CN111313056B - Data-driven fuel cell online performance evaluation method - Google Patents

Data-driven fuel cell online performance evaluation method Download PDF

Info

Publication number
CN111313056B
CN111313056B CN202010141928.7A CN202010141928A CN111313056B CN 111313056 B CN111313056 B CN 111313056B CN 202010141928 A CN202010141928 A CN 202010141928A CN 111313056 B CN111313056 B CN 111313056B
Authority
CN
China
Prior art keywords
parameter
value
fuel cell
stack
catalyst
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010141928.7A
Other languages
Chinese (zh)
Other versions
CN111313056A (en
Inventor
吴迪
汤浩
殷聪
刘汝杰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010141928.7A priority Critical patent/CN111313056B/en
Publication of CN111313056A publication Critical patent/CN111313056A/en
Application granted granted Critical
Publication of CN111313056B publication Critical patent/CN111313056B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M8/00—Fuel cells; Manufacture thereof
    • H01M8/04—Auxiliary arrangements, e.g. for control of pressure or for circulation of fluids
    • H01M8/04298—Processes for controlling fuel cells or fuel cell systems
    • H01M8/04313—Processes for controlling fuel cells or fuel cell systems characterised by the detection or assessment of variables; characterised by the detection or assessment of failure or abnormal function
    • H01M8/04537—Electric variables
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
    • H01M8/00—Fuel cells; Manufacture thereof
    • H01M8/04—Auxiliary arrangements, e.g. for control of pressure or for circulation of fluids
    • H01M8/04298—Processes for controlling fuel cells or fuel cell systems
    • H01M8/04992—Processes for controlling fuel cells or fuel cell systems characterised by the implementation of mathematical or computational algorithms, e.g. feedback control loops, fuzzy logic, neural networks or artificial intelligence
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/30—Hydrogen technology
    • Y02E60/50—Fuel cells

Landscapes

  • Engineering & Computer Science (AREA)
  • General Chemical & Material Sciences (AREA)
  • Sustainable Development (AREA)
  • Sustainable Energy (AREA)
  • Manufacturing & Machinery (AREA)
  • Chemical & Material Sciences (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Electrochemistry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Fuel Cell (AREA)
  • Automation & Control Theory (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)

Abstract

The invention discloses an online performance evaluation method for a data-driven fuel cell, and belongs to the technical field of energy. The method can update the polarization curve parameter value in real time in the whole fuel cell operation process, and the result can be applied to power scheduling, service life prediction and the like; by applying a recursion formula, the data storage space is saved, and the calculation speed is increased, so that the configuration requirement of the controller is reduced; the actual physical significance of each parameter of the polarization curve is considered, the amplitude limit is carried out on each parameter estimation value, and the algorithm stability is improved; and estimating the performance of all the single batteries and providing more detailed information for the optimization of the control strategy.

