A method and device for economic decision-making and optimization control of generator sets in power spot market

Abstract

The present invention belongs to the field of electric power automation and relates to a coal-fired unit optimization method, decision-making method, system, equipment and medium, including: searching for the same historical scenes and outputting them according to the real-time production data of DCS, and calculating the control output with the help of the established combustion system neural network model if the search fails, thereby solving the problem that the reliability of the current artificial intelligence algorithm is difficult to meet the actual production requirements, and improving the problems of the commonly used combustion test parameter determination method, such as the difficulty in applying for experimental conditions, the decrease in the accuracy of the test results over time, and the inability of the set parameters to adapt to the changes in the unit performance; by calculating the power generation cost of the unit, combining the information released by the power market, and obtaining the current optimal quotation through the Q-learning algorithm, the problems of the difficulty of data statistics for quotators, the low efficiency of information decision-making, and the high cost of quotation errors under the full liberalization of electricity prices are effectively solved.

Description

Technical Field

[0001] The invention belongs to the field of electric power automation and relates to a method and a device for economic decision-making and optimization control of a generator set in an electric power spot market. Background Art

[0002] With the construction of new power systems, the operation model of thermal power units has quietly changed. On the one hand, the proportion of new energy installed capacity has increased rapidly, and thermal power units not only need to ensure the rotational inertia and balance adjustment capabilities required by the power system, but also need to gradually transform the main body of electricity into the main body of capacity; on the other hand, the proposal of the "dual carbon" goal requires thermal power units to further deepen the low-carbon transformation and development, and "low-carbon, economic, safe, and flexible" have become the new keywords for the survival of coal-fired power plants.

[0003] However, at the current stage, the main capacity resources have much higher variable costs than new energy. In the scenario of a high proportion of new energy, it is difficult to recover all costs through existing manual decision-making methods. At the same time, technologies such as "carbon capture" are not yet mature, and the demand for deep peak regulation and rapid load change continues to increase, resulting in a decline in the economic efficiency of the units, which in turn leads to an increase in the cost of power generation. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art, in which the economic efficiency of the unit is difficult to guarantee due to the lack of technical means for optimizing the economic efficiency of the unit and incomplete power market declaration and decision-making, and to provide a method and device for economic decision-making and optimization control of power generating units in the power spot market.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: an adaptive online combustion optimization method for coal-fired units, comprising:

[0006] (1) Obtain production data from the DCS system and send the production data as input parameters to the combustion optimization control module;

[0007] (2) The combustion optimization control module converts the input parameters into an input case vector and adds a time tag for storing the case time;

[0008] (3) Based on the input case vector, a nearest neighbor algorithm is used to perform a global search in the combustion case database to compare the similarity between the target case and the historical case. If a historical case output set that meets the similarity threshold requirement is obtained, the historical case output set is weighted averaged to obtain a recommended output case set, which is then converted into corresponding oxygen content constants, secondary air damper openings at each layer, and rear air inlet damper openings; if the historical case set search return value is empty, the corresponding output value is obtained through the neural network model;

[0009] (4) The combustion optimization control module returns the output value to the DCS system, directly controls the controlled equipment or guides the operator to perform corresponding operations.

[0010] The production data include: unit load instruction, main steam pressure, main steam temperature, coal feeding instruction, reheat steam pressure, reheat steam temperature, furnace negative pressure, pulverizer operating status, coal calorific value, and ambient temperature.

[0011] The self-adaptive online combustion optimization method for coal-fired units further comprises the following steps:

[0012] The combustion optimization control module updates the calculation neural network model and case library according to a set fixed period.

[0013] In addition, the present invention also provides a power generation bidding decision method based on the Q-learning algorithm, comprising the following steps:

[0014] S1. The DCS system obtains the unit production data and sends it to the cost calculation module to calculate the current unit power generation cost;

[0015] S2. Send the corresponding parameters of the current unit power generation cost to the auxiliary bidding decision module through the cost calculation module; at the same time, obtain the prior market information released by the dispatching agency and the trading center, and import it into the auxiliary bidding decision module;

[0016] S3. Calculate the electricity price for each section based on the set operation target;

[0017] S4. Push the quotation decision result to the trading terminal, obtain the transaction result and post-market information and return it to the auxiliary bidding decision module. The auxiliary bidding decision module updates and calculates the bidding strategy parameters based on the transaction result through the Q-learning algorithm.

