Denitration catalyst performance online evaluation method, device, equipment and storage medium

Abstract

The present invention provides a method, device, equipment and storage medium for online evaluation of denitration catalyst performance, which relates to the technical field of denitration catalytic reaction control. The method comprises: obtaining the inlet reactant concentration and reaction condition data of the denitration catalytic reaction; taking the inlet reactant concentration and reaction condition data as input, predicting the theoretical reaction rate of the reactant through a reaction rate model; taking the inlet reactant concentration and reaction condition data as input, obtaining the actual reaction rate of the reactant through a reaction rate mechanism model; taking the theoretical reaction rate of the reactant as the first input, taking the actual reaction rate of the reactant as the second input, outputting the theoretical performance evaluation value and the actual performance evaluation value of the catalyst through a catalyst performance evaluation model, and determining the performance state of the current catalyst based on the theoretical performance evaluation value and the actual performance evaluation value of the catalyst. The present invention solves the problem that the existing model has low accuracy and is difficult to be practical in engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of denitration catalytic reaction control, and in particular to a denitration catalyst performance online evaluation method, a denitration catalyst performance online evaluation device, a denitration catalyst performance online evaluation equipment and a machine-readable storage medium. Background Art

[0002] Selective catalytic reduction (SCR) flue gas denitrification technology is one of the most effective methods to remove NOx from coal-fired flue gas so far. In the actual operation of thermal power plants, as environmental protection requirements for emissions are further improved, factors such as uneven ammonia injection, catalyst pore blockage and catalyst poisoning and fire have an increasingly significant impact on catalyst performance. Due to the reduced activity of the catalyst, it is easy to cause an increase in NOx emissions or even exceed the standard; in order to ensure that NOx emissions meet the standard, it is necessary to increase the amount of ammonia injection, causing excess ammonia to react with SO3 in the flue gas to produce ammonium bisulfate and ammonium sulfate, which further reduces the activity of the catalyst, causing ash accumulation, blockage and corrosion in the air preheater, affecting the safe operation of the boiler, and increasing the amount of ammonia escape, which also causes waste of operating costs and secondary environmental pollution. In order to ensure the performance of the catalyst, the catalyst needs to be regenerated to improve the activity of the catalyst, reduce the amount of ammonia injection, and reduce ammonia escape. The existing technology usually regenerates the catalyst based on the experience of technicians or regularly, which can easily lead to untimely regeneration.

[0003] At present, the SCR denitration catalyst performance modeling methods include SCR catalyst modeling based on reaction mechanism and SCR catalyst modeling based on intelligent algorithm. SCR catalyst modeling based on reaction mechanism is to establish the catalytic reaction process equation of ammonia and NOx according to the catalytic reaction mechanism of the denitration catalyst, and obtain the relevant parameters of the reaction process equation through experiments. This method requires many parameters to be determined experimentally, and the experimental conditions are very different from the actual operating process conditions. Therefore, this method is difficult to accurately apply to the performance evaluation of catalysts under actual conditions.

[0004] The SCR catalyst modeling method based on intelligent algorithm mainly adopts neural network, support vector machine, deep learning and other methods. The key of this method is to find the input and output parameters required for modeling. The input parameters are relatively easy to find, but the weight distribution of the input parameters has a certain influence on the accuracy and generalization ability of modeling. For the output parameters, the denitrification efficiency is generally used to evaluate the performance of the denitrification catalyst. However, the denitrification efficiency is affected by factors such as operating load, ammonia injection amount, and NOx inlet concentration. It is difficult to use the denitrification efficiency to accurately evaluate the performance of the catalyst. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide a method for online evaluation of denitration catalyst performance, an apparatus for online evaluation of denitration catalyst performance, an equipment for online evaluation of denitration catalyst performance and a machine-readable storage medium to solve the problem that it is difficult to accurately evaluate catalyst performance in the prior art.

[0006] In order to achieve the above object, the first aspect of the present invention provides a method for online evaluation of denitration catalyst performance, the method comprising:

[0007] Obtaining the inlet reactant concentration and reaction condition data of the denitration catalytic reaction;

[0008] The theoretical reaction rate of the reactants is predicted by a reaction rate model using the inlet reactant concentration and reaction condition data as input, wherein the reaction rate model is constructed based on a long short-term memory network algorithm;

[0009] Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model, wherein the reaction rate mechanism model includes a method for determining the actual reaction rate of the reactant in the catalytic reaction mechanism;

[0010] The theoretical reaction rate of the reactant is used as the first input, and the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst. The actual reaction rate of the reactant is used as the second input, and the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst. The performance state of the current catalyst is determined based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst.