Description

Data-driven fuel cell online performance evaluation method
Technical Field
The invention belongs to the technical field of energy, and particularly relates to an online performance evaluation method for a data-driven fuel cell.
Background
At present, the global energy crisis and the environmental pollution problem are increasingly prominent, and the proton exchange membrane fuel cell is widely concerned worldwide due to the characteristics of high efficiency, low noise and cleanness. Fuel cell performance can be represented by a polarization curve that shows the fuel cell voltage output at a given current output. In practical application, the actual voltage output of the fuel cell is lower than the ideal thermodynamic voltage, and the larger the output current is, the lower the output voltage is. The voltage loss is composed of three components, namely the activation loss due to electrochemical reactions, the ohmic loss due to ionic and electronic conduction, and the concentration loss due to mass transport. These three losses affect the characteristic shape of the polarization curve. In actual operation, due to fluctuation of cell operation conditions, different state of a cell stack and start-stop strategy and aging of a membrane electrode, a polarization curve is time-varying, and the performance of a fuel cell cannot be predicted online in real time by the polarization curve obtained through experiments. The fuel cell reaction principle is complex, a large number of accurate actual measurement parameters are needed for establishing a mechanism model for predicting the performance in real time, and a controller is required to have strong operation processing capacity, which is usually difficult to realize in actual operation.
Patent US 8214174B 2(2012.7.3) discloses an on-line adaptive polarization curve estimation algorithm for a fuel cell stack. When the fuel cell stack is operating and certain data validity criteria are met, the algorithm enters a data collection mode in which fuel cell data, such as stack current density, cell average voltage, and cell minimum voltage, are collected. When the stack is shut down, the algorithm estimates the predetermined parameters of the polarization curve model using a least squares method. If the estimated parameters meet certain termination criteria, the estimated parameters are stored for use by the system controller in calculating the polarization curve of the stack. The method collects data when the galvanic pile runs, obtains and stores an estimated value of a polarization curve parameter through batch learning when the galvanic pile stops, and estimates the polarization curve by using the parameter when the galvanic pile runs next time.
However, the above algorithm has the following disadvantages:
1) the estimation value of the polarization curve parameter is updated only after the operation process of the galvanic pile is finished once, so that the average performance of one section of operation process can be estimated only, and the parameter value cannot be updated in real time in the whole operation process of the galvanic pile.
2) Because of the use of the batch learning method, a large training data storage space and a long calculation time are required, and the algorithm time complexity and space complexity increase with the increase of the stack running time, which may result in insufficient controller resources.
3) The algorithm establishes a data collection sufficiency criterion, such as starting to collect data when the coolant temperature and humidity exceed certain set values. The data sufficiency standard discards partial measured data, the rest measured data are used as a training set for polarization curve parameter estimation, and the obtained model is local and cannot reflect the performance of the whole operation process. For example, during the startup process, the temperature is lower than the set threshold value, and the data in the stage is discarded, so the battery performance in the process is not reflected in the polarization curve model.
4) The algorithm does not consider the feasible domain of the polarization curve parameters, and the obtained result may exceed the actual physical value range, so that the performance estimation residual error is increased, and even the algorithm is not converged.
5) The average voltage and the minimum voltage are selected as training sets by the algorithm, and the result cannot reflect the polarization curve condition of each single cell of the stack.