[0018] The cost calculation data include: unit coal consumption curve, plant power consumption rate, boiler efficiency, steam turbine heat consumption rate, coal calorific value, power generation cost approved by the Energy Bureau, coal purchase unit price, startup cost, shutdown cost, no-load cost, and fixed cost.

[0019] The specific method for updating and calculating the bidding strategy parameters of the power generation bidding decision method based on the Q-learning algorithm is as follows:

[0020] S501, generate a Q value table based on the market state s and the bidding strategy a and initialize it; determine the maximum learning round N;

[0021] S502, randomly generate a selection probability r. When r is greater than p, the market member will select any quotation strategy and expand the Q value table; when r is less than p, the market member will select the optimal quotation strategy based on the existing Q value table;

[0022] S503, integrating and recording the quotation data, and then calling the clearing calculation service to obtain the market clearing electricity price and clearing electricity quantity;

[0023] S504. Calculate the revenue u of the unit according to the revenue model i (t), and calculate the Q value of the corresponding state strategy, and then update the Q value table and strategy selection probability table;

[0024] S505. If the bidding round is equal to T, determine the optimal bidding strategy set and the learning ends; otherwise, repeat steps S502 to S504.

[0025] In addition, the present invention also provides a coal-fired unit economic decision-making calculation system, comprising:

[0026] Communication module, used to convert communication protocols to obtain DCS production and operation data;

[0027] Combustion optimization module, used for online adaptive calculation of combustion system optimization control output value;

[0028] A power generation cost calculation module, used for calculating the real-time power generation cost of the computer group;

[0029] The power generation bidding decision module is used to select the current bidding strategy through iterative calculation using the Q-Learning algorithm based on the power generation cost and the set operating objectives;

[0030] The isolation module is used to isolate the production system to ensure network security and production safety.

[0031] The combustion optimization module includes a case-based reasoning calculation module and a neural network calculation module. The case-based reasoning calculation module is used to search the case library for case-based reasoning calculation to obtain the output value of the controlled quantity according to the real-time production data of the combustion system. The neural network calculation module is used to calculate the output when the case-based reasoning calculation module has no output.

[0032] The power generation cost calculation module is used to obtain data from the real-time production system and the production management system, and calculate the fuel cost, auxiliary machine power consumption cost, environmental protection cost, and operation and maintenance cost of the computer group.

[0033] In addition, the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computer program.

[0034] In addition, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention discloses an adaptive online combustion optimization method for coal-fired units. The method searches for and outputs the same historical scenes according to the real-time production data of the DCS. If the search fails, the established neural network model of the combustion system is used to calculate the control output. The method solves the problem that the reliability of the current artificial intelligence algorithm is difficult to meet the actual production requirements. The method improves the problems of the commonly used combustion test parameter determination method, such as the difficulty in applying for experimental conditions, the decrease in the accuracy of the test results over time, and the inability of the set parameters to adapt to the changes in the unit performance.

[0037] The invention discloses a power generation bidding decision method based on Q-learning algorithm. By calculating the power generation cost of a computer group and combining it with the information released by the power market, the current optimal quotation is obtained through Q-learning algorithm, which effectively solves the problems of great difficulty in data statistics for quotators, low efficiency of information decision-making, high cost of quotation errors, etc. under the comprehensive liberalization of electricity prices. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a flow chart of an adaptive online combustion optimization method for a coal-fired unit according to Embodiment 1 of the present invention;

[0039] Figure 2 This is a structural diagram of a power plant boiler combustion system model based on a neural network in the first embodiment of the present invention;

[0040] Figure 3 A detailed flow chart of a power generation bidding decision method based on a Q-learning algorithm according to an embodiment of the present invention;

[0041] Figure 4 This is a structural diagram of the economic decision-making calculation system for a coal-fired unit according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0043] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] The present invention is further described in detail below with reference to the accompanying drawings.