[0011] Optionally, the method further comprises:

[0012] Construct a catalyst performance prediction model based on the long short-term memory network algorithm;

[0013] A plurality of actual reaction rates of reactants obtained by the reaction rate mechanism model are obtained, and the catalyst performance evaluation values ​​of the next N steps are predicted by the catalyst performance prediction model using the obtained plurality of actual reaction rates as input.

[0014] Optionally, the inlet reactant concentration includes inlet NO concentration and inlet NH3 concentration, the reaction condition data includes reaction temperature, oxygen content, flue gas volume and flue gas pressure drop, and the reaction rate of the reactant includes the reaction rate of NO and the reaction rate of NH3.

[0015] Optionally, the reaction rate prediction model includes:

[0016] Forget gate, input gate and output gate:

[0017] Forget gate: f t =σ(W f ·[h t-1,x t ]+b f );

[0018] Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i );

[0019] Output gate: o t =σ(W o ·[h t-1 ,x t ]+b o );

[0020] The expression of the output value is: h t =o t *tanh(C t );

[0021] C t-1 Updated to C t The expression is: C′ t =tanh(W C ·[h t-1 ,x t ]+b C ), C t =f t *C t-1 +i t *C′ t ;

[0022] Among them, σ is the sigmoid function;

[0023] W f , W i , W C and W o They are the weights of the forget gate, input gate, tanh layer, and output gate respectively;

[0024] b f 、b i 、b C and b o They are the forget gate, input gate, tanh layer and output gate bias respectively;

[0025] h t-1 Output value at time t-1;

[0026] x t Input feature quantity for time t;

[0027] i t is the input layer vector;

[0028] C′t Cell state candidate vector;

[0029] C t is the unit state vector at time t;

[0030] o t is the output layer vector;

[0031] h t The predicted output at time t.

[0032] Optionally, the parameters in the reaction rate prediction model are determined using an adaptive momentum estimation gradient optimization algorithm.

[0033] Optionally, the reaction rate mechanism model includes:

[0034] r NO =(C′ NO -C″ NO ) / τ s ;

[0035]

[0036] τ s =V cat / V gas ;

[0037] Among them, r NO and r NH3 The total amount of NO and NH3 consumed by the denitration reaction and ammonia oxidation reaction per unit volume of catalyst per unit time, respectively, in mol / (s·m 3 );

[0038] C″ NO , C″ NH3 They are the NO concentration and NH3 concentration at the outlet of the denitration catalyst, in mol / m 3 ;

[0039] C′ NO , C′ NH3 They are the NO concentration and NH3 concentration at the inlet of the denitration catalyst, in mol / m 3 ;

[0040] τ s is the contact time, in seconds;

[0041] V cat is the catalyst volume, in m 3 ;

[0042] V gas is the flue gas volume, in Nm 3 / h.

[0043] Optionally, the catalyst performance evaluation model includes:

[0044]

[0045] Among them, η SCR is the performance evaluation value of the catalyst, is the reaction rate of NH3 at time t, is the reaction rate benchmark value of NH3, is the reaction rate of NO at time t, is the reaction rate reference value of NO, and A is the weight coefficient.

[0046] A second aspect of the present invention provides a denitration catalyst performance online evaluation device, the device comprising:

[0047] A data acquisition module configured to acquire inlet reactant concentration and reaction condition data of a denitration catalytic reaction;

[0048] A catalyst performance evaluation module is configured to use inlet reactant concentration and reaction condition data as input to predict the theoretical reaction rate of the reactant through a reaction rate model, wherein the reaction rate model is constructed based on a long short-term memory network algorithm;

[0049] Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model, wherein the reaction rate mechanism model includes a method for determining the actual reaction rate of the reactant in the catalytic reaction mechanism;

[0050] The theoretical reaction rate of the reactant is used as the first input, and the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst. The actual reaction rate of the reactant is used as the second input, and the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst. The performance state of the current catalyst is determined based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst.