The invention provides a data-driven fuel cell online performance evaluation method, which reduces algorithm complexity by a small amount of measured data and adopting a recursion method, and realizes accurate online estimation of fuel cell performance by online estimation of polarization curve model parameters.
Disclosure of Invention
The present invention is directed to overcoming the above-mentioned drawbacks of the prior art and providing a method for evaluating the on-line performance of a data-driven fuel cell.
The technical problem proposed by the invention is solved as follows:
a data-driven fuel cell online performance evaluation method comprises the following steps:
step 1, after the fuel cell is started, initializing the time value k to 1, and first setting the parameter estimator x to Rohm，iexch，ilim]TInitializing; if the fuel cell is started for the first time, the initialization x is ═ 4 × 10-4，0.01，530]TOtherwise, initializing x is the parameter x stored in the controller at the end of the last runs(i.e., x ═ x)s) Setting a learning rate lambda; wherein R isohmIs a resistance, iexchTo exchange current, ilimIs the limiting current;
step 2, collecting the output current i of the galvanic pile at the moment kstackTemperature T of catalyst layercatalystPartial pressure of oxygen
Figure BDA0002398643270000021
Partial pressure of hydrogen
Figure BDA0002398643270000022
And a cell voltage v;
step 3, filtering the sampling value at the moment k to obtain a moment k istack、Tcatalyst、
Figure BDA0002398643270000023
Estimated value of v:
Yk＝β·Xk+(1-β)·Yk-1
wherein beta is a filter coefficient, beta is more than 0 and less than 1, and XkIs k time istack、Tcatalyst、
Figure BDA0002398643270000024
value of v sampled, Yk-1Is k-1 time istack、Tcatalyst、
Figure BDA0002398643270000025
Estimate of v, Yk-1Initial value Y of0＝0；
Step 4, calculating thermodynamic voltage Ethermo：
Figure BDA0002398643270000026
Wherein, TcatalystIs the catalyst layer temperature, in K; r is an ideal gas constant, and R is 8.314J/(mol K); f is the faraday constant, F is 96485C/mol;
Figure BDA0002398643270000027
is the oxygen partial pressure in atm;
Figure BDA0002398643270000028
is the hydrogen partial pressure in atm.
Calculating two coefficients C1And C2：
Figure BDA0002398643270000031
Wherein n is the number of transmitted electrons in the reaction;
step 5, calculating the difference e between the electricity-saving compaction interval value of the single battery and the estimated value:
Figure BDA0002398643270000032
and 6, calculating the objective function gradient calculation formula of the single battery as follows:
Figure BDA0002398643270000033
step 7, calculating the parameter estimation value x of the monocell at the k +1 momentk+1：
xk+1＝xk-λg(xk)
Wherein λ ═ diag (λ)1，λ2，λ3) Diag denotes a diagonal matrix, λ1、λ2、λ3Are each Rohm、iexch、ilimThe learning rate of each parameter;
step 8, if the k +1 moment parameter estimated value xk+1(m)＜xinf，mM is more than or equal to 1 and less than or equal to 3, then let xk+1(m)＝xinf，m(ii) a If xk+1(m)＞xsup，mThen let xk+1(m)＝xsup，m(ii) a Wherein x isk+1(m) is a vector xk+1The m-th element of (2, x)inf，mIs a vector xk+1Lower bound of the mth element of (1), xsup，mIs a vector xk+1Upper bound of the mth element of (1);
and 9, estimating the performance of the single battery at the k +1 moment:
V＝Ethermo-ηohmic-ηact-ηconc
ηohmic＝Rohm·istack
Figure BDA0002398643270000034
Figure BDA0002398643270000035
wherein α is a charge transfer coefficient, ηohmicIs ohmic overvoltage, ηactTo activate overvoltage, etaconcIs concentration overvoltage;
step 10, making k equal to k + 1;
step 11, judging whether a shutdown state is entered; if yes, go to step 12; if not, turning to the step 2;
step 12, saving the estimated value of the parameter, i.e. order xs＝xkAnd the routine is ended.
The invention has the beneficial effects that:
the method can update the polarization curve parameter value in real time in the whole fuel cell operation process, and the result can be applied to power scheduling, service life prediction and the like; by applying a recursion formula, the data storage space is saved, and the calculation speed is increased, so that the configuration requirement of the controller is reduced; the actual physical significance of each parameter of the polarization curve is considered, the amplitude limit is carried out on each parameter estimation value, and the algorithm stability is improved; and estimating the performance of all the single batteries and providing more detailed information for the optimization of the control strategy.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 shows a resistor RohmThe change situation of the actual value and the estimated value is shown schematically;