[0045] Embodiment 1

[0046] See also Figure 1 In one embodiment of the present invention, an adaptive online combustion optimization method for coal-fired units is provided. The method searches for the same historical scenes and outputs them according to the real-time production data of the DCS. If the search fails, the control output is calculated by using the established neural network model of the combustion system. This solves the problem that the reliability of the current artificial intelligence algorithm is difficult to meet the actual production requirements. It also improves the problems of the commonly used combustion test parameter determination method, such as the difficulty in applying for experimental conditions, the decrease in the accuracy of the test results over time, and the inability of the set parameters to adapt to the changes in the unit performance.

[0047] Specifically, the adaptive online combustion optimization method for coal-fired units includes the following steps:

[0048] (1) The DCS system obtains production data such as unit load instructions, main steam pressure, main steam temperature, coal supply instructions, reheat steam pressure, reheat steam temperature, furnace negative pressure, pulverizer operating status, coal calorific value, ambient temperature, etc., and sends the above data as input parameters to the combustion optimization control module.

[0049] (2) The combustion optimization control module converts the input parameters into the input case vector C XYZ , and add a time tag to store the case time.

[0050]

[0051] Where: X P is the unit load;

[0052] X MA It is the operating status of coal mill A;

[0053] X MB It is the operating state of coal mill B;

[0054] X MC It is the operating state of coal mill C;

[0055] X MD It is the operating state of coal mill D;

[0056] X Q is the calorific value of coal;

[0057] X ET is the ambient temperature;

[0058] Y O2 It is the oxygen content value;

[0059] Y SAA , Y SAB The opening degree of the dampers on the A side and B side of the secondary air of the A layer;

[0060] Y SBA , Y SBB The opening degree of the dampers on the A and B sides of the secondary air of the B layer;

[0061] Y SCA , Y SCB The opening degree of the dampers on the A and B sides of the secondary air of the C layer;

[0062] Y SDA , Y SDB The opening degree of the dampers on the A and B sides of the secondary air of the D layer;

[0063] Y FA , Y FB is the opening of the secondary air baffle on the front wall of the rear air outlet;

[0064] Y RA , Y RB is the opening of the secondary air baffle on the rear wall of the rear air outlet;

[0065] T is the time tag.

[0066] (3) Based on the input case vector, a nearest neighbor algorithm is used to perform a global search in the combustion case database to calculate the target case X and the historical case X k The similarity S(X, X k ), compare it with the threshold, and obtain the output set of historical cases that meet the similarity threshold requirements.

[0067]

[0068] Where: x i The i-th eigenvalue of the target case X, X = [X P ,X MA ,X MB ,X MC ,XMD ,X Q ,X ET ]; x ik is the kth historical case X k The i-th eigenvalue of the input; X j =[X jP ,X jMA ,X jMB ,X jMC ,X jMD ,X jQ ,X jET ];ω i is the weight of the input i-th attribute parameter, the value is greater than zero and its sum is 1.

[0069] The similarity is greater than the threshold T h (T h =0.85) constitutes the historical case set H = {H XYZ}, if H is an empty set, indicating that there is no historical case matching the current target case in the case library, then calculation is performed through the neural network model; if H is not empty, further calculation is performed on the case set.

[0070] (4) The historical case output set is weighted averaged to obtain the recommended output case set, which is then converted into the corresponding oxygen content constant, the opening of the secondary air damper on each floor, and the opening of the rear air outlet damper.

[0071] Arrange the cases in the historical case set H from large to small according to their similarity, select the first m cases and perform weighted average according to the corresponding similarity to obtain the recommended output case set:

[0072]

[0073] Where: Si represents the similarity of the i-th case, and Yn represents the n-th output case.

[0074] (5) If the historical case set search return value is empty, it means that the target case has no corresponding historical case. Then, the production data obtained by the DCS system is input into the neural network model to obtain output values ​​including oxygen constant value, secondary air damper opening of each layer, and rear air outlet damper opening.

[0075] 1) Combustion system model framework

[0076] According to the correlation between the combustion system mechanism and variables, the present invention uses the unit load, coal calorific value, primary air pressure, secondary air pressure, opening of the secondary air damper of each layer of burners, flue gas oxygen content, and furnace temperature as the input values ​​of the nitrogen oxide prediction model, and uses the unit load, coal calorific value, primary air pressure, secondary air pressure, opening of the secondary air damper of each layer of burners, flue gas oxygen content, exhaust gas temperature, and fly ash carbon content as the input values ​​of boiler efficiency. The power station boiler combustion system model structure is as follows Figure 2 shown.

[0077] Figure 2 The meanings of the parameters are as follows:

[0078] LOAD is the unit load;

[0079] Q is the calorific value of coal;

[0080] P p is the primary wind pressure;

[0081] S p is the secondary wind pressure;

[0082] SE A , SE B , SE C , SE D , SE E , SE F They are the secondary air damper openings of A, B, C, D, E, and F layers respectively;

[0083] SOFA and CCOFA are the openings of the separate overburnt air and compact overburnt air dampers, respectively;

[0084] O2 is the oxygen content of flue gas;

[0085] C ash is the carbon content of fly ash;

[0086] T gas is the exhaust gas temperature;

[0087] T b is the furnace temperature;

[0088] ANN1 and ANN2 represent the neural network prediction models of boiler efficiency and nitrogen oxides, respectively;

[0089] η is the boiler efficiency;

[0090] NO x is the concentration of nitrogen oxides.

[0091] 2) Boiler efficiency model

[0092] When establishing the boiler efficiency model, the input layer of the neural network is determined to be 15 nodes and the output layer is 1 node. Two hidden layers are selected, one with 14 neurons and the other with 8 neurons. The training sample data is normalized and input into the network. The activation functions of the hidden layer and output layer of the network are set to tansig and logsig functions respectively, the network training function is traingdx, the network performance function is mse, the number of network iterations is 10,000, the expected error is 0.001, and the learning rate is 0.01.

[0093] 3) Flue gas inlet NOX concentration model

[0094] The input of the NOX concentration value model at the flue gas inlet is 14 variables, so the input layer of the neural network is determined to be 14 nodes, the output layer is 1 node, and two hidden layers are selected, one with 10 neurons and the other with 6 neurons. The training sample data is normalized and input into the network, and the activation functions of the network hidden layer and output layer are set to tansig and logsig functions respectively, the network training function is traingdx, the network performance function is mse, the number of network iterations is 10,000, the expected error is 0.001, and the learning rate is 0.01.

[0095] (6) The combustion optimization control module updates the calculation neural network model and case library according to the set fixed period. The case reasoning output is preferred because the case reasoning is a case library established based on real production data. That is, the case reasoning output is a scenario that has occurred in the past, and its safety is fully guaranteed.

[0096] (7) The combustion optimization control module returns the output value to the DCS system, directly controls the controlled equipment or guides the operator to perform corresponding operations.

[0097] Embodiment 2

[0098] See also Figure 3 In another embodiment of the present invention, a power generation bidding decision method based on a Q-learning algorithm is provided. The current optimal quotation is obtained by calculating the power generation cost of the computer group and combining it with the information released by the power market. This effectively solves the problems of difficulty in data statistics for bidders, low efficiency of information decision-making, and high cost of quotation errors under the full liberalization of electricity prices.

[0099] Specifically, the power generation bidding decision method based on the Q-learning algorithm includes the following steps:

[0100] (1) The DCS system obtains the unit production data and sends it to the cost calculation module to calculate the current unit power generation cost.

[0101] The cost of power generation usually consists of two parts: capacity cost and electricity cost. The capacity cost mainly includes equipment depreciation, construction cost, employee wages, etc. The electricity cost, as the name implies, is the cost required in the power generation process. The specific calculation formula is as follows:

[0102] C cost =C cap +C ele ; (3)

[0103] Where: C cap is the capacity cost, also known as fixed cost; C eleIt is the electricity cost, also known as the operating cost.

[0104] Capacity cost: Under the background of the new power system, my country's current power market system is difficult to guarantee the capacity benefits of traditional thermal power units. The capacity market or capacity compensation mechanism will be an important means to solve the capacity demand of the power system and ensure the healthy development of capacity resources. Therefore, accurate calculation of capacity cost is particularly important for units. The capacity cost calculation formula is as follows:

[0105] C cap =C dep +C sal +C mag ; (4)

[0106] Where: C dep is the depreciation cost; C sal is the wage cost; C mag To manage costs.

[0107] Electricity cost: Electricity cost is the main component of unit cost, accounting for a relatively high proportion; and electricity cost is directly related to unit characteristics, operating level, operating mode, etc. Electricity cost is composed of fuel cost, plant power cost, environmental protection cost and operation and maintenance cost, and the calculation formula is as follows:

[0108] C ele =C fuel +C plant +C env +C main ; (5)

[0109] Where: C fuel is the fuel cost; C plant is the factory electricity cost; C env Environmental protection cost; C main For operation and maintenance costs.

[0110] 1) Fuel cost: Fuel cost not only accounts for a large proportion, but also has a wide range of impact. On the one hand, fuel cost is affected by the price of primary energy market, especially in the context of tight global energy supply and urgent requirements for low-carbon transformation and development, the price is high and fluctuates frequently; on the other hand, fuel price is positively correlated with fuel quality, and different types of coal have a direct impact on fan power consumption, boiler efficiency, environmental protection investment, etc. The fuel cost calculation formula is as follows:

[0111] C fuel =M×S / P+C store ; (6)

[0112] Where: M is the coal supply (t / h); S is the average unit price of coal (yuan / t); P is the unit load (MW); C store The cost of coal storage.

[0113] 2) Plant power cost: Coal-fired power plants require a lot of high-energy-consuming equipment, such as coal mills, blowers, induced draft fans, water pumps, etc. In the early days, auxiliary motors were mostly operated at industrial frequency, but now most of them have been changed to variable frequency. Plant power cost accounts for a relatively high proportion of power generation costs, and its calculation method is as follows:

[0114] C plant =C mill +C fan +C desu +C deni +C dust_e ; (7)

[0115] Where: C mill is the electricity consumption cost of the milling system; C fan is the power consumption cost of the fan; C desu is the electricity consumption cost of the desulfurization system; C deni is the power consumption cost of the out-of-stock system; C dust_e The electricity consumption cost of the dust removal system.

[0116] 3) Environmental protection cost: Under the background of "dual carbon", energy saving and consumption reduction have become two major problems that coal-fired units must solve, and the two are contradictory to a certain extent. How to avoid crossing the red line of environmental protection and ensure the bottom line of economy has become a new focus of coal-fired units in production management. The formula for calculating environmental protection cost is as follows:

[0117] C env =C SO2 +C NOX +C NH3 +C smoke +C dust -C ash (8)

[0118] Among them, C SO2 It is the sewage discharge fee of the desulfurization system; C NOX NO X Sewage discharge fee; C NH3 Ammonia emission fee; C smoke The flue gas emission fee; C dust Dust emission fee; C ash For ash income.

[0119] 4) Operation and maintenance costs: Operation and maintenance costs refer to the maintenance costs incurred according to the maintenance plan. However, the incurrence of these costs has a certain probability, so the costs are usually evaluated based on historical data.

[0120] (2) The corresponding parameters of the current unit power generation cost are sent to the auxiliary bidding decision module through the cost calculation module.

[0121] (3) Obtain advance market information released by dispatching agencies and trading centers, and import it into the auxiliary bidding decision module.

[0122] (4) The electricity price reported for each section is calculated based on the set operating objectives.

[0123] (5) Push the quotation decision results to the trader.

[0124] (6) The trader returns the transaction results and post-market information released by the dispatching agency and the trading center to the auxiliary bidding decision module, and updates and calculates the bidding strategy parameters based on the transaction results through the Q-learning algorithm.