[0051] The third aspect of the present invention provides a denitration catalyst performance online evaluation device, comprising:

[0052] at least one processor;

[0053] a memory connected to the at least one processor;

[0054] The memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the above-mentioned online evaluation method for the performance of the denitrification catalyst by executing the instructions stored in the memory.

[0055] A fourth aspect of the present invention provides a machine-readable storage medium having instructions stored thereon, which, when executed by a processor, configure the processor to execute the above-mentioned online denitration catalyst performance evaluation method.

[0056] The embodiments provided by the present invention have the following beneficial effects:

[0057] Based on the coupling of the denitrification catalytic mechanism model and the intelligent algorithm, the online evaluation of the denitrification catalyst performance is realized, which solves the problem that the relevant parameters obtained under experimental conditions based on the reaction mechanism of SCR catalyst modeling are too different from the actual operating conditions, and the denitrification efficiency is affected by factors such as operating load, ammonia injection amount, and NOx inlet concentration. At the same time, it solves the problem that the existing model is not accurate and difficult to be used in engineering.

[0058] Other features and advantages of the embodiments or implementations of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0060] Figure 1 A method flow chart schematically shows a method for online evaluation of denitration catalyst performance according to an embodiment of the present invention;

[0061] Figure 2 A schematic diagram of an evaluation logic of an online evaluation method for denitration catalyst performance according to an embodiment of the present invention is schematically shown;

[0062] Figure 3 A schematic diagram of the topological structure of an LSTM according to an embodiment of the present invention is shown;

[0063] Figure 4 A flowchart of LSTM algorithm hyperparameter optimization according to an embodiment of the present invention is schematically shown;

[0064] Figure 5 A flowchart of online modeling and prediction of an LSTM algorithm according to an embodiment of the present invention is schematically shown;

[0065] Figure 6 A schematic block diagram of a device for online evaluation of denitration catalyst performance according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0066] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0067] like Figure 1 As shown, the first aspect of this embodiment provides a method for online evaluation of denitration catalyst performance, the method comprising:

[0068] Obtaining the inlet reactant concentration and reaction condition data of the denitration catalytic reaction;

[0069] The theoretical reaction rate of the reactants is predicted by the reaction rate model using the inlet reactant concentration and reaction condition data as input. The reaction rate model is constructed based on the long short-term memory network algorithm.

[0070] Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model. The reaction rate mechanism model includes the determination method of the actual reaction rate of the reactant in the catalytic reaction mechanism;

[0071] The theoretical reaction rate of the reactant is used as the first input, and the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst. The actual reaction rate of the reactant is used as the second input, and the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst. The performance state of the current catalyst is determined based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst.

[0072] In this way, this embodiment realizes online evaluation of denitrification catalyst performance based on the coupling of denitrification catalytic mechanism model and intelligent algorithm, solves the problem that the relevant parameters obtained under experimental conditions based on reaction mechanism SCR catalyst modeling are too different from the actual operating conditions, and the denitrification efficiency is affected by factors such as operating load, ammonia injection amount, and NOx inlet concentration. At the same time, it solves the problem that the existing model is not accurate and difficult to be used in engineering.

[0073] In this embodiment, the inlet reactant concentration includes the inlet NO concentration and the inlet NH3 concentration, the reaction condition data includes the reaction temperature, the oxygen content, the flue gas volume and the flue gas pressure drop, the reactants include NO and NH3, and the reaction rates of the reactants include the reaction rate of NO and the reaction rate of NH3. It can be understood that the above parameters can be directly obtained online through the SCR denitration system.

[0074] The long short-term memory network (LSTM) algorithm is suitable for processing and predicting important events with very long intervals and delays in time series, and can be used as a complex nonlinear unit to construct larger deep neural networks. Figure 2As shown, this embodiment uses the advantages of LSTM algorithm in processing time series data, and establishes a reaction rate model of the catalytic reaction of the SCR denitration catalyst based on the LSTM algorithm in advance, taking the inlet NO concentration, the inlet NH3 concentration, and the reaction condition data including the reaction temperature, oxygen content, flue gas volume and flue gas pressure drop as input, and realizes the intermediate variable r through the reaction rate model NO and r NH3 The predicted output of the reaction rate model is r NO and r NH3 As the theoretical reaction rate of NO and NH3. In this embodiment, the reaction rate mechanism model is constructed based on the catalytic denitration of NO and ammonia oxidation mechanism model, which establishes the relationship between the inlet reactant concentration and reaction condition data and the actual reaction rate of NO and NH3, so that the actual reaction rate of NO and NH3 can be determined according to the obtained reaction data through the reaction rate mechanism model.