FIG. 3 is a schematic diagram illustrating variation of estimation error of a cell voltage;
fig. 4 shows the results of evaluating the performance of the fuel cell at different times, wherein (a)100s, (b)150s, (c)200s, and (d)250 s.
Detailed Description
The invention is further described below with reference to the figures and examples.
The derivation process of the data-driven fuel cell online performance evaluation algorithm is divided into the following four steps:
(1) and selecting a performance evaluation model. The fuel cell output voltage may be represented by the following equation:
V＝Ethermo-ηohmic-ηact-ηconc (1)
wherein E isthermoThermodynamic voltage, in units of V; etaohmicIs ohmic overvoltage, in units of V; etaactFor activation ofOvervoltage, in units of V, etaconcIs the concentration overvoltage, in V.
Thermodynamic voltage EthermoCalculated according to the following formula:
Figure BDA0002398643270000041
wherein, Tcatal.tIs the catalyst layer temperature, in K; r is an ideal gas constant, and R is 8.314J/(mol K); f is the faraday constant, F is 96485C/mol;
Figure BDA0002398643270000042
is the oxygen partial pressure in atm;
Figure BDA0002398643270000043
is the hydrogen partial pressure in atm.
Ohmic overvoltage etaohmicCalculated according to the following formula:
ηohmic＝Rohm·istack (3)
wherein R isohmIs resistance, unit Ω; i.e. istackIs the stack output current, unit a.
Activation overvoltage etaactCalculated according to the following formula:
Figure BDA0002398643270000044
wherein alpha is a charge transfer coefficient and is dimensionless; n is the number of transmitted electrons in the reaction and is dimensionless; i.e. iexchFor exchange of current, unit a.
Concentration overvoltage etaconcCalculated according to the following formula:
Figure BDA0002398643270000051
wherein ilimIs the limiting current, in units a.
(2) And selecting parameters to be estimated. The parameters of the fuel cell output voltage model are selected as follows:
x＝[Rohm，iexch，ilim]T
where the superscript T denotes transpose.
(3) And determining the description of the optimization problem. The following two parameters are defined:
Figure BDA0002398643270000052
the difference e between the actual value and the estimated value of the node voltage is:
Figure BDA0002398643270000053
wherein, V is the actual measurement value of single section voltage, and unit V is through voltage inspection device on-line collection.
Obtaining an estimate x of x by minimizing the square E of the estimation error*Then the parameter estimation problem is described as the following optimization problem:
Figure BDA0002398643270000054
(4) and (4) designing an algorithm to solve an optimization problem.
The invention uses the steepest descent method to perform online learning and updating on the parameter x. I.e. the estimated parameters are corrected each time a new set of measured data is received. The objective function gradient is calculated as follows:
Figure BDA0002398643270000055
a parameter update formula in a recursive form can be obtained:
xk+1＝xk-λg(xk) (8)
where k is time, λ ═ diag (λ)1，λ2，λ3) Diag denotes a diagonal matrix, λ1，λ2，λ3And determining the correction quantity of each parameter for each iteration for the learning rate of each parameter. The size of each component of lambda is determined by the characteristics of the fuel cell, namely the component is matched with the resistance change rate of the membrane electrode, the change rate of the exchange current and the change rate of the limiting current, so that the parameter fluctuation near the optimal solution or the too slow parameter updating rate is avoided.
In sampling, if measurement noise is large, in order to avoid fluctuation of an algorithm, the measurement value needs to be filtered, and a specific algorithm of the estimated value at the time k is as follows:
Yk＝β·Xk+(1-β)·Yk-1 (9)
wherein, XkIs k time istack、Tcatalyst、
Figure BDA0002398643270000061
value of v sampled, Yk-1Is k-1 time istack、Tcatalyst、
Figure BDA0002398643270000062
The estimated value of v, β, is the filter coefficient (0 < β < 1), determining the filtering effect.
The embodiment provides an online performance evaluation method for a data-driven fuel cell, a flow chart of which is shown in fig. 1, and the method comprises the following steps:
step 1, after the fuel cell is started, initializing the time value k to 1, and first setting the parameter estimator x to Rohm，iexch，ilim]TInitializing; if the fuel cell is started for the first time, the initialization x is ═ 4 × 10-4，0.01，530]TOtherwise, initializing x is the parameter x stored in the controller at the end of the last runs(i.e., x ═ x)s) Setting a learning rate lambda; wherein R isohmIs a resistance, iexchTo exchange current, ilimIs the limiting current;
specifically, after the fuel cell system is started, the controller is powered on, and the control program defines the parameter estimation value variable x and then assigns an initial value xs. Defining a learning rate lambda based on the stack and the membraneThe electrode characteristics are given the values of the lambda components. The initial value x is an estimation result when the last time is finished, so that the parameter actual value cannot be excessively deviated, and the oscillation and divergence of the algorithm are avoided.
Step 2, collecting the output current i of the galvanic pile at the moment kstackTemperature T of catalyst layercatalystPartial pressure of oxygen
Figure BDA0002398643270000063
Partial pressure of hydrogen
Figure BDA0002398643270000064
And a cell voltage v;
specifically, the single cells are connected in series, so that the single cell current is istackThe cell performance difference is different from the v measurement value, and the cell voltage needs to be acquired so as to estimate the cell performance.
Step 3, filtering the sampling value at the moment k to obtain a moment k istack、Tcatalyst、
Figure BDA0002398643270000065
Estimated value of v:
Yk＝β·Xk+(1-β)·Yk-1
wherein beta is a filter coefficient, beta is more than 0 and less than 1, and XkIs k time istack、Tcatalyst、
Figure BDA0002398643270000066
value of v sampled, Yk-1Is k-1 time istack、Tcatalyst、
Figure BDA0002398643270000067
Estimate of v, Y0＝0；
Specifically, the invention uses a first-order digital low-pass filter to filter the measurement signal, and can obtain compromise between filtering stability and quick response by setting the filtering coefficient alpha.
Step 4, calculating thermodynamic voltage Ethermo：
Figure BDA0002398643270000071
Wherein, TcatalystIs the catalyst layer temperature, in K; r is an ideal gas constant, and R is 8.314J/(mol K); f is the faraday constant, F is 96485C/mol;
Figure BDA0002398643270000072
is the oxygen partial pressure in atm;
Figure BDA0002398643270000073
is the hydrogen partial pressure in atm.
Calculating two coefficients C1And C2：
Figure BDA0002398643270000074
Wherein n is the number of transmitted electrons in the reaction;
step 5, calculating the difference e between the electricity-saving compaction interval value of the single battery and the estimated value:
Figure BDA0002398643270000075
and 6, calculating the objective function gradient calculation formula of the single battery as follows:
Figure BDA0002398643270000076
step 7, calculating the parameter estimation value x of the monocell at the k +1 momentk+1：
xk+1＝xk-λg(xk)
Wherein λ ═ λ1，λ2，λ3]Learning rate of each parameter;
step 8, if the k +1 moment parameter estimated value xk+1(m)＜xinf，mM is more than or equal to 1 and less than or equal to 3, then let xk+1(m)＝xinf，m(ii) a If xk+1(m)＞xsup，mThen let xk+1(m)＝xsup，m(ii) a Wherein x isk+1(m) is a vector xk+1The m-th element of (2, x)inf，mIs a vector xk+1Lower bound of the mth element of (1), xsup，mIs a vector xk+1Upper bound of the mth element of (1);
specifically, in consideration of the actual physical significance, each parameter should be limited by an upper bound and a lower bound in the updating and estimating process, so that the oscillation and even non-convergence of the algorithm are avoided.
Step 9, estimating the performance of the monocell at the moment k +1, wherein the estimation result can be used for power scheduling, life prediction and the like;
V＝Ethermo-ηohmic-ηact-ηconc
ηohmic＝Rohm·istack
Figure BDA0002398643270000081
Figure BDA0002398643270000082
wherein α is a charge transfer coefficient, ηohmnicIs ohmic overvoltage, ηactTo activate overvoltage, etaconcIs concentration overvoltage;
step 10, making k equal to k + 1;
step 11, judging whether a shutdown state is entered; if yes, go to step 12; if not, turning to the step 2;
step 12, saving the estimated value of the parameter, i.e. order xs＝xkAnd the routine is ended.
FIG. 2 shows a resistor RohmThe change situation of the actual value and the estimated value is shown schematically.
FIG. 3 is a schematic diagram illustrating variation of estimation error of a cell voltage;
fig. 4 shows the results of evaluating the performance of the fuel cell at different times, wherein (a)100s, (b)150s, (c)200s, and (d)250 s.