[0125] 1) Generate a Q-value table based on the market state s and bidding strategy a and initialize it. Q is Q(s,a), which is the expected value of the profit that can be obtained by adopting strategy a under the state s at a certain moment. The main idea of ​​the algorithm is to construct a Q-value table with state s and strategy a to store the Q-value, and then select the action that can obtain the maximum profit according to the Q-value, as shown in the following table; determine the maximum learning round n.

[0126] Q value table of market member behavior simulation

[0127]

[0128] 2) Randomly generate a selection probability r. When r is greater than the state transition probability p, market members will choose any bidding strategy and expand the Q value table. When r is less than p, market members will choose the optimal bidding strategy based on the existing Q value table.

[0129] 3) The quotation data is integrated and recorded, and then the clearing calculation service is called to obtain the market clearing electricity price and clearing electricity quantity.

[0130] 4) Calculate the revenue u according to the unit revenue model i (t), and calculate the Q value of the corresponding state strategy according to equations (9) to (11), and then update the Q value table and strategy selection probability table.

[0131]

[0132]

[0133]

[0134] Among them, α is the learning rate; γ is the discount factor; is the profit of the tth round of behavior; H is the new state The historical maximum Q value. t+1 (s,a) represents the Q value of the t+1th round, P t(s,a) represents the state transition probability, and k represents the corresponding strategy capacity.

[0135] It can be seen from the formula that the greater the learning rate, the smaller the discount factor, and the more market members value current returns; the smaller the learning rate, the larger the discount factor, and the more market members rely on historical experience.

[0136] 5) If the bidding round is equal to the maximum round T, the optimal bidding strategy set is determined and the learning ends; otherwise, repeat steps 2) to 4).

[0137] The following is a system embodiment of the present invention, which can be used to implement the method embodiment of the present invention. For details not disclosed in the system embodiment, please refer to the method embodiment of the present invention.

[0138] Embodiment 3

[0139] See also Figure 4 In another embodiment of the present invention, a coal-fired unit economic decision-making calculation system is provided, which can be used to implement the above-mentioned adaptive coal-fired unit online combustion optimization method and the power generation bidding decision method based on the Q-learning algorithm. Specifically, the coal-fired unit economic decision-making calculation system includes a communication module, a combustion optimization module, a power generation cost calculation module, a power generation bidding decision module and an isolation module.

[0140] Among them, the communication module is used to convert the communication protocol to obtain DCS production and operation data; the combustion optimization module is used to calculate the combustion system optimization control output value online adaptively, including the case reasoning calculation module and the neural network calculation module - that is, according to the real-time production data of the combustion system, search the case library for case reasoning calculation to obtain the output value of the controlled quantity, if there is no corresponding case in the case library, it is output by the neural network model; the power generation cost calculation module is used to calculate the real-time power generation cost of the computer group, that is, obtain data from the real-time production system and production management system, and calculate the fuel cost, auxiliary power consumption cost, environmental protection cost, operation and maintenance cost of the computer group; the power generation bidding decision module is used to assist in the decision-making of the unit quotation strategy, that is, according to the power generation cost and the set operation target, it is iteratively calculated through the Q-Learning algorithm to select the current bidding strategy. The isolation module is used to isolate the production system and the relationship system to ensure network security and production safety.

[0141] Embodiment 4

[0142] In another embodiment of the present invention, a terminal device is provided, the terminal device includes a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding functions.

[0143] Embodiment 5

[0144] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0145] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the economic decision calculation of the coal-fired unit in the above embodiment; one or more instructions in the computer-readable storage medium are loaded and executed by the processor.

[0146] 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.

[0147] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0148] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0149] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for calculating economic decision of a coal-fired unit, characterized in that: The invention comprises an adaptive online combustion optimization method for coal-fired units and a power generation bidding decision method based on a Q-learning algorithm. The adaptive online combustion optimization method for coal-fired units comprises: (1) Obtain production data from the DCS system and send the production data as input parameters to the combustion optimization control module; (2) The combustion optimization control module converts the input parameters into an input case vector and adds a time tag to store the case time; (3) Based on the input case vector, a nearest neighbor algorithm is used to perform a global search in the combustion case database to compare the similarity between the target case and the historical case. If a historical case output set that meets the similarity threshold requirement is obtained, the historical case output set is weighted averaged to obtain a recommended output case set, which is then converted into corresponding oxygen content constants, secondary air damper openings at each layer, and rear air inlet damper openings. If the historical case set search return value is empty, the corresponding output value is obtained through the neural network model. (4) The combustion optimization control module returns the output value to the DCS system, directly controls the controlled equipment or guides the operator to perform corresponding operations; The power generation bidding decision method based on the Q-learning algorithm comprises the following steps: S1. The DCS system obtains the unit production data and sends it to the cost calculation module to calculate the current unit power generation cost; S2. Send the corresponding parameters of the current unit power generation cost to the auxiliary bidding decision module through the cost calculation module; at the same time, obtain the prior market information released by the dispatching agency and the trading center, and import it into the auxiliary bidding decision module; S3. Calculate the electricity price for each section based on the set operation target; S4. Push the quotation decision result to the trading terminal, obtain the transaction result and post-market information and return it to the auxiliary bidding decision module. The auxiliary bidding decision module updates and calculates the bidding strategy parameters according to the transaction result through the Q-learning algorithm; The specific method for updating and calculating bidding strategy parameters is as follows: S501, generate a Q value table based on the market state s and the bidding strategy a and initialize it; determine the maximum learning round N; S502, randomly generate a selection probability r. When r is greater than p, the market member will select any quotation strategy and expand the Q value table; when r is less than p, the market member will select the optimal quotation strategy based on the existing Q value table; S503, integrating and recording the quotation data, and then calling the clearing calculation service to obtain the market clearing electricity price and clearing electricity quantity; S504. Calculate the revenue u of the unit according to the revenue model i (t), and calculate the Q value of the corresponding state strategy, and then update the Q value table and strategy selection probability table; S505. If the bidding round is equal to T, determine the optimal bidding strategy set and end the learning; otherwise, repeat steps S502 to S504.

2. The method for calculating economic decision of a coal-fired unit according to claim 1, characterized in that: The production data include: unit load instruction, main steam pressure, main steam temperature, coal feeding instruction, reheat steam pressure, reheat steam temperature, furnace negative pressure, pulverizer operating status, coal calorific value, and ambient temperature.

3. The method for calculating economic decision of a coal-fired unit according to claim 1, characterized in that: The following steps are also included: The combustion optimization control module updates the calculation neural network model and case library according to a set fixed period.

4. A coal-fired unit economic decision calculation system, used to implement the coal-fired unit economic decision calculation method according to claim 1, characterized in that: include: Communication module, used to convert communication protocols to obtain DCS production and operation data; Combustion optimization module, used for online adaptive calculation of combustion system optimization control output value; A power generation cost calculation module is used to calculate the real-time power generation cost of the computer group; The power generation bidding decision module is used to select the current bidding strategy through iterative calculations using the Q-Learning algorithm based on the power generation cost and the set operating objectives; The isolation module is used to isolate the production system to ensure network security and production safety.

5. The economic decision-making calculation system for coal-fired units according to claim 4, characterized in that: The combustion optimization module includes a case-based reasoning calculation module and a neural network calculation module. The case-based reasoning calculation module is used to search the case library for case-based reasoning calculation to obtain the output value of the controlled quantity according to the real-time production data of the combustion system. The neural network calculation module is used to calculate the output when the case-based reasoning calculation module has no output. The power generation cost calculation module is used to obtain data from the real-time production system and the production management system, and calculate the fuel cost, auxiliary machine power consumption cost, environmental protection cost, and operation and maintenance cost of the computer group.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the economic decision-making calculation method for a coal-fired unit as described in any one of claims 1 to 3 when executing the computer program.

7. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method for calculating economic decision-making of a coal-fired unit according to any one of claims 1 to 3.

Images