[0075] After obtaining the theoretical reaction rate and actual reaction rate of NO and NH3, the obtained theoretical reaction rate and actual reaction rate are respectively input into the catalyst performance evaluation model to obtain the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst. By comparing the difference evaluation values ​​between the two, and comparing the ideal state catalyst performance and the current state catalyst performance, the performance state of the current catalyst can be evaluated.

[0076] The method of this embodiment also includes: constructing a catalyst performance prediction model based on a long short-term memory network algorithm; obtaining multiple actual reaction rates of reactants obtained through a reaction rate mechanism model, using the obtained multiple actual reaction rates as input, and predicting the catalyst performance evaluation values ​​for the next N steps through the catalyst performance prediction model.

[0077] For example, collect a set of input data every 10 days, that is, the inlet reactant concentration and reaction condition data, and collect a total of 10 sets of input data; use the reaction rate mechanism model to calculate the catalyst performance of the 10 sets of data; then use the 10 sets of catalyst performance as the input of the catalyst performance prediction model built based on the LSTM intelligent algorithm to predict the catalyst performance of the next 5 steps, that is, predict the catalyst performance of the next 50 days. It can be understood that it is also possible to predict the catalyst performance of more steps in the future by increasing the number of input arrays, such as predicting the catalyst performance of the next 9 steps, that is, the catalyst performance of the next 90 days, thereby further realizing the prediction of future catalyst performance.

[0078] In this embodiment, the reaction rate prediction model adopts the long short-term memory network algorithm and is obtained by training it with samples. The construction process of the reaction rate prediction model mainly includes: model construction, sample acquisition, parameter tuning, model verification, online deployment and other stages. Figure 3As shown, the LSTM algorithm forget gate, input gate, output gate and output value are respectively expressed as follows:

[0079] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f );

[0080] Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i );

[0081] Output gate: o t =σ(W o ·[h t-1 ,x t ]+b o );

[0082] The expression of the output value is: h t =o t *tanh(C t );

[0083] C t-1 Updated to C t The expression is: C′ t =tanh(W C ·[h t-1 ,x t ]+b C ), C t =f t *C t-1 +i t *C′ t ; where σ is the sigmoid function; W f , W i , W C and W o are the weights of the forget gate, input gate, tanh layer and output gate respectively; b f 、b i 、b C and b o They are the forget gate, input gate, tanh layer and output gate bias respectively; h t-1 is the output value at time t-1; x t Input feature quantity at time t; i t is the input layer vector; C′ t Unit state candidate vector; C t is the unit state vector at time t; o t is the output layer vector; h tThe predicted output at time t.

[0084] like Figure 4 and Figure 5 As shown, the LSTM algorithm in this embodiment uses weights and hyperparameters to optimize simultaneously. The hyperparameters of the LSTM algorithm are optimized using a genetic algorithm (GA), and the weight optimization is optimized using an adaptive momentum estimation (Adam) gradient optimization algorithm. Historical data is collected through the PI system, and data standardization is achieved through the following relationship:

[0085]

[0086] Among them, x, Respectively represent arbitrary process data and normalized data; x max ,x min Respectively represent the maximum and minimum values ​​in the sample set of any process data. The collected data is divided into training data and test data. The training data is used to train the neural network module, and the test data is used to test the accuracy of the neural network. Specifically, the training process includes: inputting the training data, the set hyperparameters and the initialized weights into the LSTM algorithm, using the Adam algorithm to update and optimize the weights to obtain the optimal weights of the current hyperparameters, and then inputting the current optimal weights and initial hyperparameters into the LSTM algorithm. After the training data is calculated by the LSTM algorithm, the training results are output and compared with the actual results to test the training error. If the requirements are met, the optimal hyperparameters and optimal weights are output; if the error requirements are not met, the GA algorithm is used to update and optimize the hyperparameters, and the cycle is iterated until the error requirements are met, and the optimal hyperparameters and optimal weights are output. Among them, the error between the training results and the actual results is expressed by the root mean square error, and the root mean square error calculation formula is shown in the following formula:

[0087]

[0088] Among them, y i represents the actual value of the i-th dependent variable; Represents the predicted value of the ith dependent variable. The calculated root mean square error is compared with the set root mean square error value to decide whether to update the weight or output the optimal weight.

[0089] It can be understood that the catalyst performance prediction model can also use the above method to train the model and update related parameters.

[0090] Based on the analysis of the catalytic denitrification and NO removal and ammonia oxidation mechanism model, the reaction rates of the catalytic denitrification and NO removal and ammonia oxidation mechanism are expressed as follows:

[0091]

[0092]

[0093]

[0094]

[0095] Among them, r a is the adsorption rate, in s –1 ;k a is the adsorption reaction coefficient, in m 3 / (s·mol); r d is the desorption rate, in s –1 ;k d is the desorption reaction coefficient, in s –1 ; r NO The amount of NO consumed by the denitration reaction per unit volume of catalyst per unit time, in mol / (s·m 3 );r Ox The amount of NH3 consumed by the denitrification oxidation reaction per unit volume of catalyst per unit time, in mol / (s·m 3 );k NO is the denitrification reaction rate constant, in m 3 / (mol·s); k Ox is the reaction rate constant of ammonia oxidation, in m 3 / (mol·s); C′ NO is the NO concentration at the inlet of the denitration catalyst, in mol / m 3 θ NH3 is the NH3 coverage rate on the catalyst surface. It can be understood that the above parameters belong to reaction condition data.

[0096] According to the mass balance of NO and ammonia at the inlet and outlet of the denitrification catalyst, the above mechanism model is simplified to obtain the reaction rate mechanism model:

[0097] r NO =(C′ NO -C″ NO ) / τ s ;

[0098]

[0099] Among them, τ s =V cat / V gas ;

[0100] From the above formula, we can get r SCR , that is, r NO and r NH3 , where r NO and r NH3The total amount of NO and NH3 consumed by the denitration reaction and ammonia oxidation reaction per unit volume of catalyst per unit time, respectively, in mol / (s·m 3 );C″ NO , C″ NH3 They are the NO concentration and NH3 concentration at the outlet of the denitration catalyst, in mol / m 3 ; C′ NO , C′ NH3 They are the NO concentration and NH3 concentration at the inlet of the denitration catalyst, in mol / m 3 ; τ s is the contact time, in seconds; V cat is the catalyst volume, in m 3 ; V gas is the flue gas volume, in Nm 3 / h.

[0101] In this embodiment, the simplified mechanism model is used to obtain r NO and r NH3 The weighted summation gives the catalyst performance evaluation index η SCR ,Specifically, the catalyst performance evaluation model is as follows:

[0102]

[0103] Among them, η SCR is the performance evaluation value of the catalyst, is the reaction rate of NH3 at time t, is the reaction rate benchmark value of NH3, is the reaction rate of NO at time t, is the reaction rate reference value of NO, and A is the weight coefficient.

[0104] This implementation method simplifies the catalyst catalytic reaction mechanism model to avoid dependence on relevant parameters determined under laboratory conditions, making it practical for engineering; by coupling the catalyst catalytic reaction mechanism model and the intelligent algorithm model, an SCR denitration catalyst performance online evaluation and prediction model is established, and the NOx catalytic reaction rate and the ammonia catalytic reaction rate are weighted to obtain a catalyst performance evaluation index, thereby effectively solving the problem that the denitration efficiency as the output of the intelligent algorithm is difficult to accurately predict the catalyst performance, and realizing accurate online evaluation of the SCR catalyst performance and prediction of future performance.

[0105] By online collecting the operating parameters related to the NO concentration and ammonia slip at the outlet of the denitrification system, and using the optimal weights and hyperparameters, a catalyst performance evaluation model is established, which can adapt to the full-load operation of the boiler and the mixed combustion of multiple fuel types.

[0106] like Figure 6As shown, the second aspect of the present invention provides a denitration catalyst performance online evaluation device, the device comprising:

[0107] A data acquisition module configured to acquire inlet reactant concentration and reaction condition data of a denitration catalytic reaction;

[0108] A catalyst performance evaluation module is configured to use inlet reactant concentration and reaction condition data as inputs and predict theoretical reaction rates of reactants using a reaction rate model, where the reaction rate model is constructed based on a long short-term memory network algorithm;

[0109] Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model. The reaction rate mechanism model includes the determination method of the actual reaction rate of the reactant in the catalytic reaction mechanism;

[0110] The theoretical reaction rate of the reactant is used as the first input, and the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst. The actual reaction rate of the reactant is used as the second input, and the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst. The performance state of the current catalyst is determined based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst.

[0111] The specific definition of each functional module in the above-mentioned denitration catalyst performance online evaluation device can be found in the above-mentioned definition of the denitration catalyst performance online evaluation method, which will not be repeated here. Each module in the above-mentioned device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0112] The third aspect of the present invention provides an online denitration catalyst performance evaluation device, comprising: at least one processor; a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the above-mentioned online denitration catalyst performance evaluation method by executing the instructions stored in the memory.

[0113] It can be understood that the processor here has the functions of numerical calculation and logical operation, and it has at least a central processing unit CPU with data processing capability, a random access memory RAM, a read-only memory ROM, multiple I / O ports and an interrupt system. The processor includes a kernel, and the kernel calls the corresponding program unit from the memory. One or more kernels can be set, and the aforementioned method is implemented by adjusting the kernel parameters. The memory may include a non-permanent memory in a computer-readable medium, a random access memory (RAM) and / or a non-volatile memory, such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one storage chip.

[0114] A fourth aspect of the present invention provides a machine-readable storage medium having instructions stored thereon, which, when executed by a processor, configure the processor to execute the above-mentioned online denitration catalyst performance evaluation method.

[0115] In summary, the denitration catalyst performance evaluation model established in this embodiment is based on the denitration catalytic mechanism model and the intelligent algorithm. By analyzing the historical data through the established model, the online evaluation of the denitration catalyst performance and the prediction of future development trends are realized, and the high accuracy of the performance evaluation and the adaptability to all working conditions are guaranteed. At the same time, the problem that the relevant parameters obtained under experimental conditions based on the reaction mechanism of SCR catalyst modeling are too different from the actual operating conditions is solved, and the problem that the denitration efficiency is affected by factors such as operating load, ammonia injection amount, and NOx inlet concentration is solved.

[0116] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0117] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0118] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for online evaluation of denitration catalyst performance, characterized in that: The method comprises: Obtaining the inlet reactant concentration and reaction condition data of the denitration catalytic reaction; Taking the inlet reactant concentration and reaction condition data as input, the theoretical reaction rate of the reactant is predicted by a reaction rate prediction model, wherein the reaction rate prediction model is constructed based on a long short-term memory network algorithm; Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model, wherein the reaction rate mechanism model includes a method for determining the actual reaction rate of the reactant in the catalytic reaction mechanism; Taking the theoretical reaction rate of the reactant as the first input, the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst, taking the actual reaction rate of the reactant as the second input, the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst, and determining the performance state of the current catalyst based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst; The reaction rate mechanism model includes: r NO =(C′ NO -C′ N ′ O ) / t s ; r NH3 =(C′ NH3 -C′ N ′ H3 ) / t s ; τ s =V cat / V gas ; Among them, r NO and r NH3 The total amount of NO and NH3 consumed by the denitration reaction and ammonia oxidation reaction per unit volume of catalyst per unit time, respectively, in mol / (s·m 3 ); C″ NO , C″ NH3 They are the NO concentration and NH3 concentration at the outlet of the denitration catalyst, in mol / m 3 ; C′ NO , C′ NH3 They are the NO concentration and NH3 concentration at the inlet of the denitrification catalyst, in mol / m 3 ; τ s is the contact time, in seconds; V cat is the catalyst volume, in m 3 ; V gas is the flue gas volume, in Nm 3 / h; The catalyst performance evaluation model includes: Among them, η SCR is the performance evaluation value of the catalyst, is the reaction rate of NH3 at time t, is the reaction rate benchmark value of NH3, is the reaction rate of NO at time t, is the reaction rate reference value of NO, and A is the weight coefficient.

2. The method for online evaluation of denitration catalyst performance according to claim 1, characterized in that: The method further comprises: Construct a catalyst performance prediction model based on the long short-term memory network algorithm; A plurality of actual reaction rates of reactants obtained by the reaction rate mechanism model are obtained, and the catalyst performance evaluation values ​​of the next N steps are predicted by the catalyst performance prediction model using the obtained plurality of actual reaction rates as input.

3. The method for online evaluation of denitration catalyst performance according to claim 1, characterized in that: The inlet reactant concentration includes the inlet NO concentration and the inlet NH3 concentration, the reaction condition data includes the reaction temperature, oxygen content, flue gas volume and flue gas pressure drop, and the reaction rate of the reactant includes the reaction rate of NO and the reaction rate of NH3.

4. The method for online evaluation of denitration catalyst performance according to claim 1, characterized in that: The reaction rate prediction model includes: Forget gate, input gate and output gate: Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f ); Input gate: i t =σ(W i ·[h t-1 ,x t ]+b i ); Output gate: o t =σ(W o ·[h t-1 ,x t ]+b o ); The expression of the output value is: h t =o t *tanh(C t ); C t-1 Updated to C t The expression is: C t ′=tanh(W C ·[h t-1 ,x t ]+b C ), C t =f t *C t-1 +i t *C t ′; Among them, σ is the sigmoid function; W f , W i , W C and W o They are the weights of the forget gate, input gate, tanh layer, and output gate respectively; b f , b i , b C and b o They are the forget gate, input gate, tanh layer and output gate bias respectively; h t-1 Output value at time t-1; x t Input feature quantity for time t; i t is the input layer vector; C′ t Cell state candidate vector; C t is the unit state vector at time t; o t is the output layer vector; h t The predicted output at time t.

5. The method for online evaluation of denitration catalyst performance according to claim 4, characterized in that: The parameters in the reaction rate prediction model are determined by using a gradient optimization algorithm based on adaptive momentum estimation.

6. A denitration catalyst performance online evaluation device, characterized in that: The device comprises: A data acquisition module configured to acquire inlet reactant concentration and reaction condition data of a denitration catalytic reaction; A catalyst performance evaluation module is configured to use inlet reactant concentration and reaction condition data as input to predict the theoretical reaction rate of the reactant through a reaction rate prediction model, wherein the reaction rate prediction model is constructed based on a long short-term memory network algorithm; Taking the inlet reactant concentration and reaction condition data as input, the actual reaction rate of the reactant is obtained through the reaction rate mechanism model, wherein the reaction rate mechanism model includes a method for determining the actual reaction rate of the reactant in the catalytic reaction mechanism; Taking the theoretical reaction rate of the reactant as the first input, the catalyst performance evaluation model outputs the theoretical performance evaluation value of the catalyst, taking the actual reaction rate of the reactant as the second input, the catalyst performance evaluation model outputs the actual performance evaluation value of the catalyst, and determining the performance state of the current catalyst based on the theoretical performance evaluation value of the catalyst and the actual performance evaluation value of the catalyst; The reaction rate mechanism model includes: r NO =(C′ NO -C′ N ′ O ) / t s ; r NH3 =(C′ NH3 -C′ N ′ H3 ) / t s ; τ s =V cat / V gas ; Among them, r NO and r NH3 The total amount of NO and NH3 consumed by the denitration reaction and ammonia oxidation reaction per unit volume of catalyst per unit time, respectively, in mol / (s·m 3 ); C″ NO , C″ NH3 They are the NO concentration and NH3 concentration at the outlet of the denitration catalyst, in mol / m 3 ; C′ NO , C′ NH3 They are the NO concentration and NH3 concentration at the inlet of the denitration catalyst, in mol / m 3 ; τ s is the contact time, in seconds; V cat is the catalyst volume, in m 3 ; V gas is the flue gas volume, in Nm 3 / h; The catalyst performance evaluation model includes: Among them, η SCR is the performance evaluation value of the catalyst, is the reaction rate of NH3 at time t, is the reaction rate benchmark value of NH3, is the reaction rate of NO at time t, is the reaction rate reference value of NO, and A is the weight coefficient.

7. A denitration catalyst performance online evaluation device, characterized in that: include: at least one processor; a memory connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the online evaluation method for the performance of a denitration catalyst as described in any one of claims 1 to 5 by executing the instructions stored in the memory.

8. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions, which, when executed by a processor, configure the processor to execute the online denitration catalyst performance evaluation method according to any one of claims 1 to 5.

Images