Claims (1)

1. A method for evaluating the on-line performance of a data-driven fuel cell is characterized by comprising the following steps:
step 1, after the fuel cell is started, initializing the time value k to 1, and first setting the parameter estimator x to Rohm，iexch，ilim]TInitializing; if the fuel cell is started for the first time, the initialization x is ═ 4 × 10-4，0.01，530]TOtherwise, initializing x is the parameter x stored in the controller at the end of the last runs(ii) a Setting a learning rate lambda; wherein R isohmIs a resistance, iexchTo exchange current, ilimFor limiting current, superscript T denotes transposition;
step 2, collecting the output current i of the galvanic pile at the moment kstackTemperature T of catalyst layercatalystPartial pressure of oxygen
Figure FDA00026948899500000112
Partial pressure of hydrogen
Figure FDA0002694889950000011
And a cell voltage v;
step 3, filtering the sampling value at the moment k to obtain a moment k istack、Tcatalyst、
Figure FDA00026948899500000113
Estimated value of v:
Yk＝β·Xk+(1-β)·Yk-1
wherein beta is a filter coefficient, beta is more than 0 and less than 1, and XkIs k time istack、Tcatalyst、
Figure FDA0002694889950000013
value of v sampled, Yk-1Is k-1 time istack、Tcatalyst、
Figure FDA0002694889950000014
Estimate of v, Yk-1Initial value Y of0＝0；
Step 4, calculating thermodynamic voltage Ethermo：
Figure FDA0002694889950000015
Wherein, TcatalystIs the catalyst layer temperature, in K; r is an ideal gas constant, and R is 8.314J/(mol K); f is the faraday constant, F is 96485C/mol;
Figure FDA0002694889950000016
is the oxygen partial pressure in atm;
Figure FDA0002694889950000017
hydrogen partial pressure in atm;
calculating two coefficients C1And C2：
Figure FDA0002694889950000018
Wherein n is the number of transmitted electrons in the reaction;
step 5, calculating the difference e between the electricity-saving compaction interval value of the single battery and the estimated value:
Figure FDA0002694889950000019
and 6, calculating the objective function gradient calculation formula of the single battery as follows:
Figure FDA00026948899500000110
step 7, calculating the parameter estimation value x of the monocell at the k +1 momentk+1：
xk+1＝xk-λg(xk)
Wherein λ ═ diag (λ)1，λ2，λ3) Diag denotes a diagonal matrix, λ1、λ2、λ3Are each Rohm、iexch、ilimThe learning rate of each parameter;
step 8, if the k +1 moment parameter estimated value xk+1(m)＜xinf，mThen let xk+1(m)＝xinf，m(ii) a If xk+1(m)＞xsup，mThen let xk+1(m)＝xsup，m(ii) a Wherein m is more than or equal to 1 and less than or equal to 3, xk+1(m) is a vector xk+1The m-th element of (2, x)inf，mIs a vector xk+1Lower bound of the mth element of (1), xsup，mIs a vector xk+1Upper bound of the mth element of (1);
and 9, estimating the performance of the single battery at the k +1 moment:
V＝Ethermo-ηohmic-ηact-ηconc
ηohmic＝Rohm·istack
Figure FDA0002694889950000021
Figure FDA0002694889950000022
wherein α is a charge transfer coefficient, ηohmicIs ohmic overvoltage, ηactTo activate overvoltage, etaconcIs concentration overvoltage;
step 10, making k equal to k + 1;
step 11, judging whether a shutdown state is entered; if yes, go to step 12; if not, turning to the step 2;
step 12, saving the estimated value of the parameter, i.e. order xs＝xkAnd the routine is ended.
CN202010141928.7A 2020-03-03 2020-03-03 Data-driven fuel cell online performance evaluation method Active CN111313056B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010141928.7A CN111313056B (en) 2020-03-03 2020-03-03 Data-driven fuel cell online performance evaluation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010141928.7A CN111313056B (en) 2020-03-03 2020-03-03 Data-driven fuel cell online performance evaluation method

