time, further analyze the joint probability density relationship between the parameters. A random variable uses a one-dimensional probability density function. According to the actual situation, several random variables are considered to form a multi-parameter multi-dimensional joint probability density function f(r) of the subsystem, which is used to describe the distribution of all unknown parameters and the relationship between them, and express the relationship between these random variables.

[0053] In this embodiment, the polynomial used is

[0054] f = a n r n +a n-1 r n-1 +…+a2r 2 +a1r+a0

[0055] To describe the probability distribution of each acquisition parameter.

[0056] The collected parameters include external random interference data: external ambient temperature, external humidity, ambient temperature change; internal random interference data: core temperature, air flow change, load change and mode switching, etc., to obtain the six-dimensional probability density function f(r);

[0057] S4, using the six-dimensional probability density function obtained by the S3 data analysis, the Gram-Schmit method is used to construct a chaotic polynomial orthogonal basis function specific to hydrogen fuel cells;

[0058] Since the internal parameters of the hydrogen fuel cell subsystem are affected by the operating environment temperature, air pressure, etc., they are highly random, their distribution is uncertain, and they are non-Gaussian. Therefore, the basis function cannot be obtained by conventional division. It is necessary to use the six-dimensional probability density function obtained by S3 to construct the unique basis function set {′ i (r)}.

[0059] The basis function set {′ i The basis functions in (r)} need to satisfy:

[0060]

[0061] In the formula, δ mn represents the Kronecker function. If m = n, δ mn =1; otherwise 0; <Φ m (r),Φ n (r)> defines the inner product:

[0062]

[0063] Where f is the six-dimensional probability density function obtained by S4.

[0064] The basis function set {Φ i (r)} are constructed so that they are mutually orthogonal with respect to any probability density function. Starting from an arbitrary basis, the Gram-Schmidt orthogonalization method can be used to construct basis functions that are orthogonal to a given inner product. To this end, it is necessary to first construct a set of linearly independent polynomials as the starting basis. In this embodiment, the starting basis is a univariate polynomial A collection of .

[0065] According to the Gram-Schmidt method, the arbitrary chaotic polynomial basis function is constructed as:

[0066] Φ0(v)=1

[0067]

[0068] In the formula, the coefficient c jk Defined as:

[0069]

[0070] S5. Use the orthogonal basis functions obtained in S4 to propagate uncertainty in the hydrogen fuel cell system. In view of the high-dimensional, strongly coupled and strongly nonlinear characteristics of the hydrogen fuel cell system, use the least squares estimation method in a non-invasive manner to reduce the amount of calculation and effectively improve the control efficiency.

[0071] The hydrogen fuel cell system model equation is random, so it is considered a random process. Most previous work assumes that the system state x is random, but the control input u is deterministic. In fact, due to the existence of uncertainty, the system state x and control input u are random at every moment. Therefore, it is assumed that both the system state and control are random. Next, the idea of ​​chaotic polynomial expansion will be used to transform the hydrogen fuel cell system model random equation into a deterministic equation.

[0072] The present invention uses the polynomial chaos expansion method to propagate the uncertainty of the hydrogen fuel cell system parameters in the constructed model to quantify the uncertainty of the hydrogen fuel cell system output. At the same time, the polynomial chaos expansion method allows the model predictive control formula to be executed, the set value of the expected value is tracked, and the variance of the system output is minimized. Convergent expansion approximation The polynomial chaos expansion model of state variables and control variables is:

[0073]

[0074] In the formula, x p,il and u p,jl is the expansion coefficient, Φ l (r) is the m-order polynomial chaotic basis function with respect to the uncertainty parameter.

[0075] These basis functions belong to the Askey format of polynomials, which consists of a set of orthogonal basis functions defined by uncertainty parameter support in Hilbert space. Therefore, the basis functions satisfy <Φ i (r),Φ j (r)>=<Φ i (r) 2 >δ ij ,in<h(r),g(r)> =∫h(r)g(r)f θ dr = E[h(r)g(r)] represents the inner product of the sum over the domain Ω with respect to the weight f. For practical reasons, the chaotic polynomial must be truncated to a finite number of terms. The total number of terms is Dependence on the number of uncertain parameters n in the truncated expansion v And polynomial basis functions of order up to m are preserved in the expansion:

[0076]

[0077] The orthogonality of multivariate polynomials allows efficient computation of x i The PDF statistic of (t,r). For example, x i The first and second central moments of (t,r):

[0078] E[xp,i (t,r)]=x i0

[0079]

[0080] Since the hydrogen fuel cell system model is highly complex and can only be used as a black box physical model, the least squares estimation method in the non-intrusive scheme is selected. is a sample of M random variables, called an experimental design, which is implemented using space-filling sampling schemes such as Latin hypercube sampling and random Sobol sequences. The original model solver is then executed at each sample point to collect the corresponding target response The determination of polynomial chaotic expansion coefficients involves the problem of minimizing the residual l2 norm.

[0081]

[0082] At the same time, there are:

[0083]

[0084] Therefore, the least squares solution is:

[0085]

[0086] Among them, the coefficient u jl (t) can also be calculated using the same method.

[0087] Bringing the polynomial chaos expansion model into the hydrogen fuel cell subsystem model, we get:

[0088] To facilitate subsequent derivations, the following symbols are defined for the coefficients:

[0089]

[0090] Therefore, the hydrogen fuel cell subsystem model can be reformulated as:

[0091]

[0092] S6. Through S5, the random control problem of the hydrogen fuel cell system is transformed into a deterministic control problem, the impact of parameter randomness on the system operation efficiency is solved, and the multi-objective operation optimization control task of the hydrogen fuel cell system is formulated according to demand.

[0093] S7. Formulate multi-objective operation optimization control tasks for the hydrogen fuel cell system according to demand to achieve efficient range-extending control of the hydrogen fuel cell system.

[0094] The above is only an embodiment of the present invention, and common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that the technical means for solving the problems in the above-mentioned embodiments of the present invention can be used in combination to solve multiple technical problems at the same time. For those skilled in the art, several modifications and improvements can be made without departing from the technical solution of the present invention, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A control method for high-efficiency range extension of a hydrogen fuel cell, characterized in that: The steps include: S1. Establish a hydrogen fuel cell system model; S2, collecting the operating data of the hydrogen fuel cell system model established in S1, wherein the operating data includes external random interference data and internal random interference data; S3, processing data, sorting and classifying the data collected in S2 according to the source of the collected data, obtaining the data of each parameter, and processing the data of each parameter separately to establish a multidimensional probability density function; S4, using the multidimensional probability density function obtained in S3, construct the chaotic polynomial orthogonal basis function specific to hydrogen fuel cells, and obtain the basis function set {Φ i (r)}; S5, using the chaotic polynomial orthogonal basis function obtained in S4 to propagate uncertainty in the hydrogen fuel cell system, and using a non-invasive calculation method and the chaotic polynomial orthogonal basis function to perform chaotic expansion on the hydrogen fuel cell system to handle randomness and obtain a deterministic model; S6, constructing a multi-objective operation optimization control task for a hydrogen fuel cell system, and transforming the random control problem of the hydrogen fuel cell system into a deterministic control problem through the deterministic model obtained in S5; S7. Formulate multi-objective operation optimization control tasks for the hydrogen fuel cell system according to demand to achieve efficient range-extending control of the hydrogen fuel cell system.

2. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 1, characterized in that: The hydrogen fuel cell includes a single stack battery. In S1, the hydrogen fuel cell system model includes a battery monomer subsystem model that simulates the working conditions of the single stack battery. The battery monomer subsystem model is x p (k+1)=f(x p ,u p ,r p ).

3. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 2, characterized in that: The external random interference data in S2 include: one or more of ambient temperature, ambient humidity, ambient pressure changes, etc.; the internal random interference data include: one or more of core temperature, gas flow changes, load changes, mode switching, etc.

4. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 3, characterized in that: The processing method of each data in S3 is: the distribution of the collected data is analyzed by polynomial fitting, and the polynomial is d=a n r n +a n-1 r n-1 +…+a2r 2 +a1r+a0.

5. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 4, characterized in that: In S4, the basis function set {Φ i The basis functions in (r)} need to satisfy:

6. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 5, characterized in that: It also includes a starting base, which is a set of linearly independent polynomials, wherein the linearly independent polynomials are: a univariate polynomial A collection of .

7. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 6, characterized in that: According to the Gram-Schmidt method, the arbitrary chaotic polynomial basis function is constructed as:

8. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 7, characterized in that: In S5, the non-intrusive method is: least squares estimation method.

9. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 2, characterized in that: The hydrogen fuel cell includes a plurality of single-stack batteries, and the plurality of single-stack batteries constitute a plurality of stack batteries. A plurality of battery cell subsystem models are arranged in a distributed manner to establish a plurality of hydrogen fuel cell system models simulating the working conditions of the plurality of stack batteries.

10. A control method for high-efficiency range extension of a hydrogen fuel cell according to claim 1, characterized in that: The method for constructing chaotic polynomial orthogonal basis functions is the Gram-Schmit method.

Images