Publications (2)

Publication Number Publication Date
CN111313056A CN111313056A (en) 2020-06-19
CN111313056B true CN111313056B (en) 2020-12-29

Family

ID=71147872

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010141928.7A Active CN111313056B (en) 2020-03-03 2020-03-03 Data-driven fuel cell online performance evaluation method

Country Status (1)

Country Link
CN (1) CN111313056B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114566686B (en) * 2020-11-27 2023-11-14 中国科学院大连化学物理研究所 Method for evaluating state and predicting service life of fuel cell
CN112563542B (en) * 2020-12-08 2022-02-11 上海重塑能源科技有限公司 Fuel cell online detection method and detection system
CN114628742A (en) * 2020-12-11 2022-06-14 中国科学院大连化学物理研究所 Measurement system and measurement method for fuel cell polarization curve
CN114335628B (en) * 2021-12-29 2023-11-14 上海重塑能源科技有限公司 Non-storage type online iterative estimation method of polarization curve and fuel cell system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8214174B2 (en) * 2007-01-31 2012-07-03 GM Global Technology Operations LLC Algorithm for online adaptive polarization curve estimation of a fuel cell stack
CN104051755A (en) * 2013-03-15 2014-09-17 通用汽车环球科技运作有限责任公司 Systems and methods for predicting polarization curves in a fuel cell system
CN108226789A (en) * 2017-11-21 2018-06-29 东南大学 A kind of performance test methods of intermediate temperature solid oxide fuel cell
DE102018129845A1 (en) * 2017-11-29 2019-05-29 GM Global Technology Operations LLC Fuel cell stack cell voltage sensor diagnostics
CN110277572A (en) * 2018-03-14 2019-09-24 现代自动车株式会社 For estimating the method and system of the available output of fuel cell

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8214174B2 (en) * 2007-01-31 2012-07-03 GM Global Technology Operations LLC Algorithm for online adaptive polarization curve estimation of a fuel cell stack
CN104051755A (en) * 2013-03-15 2014-09-17 通用汽车环球科技运作有限责任公司 Systems and methods for predicting polarization curves in a fuel cell system
CN108226789A (en) * 2017-11-21 2018-06-29 东南大学 A kind of performance test methods of intermediate temperature solid oxide fuel cell
DE102018129845A1 (en) * 2017-11-29 2019-05-29 GM Global Technology Operations LLC Fuel cell stack cell voltage sensor diagnostics
CN110277572A (en) * 2018-03-14 2019-09-24 现代自动车株式会社 For estimating the method and system of the available output of fuel cell

Also Published As

Publication number Publication date
CN111313056A (en) 2020-06-19

Similar Documents

Publication Publication Date Title
CN111313056B (en) Data-driven fuel cell online performance evaluation method
CN111948560A (en) Lithium battery health state estimation method based on multi-factor evaluation model
CN112505568B (en) Multi-stack solid oxide fuel cell system service life prediction method
CN111044906B (en) Lithium ion battery energy state estimation method based on maximum likelihood criterion
CN205898985U (en) Power lithium cell SOC estimates system for pure electric vehicles
CN112858920B (en) SOC estimation method of all-vanadium redox flow battery fusion model based on adaptive unscented Kalman filtering
CN112462282B (en) Method for determining real-time state of charge of battery pack based on mechanism model
CN112687926A (en) Method and system for predicting attenuation trend and predicting remaining service life of fuel cell
CN112816879A (en) Online estimation method for power battery SoE for variable-working-condition vehicle
CN113484771A (en) Method for estimating wide-temperature full-life SOC and capacity of lithium ion battery
CN111063919B (en) Fuel cell anode state monitoring method
CN116259797A (en) Health assessment and prediction method and system for solid oxide fuel cell system
CN114335628B (en) Non-storage type online iterative estimation method of polarization curve and fuel cell system
CN115267585A (en) Improved single particle electrochemical model modeling method based on gradient enhanced decision tree
CN107402356B (en) EKF estimation lead-acid battery SOC method based on dynamic parameter identification
CN113125969B (en) Battery data processing method, device and medium based on AUKF
CN115693916A (en) Intelligent online monitoring method and system for direct-current power supply of transformer substation
CN115422846A (en) Aging modeling method, apparatus, device and medium for fuel cell
CN115015792A (en) Online parameter identification method for energy storage battery sub-module
CN114566686B (en) Method for evaluating state and predicting service life of fuel cell
CN116859251B (en) Multi-step hybrid prediction method and system for fuel cell considering recovery voltage
CN116973791A (en) Proton exchange membrane fuel cell life prediction method and system considering uncertainty
CN113447818B (en) Identification method and system of battery equivalent circuit model
CN115079003A (en) Method and system for estimating state of charge value of battery, electronic equipment and vehicle
CN115421057A (en) Lithium battery SOE estimation method and system based on adaptive unscented Kalman filtering

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant