A positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method

Abstract

The present application discloses a positive pressure biomass gasification furnace and a coal-fired boiler coupling power generation method, which relates to the field of biomass energy processing. The method comprises: obtaining first furnace environment information in the positive pressure biomass gasification furnace, the first furnace environment information comprising first pressure information, first temperature information and first gas composition information; obtaining second furnace environment information in the coal-fired boiler, the second furnace environment information comprising second pressure information, second temperature information and second gas composition information; adjusting the flow of a gas delivery pipeline according to the first furnace gas environment information and the second furnace gas environment information, the gas delivery pipeline being a gas delivery pipeline connecting the positive pressure biomass gasification furnace and the coal-fired boiler, and a gas temporary storage tank being further arranged in the gas delivery pipeline; after the flow of the flow delivery pipeline is adjusted for a preset time, adjusting the first temperature information and the second temperature information according to the first furnace gas environment information and the second furnace gas environment information.

Description

Technical Field

[0001] The present specification relates to biomass energy processing, and more specifically, the present application relates to a positive pressure biomass gasification furnace and a coal-fired boiler coupling power generation method. Background Art

[0002] In recent years, with the growth of energy demand and the enhancement of environmental protection awareness, the use of various energy sources such as biomass energy and coal to achieve efficient power generation has gradually become a research hotspot. Among them, positive pressure biomass gasification technology is widely used because it can convert biomass into combustible gases (such as hydrogen and carbon monoxide), while coal-fired boilers have always occupied an important position in cogeneration due to their maturity and high energy density.

[0003] In the coupled power generation system, the coupling of the positive pressure biomass gasifier and the coal-fired boiler can make full use of the complementary characteristics of the two energy sources: the positive pressure biomass gasifier can provide clean combustion-supporting gas for the coal-fired boiler, improve combustion efficiency, and reduce coal consumption. However, the coupled operation of this system also faces the following technical challenges:

[0004] 1. The coupling control of the operating parameters of the two furnaces is complex: the operating conditions of the positive pressure biomass gasifier and the coal-fired boiler are quite different. For example, the gasifier requires a lower temperature and a higher pressure to ensure the gasification reaction, while the coal-fired boiler requires a high temperature environment to ensure the complete combustion of coal. How to coordinate the operating parameters of the two to achieve stable coupling operation becomes the key.

[0005] 2. Instability of gas pipeline flow: During the gas transportation process between the gasifier and the boiler, the gas flow is easily affected by the fluctuation of the gasification furnace gas production rate and the change of boiler combustion demand. If the flow fluctuation is too large, it may lead to instability in the operation of the coal-fired boiler.

[0006] 3. Poor adaptability to dynamic environment: The operating environment of the two furnaces (such as pressure, temperature and gas composition) changes with changes in external conditions and load requirements. The existing technology often lacks a dynamic adjustment method based on real-time environmental information, which easily causes system response hysteresis or instability.

[0007] 4. Limited efficiency and stability of coupled power generation: The lack of a systematic control strategy in coupled operation may cause the gasifier to be unable to operate within the optimal temperature range, and the combustion efficiency of the coal-fired boiler will also be difficult to achieve the optimal level, thereby reducing the overall power generation efficiency.

[0008] Therefore, it is necessary to propose a positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method to at least solve some of the above problems. Summary of the invention

[0009] A series of simplified concepts are introduced in the Summary of the Invention section, which will be further described in detail in the Detailed Description of the Invention section. The Summary of the Invention section of this application does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the scope of protection of the claimed technical solution.

[0010] In the first aspect, the present application proposes a positive pressure biomass gasification furnace and a coal-fired boiler coupling power generation method, the method comprising:

[0011] Acquire first furnace environment information in the positive pressure biomass gasification furnace, wherein the first furnace environment information includes first pressure information, first temperature information and first gas composition information;

[0012] Acquire second furnace environment information in the coal-fired boiler, wherein the second furnace environment information includes second pressure information, second temperature information and second gas composition information;

[0013] According to the gas environment information in the first furnace and the gas environment information in the second furnace, the flow rate of the gas delivery pipeline is adjusted, wherein the gas delivery pipeline is a gas delivery pipeline connecting the positive pressure biomass gasification furnace and the coal-fired boiler, and a gas temporary storage tank is also provided in the gas delivery pipeline;

[0014] After the flow rate regulation operation of the gas delivery pipeline is completed for a preset time period, the first temperature information and the second temperature information are adjusted according to the first furnace gas environment information and the second furnace gas environment information.

[0015] In a feasible implementation manner, the above-mentioned adjusting the flow rate of the gas delivery pipeline according to the above-mentioned first furnace gas environment information and the above-mentioned second furnace gas environment information includes:

[0016] The flow rate of the gas delivery pipeline is adjusted through a reinforcement learning algorithm according to the above-mentioned first furnace gas environment information and the above-mentioned second furnace gas environment information.

[0017] In a feasible implementation manner, the gas delivery pipeline flow rate includes a first gas flow rate corresponding to the positive pressure biomass gasification furnace and a second gas flow rate corresponding to the coal-fired boiler;

[0018] The method of adjusting the flow rate of the gas delivery pipeline by using a reinforcement learning algorithm according to the first furnace gas environment information and the second furnace gas environment information includes:

[0019] The state vector is constructed based on the first furnace gas environment information and the second furnace gas environment information. :

[0020] ;

[0021] is the first pressure information, is the first temperature information, is the first gas composition information, is the second pressure information, is the second temperature information, is the above-mentioned second gas composition information;

[0022] According to the first gas flow rate and the second gas flow rate Constructing pipeline flow regulation actions :

[0023] ;

[0024] in, is the first gas flow rate, is the second gas flow rate mentioned above;.

[0025] Define the reward function based on the control objective :

[0026] ;

[0027] in, is the target value of the first pressure information or the target value of the second pressure information, is the target value of the first temperature information or the target value of the second temperature information, is the target value of the first gas composition information or the target value of the second gas composition information, Coupling_Error is the coupling error between furnaces, is the first weight information corresponding to the target value of the first pressure information or the target value of the second pressure information, is a second weight coefficient corresponding to the target value of the first temperature information or the target value of the second temperature information, is the weight coefficient corresponding to the target value of the first gas composition information or the target value of the second gas composition information, is the weight coefficient corresponding to the coupling error;

[0028] Construct a coupled model of the positive pressure biomass gasifier and the coal-fired boiler:

[0029] ;

[0030] in, is the value of the first temperature information at the next moment, is the value of the first pressure information at the next moment, is the value of the first gas composition information at the next moment, is the value of the second temperature information at the next moment, is the value of the second pressure information at the next moment, is the value of the second gas composition information at the next moment, is the current value of the first temperature information, is the current value of the first pressure information, is the current value of the first gas composition information, is the current value of the second temperature information, is the current value of the second pressure information, The current value of the second gas composition information, is the first gas flow rate, is the second gas flow rate, is random noise;

[0031] The above coupling model and historical experimental data are reinforced by reinforcement learning algorithm to improve the above reward function Corresponding cumulative rewards and obtaining the optimized strategy network;

[0032] The flow rate of the gas delivery pipeline is adjusted according to the gas environment information in the first furnace and the gas environment information in the second furnace through the above-mentioned optimization strategy network.

[0033] In one possible implementation, the above-mentioned reinforcement algorithm includes a deep deterministic policy gradient algorithm.

[0034] In a feasible implementation, the above coupling model and historical experimental data are reinforced by reinforcement learning algorithm to improve the above reward function The corresponding cumulative rewards and the optimized strategy network are obtained, including:

[0035] Create a policy network and a Q network based on the above coupling model

[0036] Create an experience replay pool based on the above historical experimental data ;

[0037] According to the current status , generate actions through the above strategy network ;

[0038] Execute an action , record the reward function and the next state ;

[0039] Will Stored in the above experience replay pool D;

[0040] Calculate target value :

[0041] ;

[0042] in, is the above reward function at time t, is the discount factor, For Q network in the next state and the action output by the policy network The predicted cumulative reward value under is the weight of the Q network;

[0043] Calculate the loss function of the Q network based on the above target value y , to update the Q network parameters :

[0044] ;

[0045] in, are the parameters of the Q network, Represents experience pool The expected value is calculated for the samples in is the target value With the Q network in its current state and the action output by the policy network The predicted cumulative reward value under , , is the weight of the Q network;

[0046] Calculate the gradient of the policy network to maximize the Q value and update the policy network parameters :

[0047] ;

[0048] in, is the gradient of the policy network, Represents a random sampling state from the experience replay pool D And the expected value of the gradient of these samples, Indicates the output value of the Q network action Find the partial derivative of, Represents the policy network output parameter The partial derivative of are the parameters of the policy network;

[0049] Update the target Q network according to the preset update coefficient and target strategy network , to obtain the above optimization strategy network:

[0050] ;

[0051] in, It is the main Q network The delayed version of the target policy network Main Strategy Network A delayed version of are the parameters of the target Q network, The parameters of the target policy network, are the parameters of the main network, are the parameters of the main strategy network, is the soft update coefficient, Represents an update operation.

[0052] In a feasible implementation, the gas delivery pipeline flow rate includes the inlet flow rate of the positive pressure biomass gasification furnace and the outlet flow rate of the coal-fired boiler.

[0053] The above method further includes:

[0054] When a first difference between the inlet flow rate of the positive pressure biomass gasification furnace and the outlet flow rate of the coal-fired boiler is greater than a preset difference, obtaining excess flow difference information between the first difference and the preset difference;

[0055] The gas storage flow corresponding to the gas temporary storage tank is controlled based on the excess flow difference information.

[0056] In a feasible implementation manner, the adjusting the first temperature information and the second temperature information according to the first furnace gas environment information and the second furnace gas environment information includes:

[0057] The first temperature information and the second temperature information are predicted and controlled according to the first furnace gas environment information, the second furnace gas environment information and the target neural network model.

[0058] In a feasible implementation, the target neural network model includes an input layer, a feature extraction layer, an intermediate fusion layer, a dynamic time series layer, a fully connected layer and an output layer;

[0059] The above-mentioned feature extraction layer includes a temperature feature branch, a pressure feature branch, a gas composition variable branch and a control variable branch;

[0060] The intermediate fusion layer is used to perform a vector fusion operation on the feature vectors extracted from the temperature feature branch, the pressure feature branch, the gas composition variable branch and the control variable branch to form an intermediate feature vector;

[0061] The above dynamic time series layer generates dynamic feature vectors based on historical time series based on long short-term memory network;

[0062] The fully connected layer includes a multi-layer fully connected network, and the fully connected layer is used to predict and generate the first temperature information and the second temperature information according to the intermediate feature vector and the dynamic feature vector;

[0063] The output layer is used to output the first temperature information and the second temperature information.

[0064] In a feasible implementation, the vector fusion operation is a feature weighted fusion operation, the long short-term memory network is a bidirectional long short-term memory network, and each fully connected network of the fully connected layer includes a Dropout module and a Batch Normalization module.

[0065] In a feasible implementation manner, the above method further includes:

[0066] Acquire temperature difference information between the first temperature information and the second temperature information;

[0067] When the temperature difference information exceeds the preset temperature difference, the flow rate of the gas delivery pipeline is adjusted again according to the first furnace gas environment information and the second furnace gas environment information.

[0068] In summary, the present application provides a method for coupling power generation between a positive pressure biomass gasification furnace and a coal-fired boiler. By acquiring the gas environment information in the first furnace and the gas environment information in the second furnace in real time, the comprehensive perception of the operating status of the two furnaces is realized. The environmental information of the two furnaces is accurately collected, which can reflect the output capacity of the gasifier and the combustion demand of the boiler, and provide reliable data support for subsequent flow regulation and temperature control. The coordination of the operation of the two furnaces is significantly improved, ensuring the dynamic balance and stability of the system. The present application designs a flow regulation method based on a gas delivery pipeline. By setting a gas temporary storage tank and a flow regulating valve, the flow rate of gas delivered from the gasifier to the boiler is adjusted to ensure that the pressure and temperature of the coal-fired boiler remain within a reasonable range. The design of the gas temporary storage tank reduces the instantaneous impact of the gasification furnace gas production fluctuation on the coal-fired boiler, and the flow regulating valve ensures the precise control of gas delivery. The operating stability of the coal-fired boiler is improved, avoiding combustion problems caused by insufficient or excessive gas supply, thereby providing a stable operating environment for coupled power generation. After completing the gas flow regulation operation, the present application further adjusts the first temperature information and the second temperature information based on the real-time updated environmental information of the first furnace and the second furnace. By adjusting the fuel input, gas flow or reaction temperature of the positive pressure biomass gasifier, and the coal input, combustion-aiding gas input or combustion air ratio of the coal-fired boiler, it is possible to dynamically adapt to environmental changes and optimize the operating parameters of the two furnaces. The adaptability of the system to external disturbances is improved, ensuring that both furnaces can maintain efficient operation under different working conditions. This application case comprehensively adjusts the flow and temperature parameters of the two furnaces so that the temperature of the first furnace is always kept in the optimal range suitable for the gasification reaction, and the temperature of the second furnace meets the efficient combustion requirements of the coal-fired boiler, thereby maximizing the energy efficiency of the entire coupling system. While ensuring efficient power generation, the combustion efficiency of the coal-fired boiler is optimized, and the use of coal and pollutant emissions are reduced. The power generation efficiency and environmental protection performance of the system are improved, and the deep coupling application of clean energy and traditional energy is promoted. This application provides a positive pressure biomass gasifier and coal-fired boiler coupling power generation method based on real-time environmental information, which significantly improves the coordination, dynamic adaptability and operational stability of the system. By optimizing the gas delivery flow and temperature regulation, the optimal operating state of the gasifier and the coal-fired boiler is guaranteed, and the efficient, stable and environmentally friendly operation of the coupled power generation system is achieved.

[0069] The present application proposes a method for coupling a positive pressure biomass gasification furnace with a coal-fired boiler for power generation. Other advantages, objectives and features of the present application will be partially reflected through the following description, and will also be partially understood by technicians in this field through research and practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present specification. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0071] Figure 1 A schematic diagram of a process for coupling a positive pressure biomass gasification furnace and a coal-fired boiler to generate electricity provided in an embodiment of the present application;

[0072] Figure 2 A schematic diagram of a process for regulating the flow of a gas delivery pipeline by a reinforcement learning algorithm provided in an embodiment of the present application;

[0073] Figure 3 A schematic diagram of a process for obtaining an optimization strategy network provided in an embodiment of the present application;

[0074] Figure 4 A schematic diagram of the specific steps of a positive pressure biomass gasification furnace and a coal-fired boiler coupled power generation method provided in an embodiment of the present application;

[0075] Figure 5 A schematic diagram of the structure of a target neural network model provided in an embodiment of the present application;

[0076] Figure 6 Another specific step flow diagram of a positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0077] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application 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 interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" and "having" and any of their variations 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. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0078] See also Figure 1, which is a schematic diagram of a process of coupling a positive pressure biomass gasification furnace and a coal-fired boiler to generate electricity provided in an embodiment of the present application, and may specifically include:

[0079] S110, obtaining first furnace environment information in the positive pressure biomass gasification furnace, wherein the first furnace environment information includes first pressure information, first temperature information and first gas composition information;

[0080] Exemplarily, the first pressure information represents the current pressure of the gas inside the gasifier in the positive pressure biomass gasifier. The first pressure information can be measured by a sensor in real time to obtain the gas pressure data inside the first furnace as one of the key environmental variables for subsequent control and adjustment.

[0081] The first temperature information indicates the current temperature in the positive pressure biomass gasification furnace. The gas temperature information in the first furnace collected by the temperature sensor can reflect the thermodynamic state in the gasification reaction process.

[0082] The first gas composition information indicates the composition ratio of the internal generated gas in the positive pressure biomass gasifier. The concentration and ratio of the generated gas (such as carbon monoxide, hydrogen and methane, etc.) in the gasifier can be measured by a gas analyzer to reflect the efficiency and quality of the gasification reaction.

[0083] S120, obtaining second furnace environment information in the coal-fired boiler, wherein the second furnace environment information includes second pressure information, second temperature information, and second gas composition information;

[0084] Exemplarily, the second pressure information indicates the current pressure of the gas inside the coal-fired boiler. The gas pressure inside the coal-fired boiler is measured in real time by a pressure sensor inside the boiler as an important parameter for evaluating the operating condition of the boiler.

[0085] The second temperature information indicates the current temperature of the gas inside the coal-fired boiler. The gas temperature information inside the boiler collected by the temperature sensor can reflect the state of the combustion reaction and the heat exchange efficiency.

[0086] The second gas composition information indicates the composition ratio of the gas inside the coal-fired boiler. The composition of the gas in the boiler (such as carbon dioxide, oxygen and water vapor) is analyzed by a gas analyzer to reflect the integrity of the combustion reaction and the fuel utilization efficiency.

[0087] S130, adjusting the flow rate of the gas delivery pipeline according to the gas environment information in the first furnace and the gas environment information in the second furnace, wherein the gas delivery pipeline is a gas delivery pipeline connecting the positive pressure biomass gasification furnace and the coal-fired boiler, and a gas temporary storage tank is also provided in the gas delivery pipeline;

[0088] Exemplarily, the gas delivery pipeline is a pipeline connecting the positive pressure biomass gasification furnace and the coal-fired boiler, and is used to deliver the synthesis gas generated by the gasification furnace to the coal-fired boiler as a combustion-supporting gas.

[0089] The gas temporary storage tank is installed in the gas transmission pipeline, which is used to temporarily store the gas transmitted from the gasifier, ensure the stability of gas flow and reduce the impact of transient fluctuations on the coal-fired boiler.

[0090] The gas environment information (pressure, temperature, gas composition) in the first furnace reflects the output capacity and operating status of the gasifier. The gas environment information (pressure, temperature, gas composition) in the second furnace reflects the demand for gas input by the coal-fired boiler. The gas flow rate in the gas delivery pipeline is controlled by the flow regulating valve to ensure that the pressure and temperature of the coal-fired boiler remain within a reasonable range, and to ensure the stability of the gas supply, to prevent the coal-fired boiler from operating unstably due to excessive or insufficient gas supply, thereby achieving a dynamic balance of gas delivery between the gasifier and the coal-fired boiler, and laying the foundation for the stable operation of coupled power generation.

[0091] S140. After the flow rate adjustment operation of the gas delivery pipeline is completed for a preset time, the first temperature information and the second temperature information are adjusted according to the first furnace gas environment information and the second furnace gas environment information.

[0092] Exemplarily, after the flow rate adjustment operation is completed, the temperature is further adjusted according to the environmental information of the two furnaces. After the flow rate adjustment is completed, it is necessary to wait for a preset time period (such as a few minutes or longer) to ensure that the effect of the gas delivery adjustment can be fully reflected in the environmental parameters of the two furnaces. By adjusting the operating parameters of the positive pressure biomass gasifier (such as fuel input, gas flow or reaction temperature), the temperature of the gasifier is adjusted to achieve the expected target. By adjusting the operating parameters of the coal-fired boiler (such as coal input, combustion-supporting gas input or combustion air ratio), the temperature of the boiler is adjusted. The adjustment basis includes the current second temperature information, the second pressure information and the second gas composition information to ensure the stability and thermal efficiency of the boiler combustion. Through temperature adjustment, it is ensured that the temperatures of the two furnaces reach the preset targets, the temperature of the first furnace is suitable for the optimal temperature range of the gasification reaction, and the temperature of the second furnace meets the efficient combustion requirements of the coal-fired boiler.

[0093] In summary, the present application provides a method for coupling power generation between a positive pressure biomass gasification furnace and a coal-fired boiler, and realizes comprehensive perception of the operating status of the two furnaces by obtaining the gas environment information in the first furnace and the gas environment information in the second furnace in real time. The environmental information of the two furnaces is accurately collected, which can reflect the output capacity of the gasifier and the combustion demand of the boiler, and provides reliable data support for subsequent flow regulation and temperature control. The coordination of the operation of the two furnaces is significantly improved, ensuring the dynamic balance and stability of the system. The present application designs a flow regulation method based on a gas delivery pipeline, and adjusts the flow rate of gas delivered from the gasifier to the boiler by setting a gas temporary storage tank and a flow regulating valve to ensure that the pressure and temperature of the coal-fired boiler remain within a reasonable range. The design of the gas temporary storage tank reduces the instantaneous impact of the gasification furnace gas production fluctuation on the coal-fired boiler, and the flow regulating valve ensures the precise control of gas delivery. The operating stability of the coal-fired boiler is improved, avoiding combustion problems caused by insufficient or excessive gas supply, thereby providing a stable operating environment for coupled power generation. After completing the gas flow regulation operation, the present application further adjusts the first temperature information and the second temperature information based on the real-time updated environmental information of the first furnace and the second furnace. By adjusting the fuel input, gas flow or reaction temperature of the positive pressure biomass gasifier, and the coal input, combustion-aiding gas input or combustion air ratio of the coal-fired boiler, it is possible to dynamically adapt to environmental changes and optimize the operating parameters of the two furnaces. The adaptability of the system to external disturbances is improved, ensuring that both furnaces can maintain efficient operation under different working conditions. This application case comprehensively adjusts the flow and temperature parameters of the two furnaces so that the temperature of the first furnace is always kept in the optimal range suitable for the gasification reaction, and the temperature of the second furnace meets the efficient combustion requirements of the coal-fired boiler, thereby maximizing the energy efficiency of the entire coupling system. While ensuring efficient power generation, the combustion efficiency of the coal-fired boiler is optimized, and the use of coal and pollutant emissions are reduced. The power generation efficiency and environmental protection performance of the system are improved, and the deep coupling application of clean energy and traditional energy is promoted. This application provides a positive pressure biomass gasifier and coal-fired boiler coupling power generation method based on real-time environmental information, which significantly improves the coordination, dynamic adaptability and operational stability of the system. By optimizing the gas delivery flow and temperature regulation, the optimal operating state of the gasifier and the coal-fired boiler is guaranteed, and the efficient, stable and environmentally friendly operation of the coupled power generation system is achieved.

[0094] In some examples, adjusting the flow rate of the gas delivery pipeline according to the first furnace gas environment information and the second furnace gas environment information includes:

[0095] The flow rate of the gas delivery pipeline is adjusted through a reinforcement learning algorithm according to the above-mentioned first furnace gas environment information and the above-mentioned second furnace gas environment information.

[0096] For example, the gas delivery pipeline flow is adjusted through the reinforcement learning algorithm to achieve efficient coupling between the positive pressure biomass gasifier and the coal-fired boiler, ensuring the dynamic balance and stability of the system operation. Pipeline flow adjustment based on the environmental information of the two furnaces can ensure that gas delivery meets the needs of the coal-fired boiler while maintaining the stable operation of the gasifier and dynamically adjusting the gas delivery flow to adapt to changes in environmental parameters between the two furnaces. Through the reinforcement learning algorithm, learn how to optimize the gas flow regulation strategy under different environmental conditions to maximize system efficiency and stability.

[0097] In some examples, Figure 2 A schematic diagram of a process for adjusting the flow of a gas delivery pipeline by a reinforcement learning algorithm provided in an embodiment of the present application, wherein the flow of the gas delivery pipeline includes a first gas flow corresponding to a positive pressure biomass gasifier and a second gas flow corresponding to a coal-fired boiler;

[0098] The method of adjusting the flow rate of the gas delivery pipeline by using a reinforcement learning algorithm according to the first furnace gas environment information and the second furnace gas environment information includes:

[0099] S1301: construct a state vector according to the first furnace gas environment information and the second furnace gas environment information :

[0100] ;

[0101] is the first pressure information, is the first temperature information, is the first gas composition information, is the second pressure information, is the second temperature information, is the above-mentioned second gas composition information;

[0102] S1302, according to the first gas flow and the second gas flow rate Constructing pipeline flow regulation actions :

[0103] ;

[0104] in, is the first gas flow rate, is the second gas flow rate;

[0105] S1303, define reward function according to control target :

[0106] ;

[0107] in, is the target value of the first pressure information or the target value of the second pressure information, is the target value of the first temperature information or the target value of the second temperature information, is the target value of the first gas composition information or the target value of the second gas composition information, Coupling_Error is the coupling error between furnaces, is the first weight information corresponding to the target value of the first pressure information or the target value of the second pressure information, is a second weight coefficient corresponding to the target value of the first temperature information or the target value of the second temperature information, is the weight coefficient corresponding to the target value of the first gas composition information or the target value of the second gas composition information, is the weight coefficient corresponding to the coupling error;

[0108] S1304, constructing a coupling model of the positive pressure biomass gasification furnace and the coal-fired boiler:

[0109] ;

[0110] in, is the value of the first temperature information at the next moment, is the value of the first pressure information at the next moment, is the value of the first gas composition information at the next moment, is the value of the second temperature information at the next moment, is the value of the second pressure information at the next moment, is the value of the second gas composition information at the next moment, is the current value of the first temperature information, is the current value of the first pressure information, is the current value of the first gas composition information, is the current value of the second temperature information, is the current value of the second pressure information, The current value of the second gas composition information, is the first gas flow rate, is the second gas flow rate, is random noise;

[0111] S1305: Perform reinforcement learning on the coupling model and historical experimental data through reinforcement learning algorithm to improve the reward function Corresponding cumulative rewards and obtaining the optimized strategy network;

[0112] S1306. Adjust the flow rate of the gas delivery pipeline according to the gas environment information in the first furnace and the gas environment information in the second furnace through the above-mentioned optimization strategy network.

[0113] For example, the state vector Including the temperature information of the first furnace , pressure information , Gas composition information , Temperature information of the second furnace , pressure information and gas composition information The state vector It constitutes the state description of the current environment, indicating the gas environment characteristics of the two furnaces at a certain moment, for the subsequent reinforcement learning model to evaluate the current environmental state.

[0114] Motion Vector Including first gas flow and the second gas flow , by adjusting the first gas flow and the second gas flow These two parameters can directly control the gas inflow, thereby affecting key indicators such as temperature, pressure and gas composition in the furnace.

[0115] Reward Function Designed as a weighted sum of multiple objective functions, including: Temperature target deviation: , pressure target deviation: , gas composition target deviation: , furnace coupling error: Coupling_Error. Each part has its corresponding weight , the size of the weight reflects its relative importance in the overall reward. By minimizing these deviations and coupling errors, reinforcement learning can learn how to optimize the gas flow to achieve the global goal of the system.

[0116] The coupling model predicts the next state To describe the dynamic relationship of the system, is the value of the first temperature information at the next moment, is the value of the first pressure information at the next moment, is the value of the first gas composition information at the next moment, is the value of the second temperature information at the next moment, is the value of the second pressure information at the next moment, is the value of the second gas composition information at the next moment, is the current value of the first temperature information, is the current value of the first pressure information, is the current value of the first gas composition information, is the current value of the second temperature information, is the current value of the second pressure information, The current value of the second gas composition information. These values ​​are based on the current state and actions , is the first gas flow rate, is the second gas flow rate mentioned above, and considering random noise The model is passed through the function The state transfer process is described, capturing the impact of gas flow rate adjustment on the furnace state.

[0117] The coupled model and historical experimental data are trained through reinforcement learning algorithms to improve the reward function Reinforcement learning optimizes the gas flow regulation strategy by interactively learning the policy network with the environment and optimizing the strategy based on historical data to achieve optimal control.

[0118] After the reinforcement learning training is completed, the optimization strategy network can be used in real time according to the current furnace gas environment information , dynamically adjust the gas flow , to achieve optimized control of gas transmission pipelines.

[0119] In summary, the reward function provided according to the embodiment of the present application integrates multiple target variables (temperature, pressure, composition, coupling error) to achieve all-round optimization of the gas environment in the furnace. The flow regulation strategy can reduce energy waste and improve combustion efficiency. The optimized control strategy can ensure that the system operates within a stable range, avoid drastic fluctuations in temperature and pressure, and reduce the risk of accidents. The minimization of coupling errors ensures the coordinated work between the two furnaces and avoids system imbalance. Automatic regulation of gas flow is achieved through reinforcement learning, reducing dependence on manual intervention. Precise control of flow regulation helps to improve energy utilization efficiency, reduce gas consumption and emissions, and reduce operating costs. This method can be applied to various coupled furnace systems (such as biomass gasifiers and coal-fired boilers), and is suitable for complex, multi-objective industrial control scenarios.

[0120] In some examples, the reinforcement algorithm described above includes a deep deterministic policy gradient algorithm.

[0121] In some examples, Figure 3 A schematic diagram of a process for obtaining an optimization strategy network provided in an embodiment of the present application, wherein the coupling model and historical experimental data are reinforced by a reinforcement learning algorithm to improve the reward function The corresponding cumulative rewards and the optimized strategy network are obtained, including:

[0122] S13051. Create a strategy network and a Q network according to the above coupling model;

[0123] S13052. Create an experience replay pool based on the above historical experimental data ;

[0124] S13053, according to the current status , generate actions through the above strategy network ;

[0125] S13054, Execution Action , record the reward function and the next state ;

[0126] S13055、 Stored in the above experience replay pool D;

[0127] S13056, Calculate target value :

[0128] ;

[0129] in, is the above reward function at time t, is the discount factor, For Q network in the next state and the action output by the policy network The predicted cumulative reward value under is the weight of the Q network;

[0130] S13057. Calculate the loss function of the Q network based on the above target value y , to update the Q network parameters :

[0131] ;

[0132] in, are the parameters of the Q network, Represents experience pool The expected value is calculated for the samples in is the target value With the Q network in its current state and the action output by the policy network The difference between the predicted cumulative rewards under , is the weight of the Q network;

[0133] S13058, calculate the gradient of the policy network to maximize the Q value and update the policy network parameters :

[0134] ;

[0135] in, is the gradient of the policy network, Represents a random sampling state from the experience replay pool D And the expected value of the gradient of these samples, Indicates the action on the Q network output value Find the partial derivative of, Represents the policy network output parameter The partial derivative of are the parameters of the policy network;

[0136] S13059, update the target Q network according to the preset update coefficient and target strategy network , to obtain the above optimization strategy network:

[0137] ;

[0138] in, It is the main Q network The delayed version of the target policy network Main Strategy Network A delayed version of are the parameters of the target Q network, The parameters of the target policy network, are the parameters of the main network, are the parameters of the main strategy network, is the soft update coefficient, Represents an update operation.

[0139] Exemplary, policy network and Q Network It is the core component of the deep deterministic policy gradient algorithm. The policy network is used to Output the best action The Q network is used to estimate the current state and actions The cumulative reward value of . Based on the coupling model, the policy network and Q network are initialized, and the parameters of the network will be continuously optimized through the subsequent training process.

[0140] Experience Replay Pool It is used to store data (state, action, reward, and next state) generated by the interaction between the system and the environment. The experience data in the experience replay pool provides samples for subsequent training, avoiding the time correlation problem of samples and improving the stability of training.

[0141] According to the current status , through the policy network Calculate Action , which is the flow adjustment value of the current pipeline. This action will affect the state of the system and will be used for subsequent optimization.

[0142] Perform actions in real environment After that, the system transfers to the next state At the same time, calculate the instant reward , used to measure the effect of the current action.

[0143] The current state, action, reward and next state Store in the experience replay pool to provide rich sample data for subsequent network training.

[0144] Target value The calculation formula is:

[0145] ;

[0146] It is an immediate reward that measures the direct benefit brought by the current action. The discount factor is what you need to weigh the relative importance of current rewards and future accumulated rewards. The network's cumulative reward prediction for the next state and next action.

[0147] The Q network minimizes the following loss function To update the parameters:

[0148] ;

[0149] is the target value, is the cumulative reward currently predicted by the Q network. Updated by backpropagation The parameters of make the cumulative reward predicted by the Q network closer to the target value.

[0150] The policy network updates parameters using the following gradient formula:

[0151] ;

[0152] for , reflecting the action Impact on cumulative rewards.

[0153] Policy network output pair parameters The partial derivative of describes the action in the current state Parameters sensitivity.

[0154] The target network is used to improve the stability of training and gradually approach the main network through a soft update mechanism:

[0155] ;

[0156] Based on the policy network and Q network in the deep deterministic policy gradient algorithm, the system can automatically learn the optimal flow regulation strategy in a dynamic environment without manual intervention. By optimizing gas flow regulation in real time, energy waste can be reduced, combustion efficiency can be improved, and operating costs can be reduced. The use of target networks and experience replay pools improves the stability of algorithm training and enhances the adaptability to environmental uncertainty and random noise. The reinforcement learning algorithm can dynamically adjust the control strategy according to changes in system states (such as temperature, pressure, gas composition, etc.) to adapt to complex and changing industrial environments. The reward function comprehensively considers the target values ​​of temperature, pressure, gas composition, and coupling errors, so that the optimization results can meet multiple industrial control objectives at the same time.

[0157] In some examples, Figure 4 A schematic diagram of the specific steps of a positive pressure biomass gasification furnace and a coal-fired boiler coupling power generation method provided in an embodiment of the present application, wherein the gas delivery pipeline flow rate includes the positive pressure biomass gasification furnace inlet flow rate and the coal-fired boiler outlet flow rate,

[0158] The above method further includes:

[0159] S210, when a first difference between the inlet flow rate of the positive pressure biomass gasification furnace and the outlet flow rate of the coal-fired boiler is greater than a preset difference, obtaining excess flow difference information between the first difference and the preset difference;

[0160] S220. Control the gas storage flow corresponding to the gas temporary storage tank based on the excess flow difference information.

[0161] For example, the gas delivery pipeline between the positive pressure biomass gasifier and the coal-fired boiler needs to maintain a balance of flow. If the difference between the inlet flow (biomass gasifier) ​​and the outlet flow (coal-fired boiler) is too large, it will cause abnormal operation of the system, for example: insufficient or excessive inlet flow of the gasifier. The pressure of the furnace outlet flow is unstable, affecting the combustion efficiency. In order to alleviate this unbalanced state, this embodiment introduces a gas temporary storage tank for dynamically balancing the flow.

[0162] Positive pressure biomass gasification furnace inlet flow and coal-fired boiler outlet flow The first difference calculation formula is:

[0163]

[0164] when Calculate the excess flow difference when:

[0165]

[0166] Excess flow difference Indicates the unbalanced flow that needs to be alleviated. When the inlet flow is excessive (i.e. ), part of the gas is introduced into the temporary storage tank to reduce the gas flow entering the gasifier:

[0167]

[0168] Setting up a temporary gas storage tank can help reduce the burden on the gasifier inlet. The temporary gas storage tank will automatically adjust the storage and release of gas according to the real-time flow difference information. If the outlet flow of the coal-fired boiler increases and exceeds the inlet flow at a certain moment in the future, the temporary storage tank can release gas to make up for the flow gap and ensure the dynamic balance of the flow. Through the dynamic adjustment of the temporary storage tank, the gasifier inlet flow and the boiler outlet flow are balanced, and the system instability caused by flow imbalance is reduced. The dynamic balance of inlet and outlet flow improves the operating efficiency of the gasifier and boiler and reduces the impact of flow fluctuations on system operation. Using a temporary storage tank as a buffer device can absorb flow fluctuations and avoid the impact of instantaneous large flow on equipment (such as gasifiers), thereby extending the service life of the equipment. Dynamically adjusting the excess flow helps to reduce energy losses caused by gas waste, thereby reducing the overall system energy consumption. This method can flexibly adjust the operation strategy of the temporary storage tank according to actual production needs, is applicable to a variety of industrial scenarios, and has strong versatility. By automatically controlling the gas storage flow of the temporary storage tank, intelligent management of the gas delivery system is realized, reducing manual intervention.

[0169] In some examples, adjusting the first temperature information and the second temperature information according to the first furnace gas environment information and the second furnace gas environment information includes:

[0170] The first temperature information and the second temperature information are predicted and controlled according to the first furnace gas environment information, the second furnace gas environment information and the target neural network model.

[0171] In some examples, Figure 5 A structural schematic diagram of a target neural network model provided in an embodiment of the present application, wherein the target neural network model includes an input layer, a feature extraction layer, an intermediate fusion layer, a dynamic time series layer, a fully connected layer, and an output layer;

[0172] The above-mentioned feature extraction layer includes a temperature feature branch, a pressure feature branch, a gas composition variable branch and a control variable branch;

[0173] The intermediate fusion layer is used to perform a vector fusion operation on the feature vectors extracted from the temperature feature branch, the pressure feature branch, the gas composition variable branch and the control variable branch to form an intermediate feature vector;

[0174] The above dynamic time series layer generates dynamic feature vectors based on historical time series based on long short-term memory network;

[0175] The fully connected layer includes a multi-layer fully connected network, and the fully connected layer is used to predict and generate the first temperature information and the second temperature information according to the intermediate feature vector and the dynamic feature vector;

[0176] The output layer is used to output the first temperature information and the second temperature information.

[0177] Exemplarily, this embodiment proposes a target neural network model, which predicts and adjusts the temperature information of the two furnaces by utilizing multiple gas environment feature inputs and combining a deep learning model (including feature extraction, time series processing, and prediction generation modules). Using the target neural network model as a controller, the temperature information of the two furnaces (i.e., the first temperature information and the second temperature information) is dynamically predicted and controlled by learning the features of historical and real-time data. The neural network realizes data processing, fusion, and time series modeling through multi-layer modules.

[0178] The target neural network model includes an input layer, a feature extraction layer, an intermediate fusion layer, a dynamic time series layer, a fully connected layer and an output layer. The input information of the input layer includes the first furnace gas environment information and the second furnace gas environment information.

[0179] The feature extraction layer includes four branches, a temperature feature branch, the pressure feature branch, the gas composition variable branch, and a control variable branch, each branch specializing in processing one type of feature.

[0180] The temperature feature branch is used to extract temperature-related features, such as the current temperature of each furnace. A convolutional neural network (CNN) or a fully connected network is used to extract its latent features.

[0181] The pressure feature branch is used to process the pressure information of each furnace and extract the influence of pressure on the system state change.

[0182] The gas composition variable branch is used to extract features related to gas composition (such as the CO content in the gas). Gas composition directly affects the reaction in the furnace and is of high importance.

[0183] The control variable branch is used to analyze the current control variables (such as gas flow, valve opening). These variables are used to describe the direct impact of operating behavior on temperature. Each branch extracts the corresponding feature vector through a dedicated network module (such as CNN or RNN).

[0184] The intermediate fusion layer fuses the feature vectors extracted from the temperature, pressure, gas composition, and control variable branches to form a unified intermediate feature vector. The fusion operation can be implemented through splicing, weighting, or attention mechanism. The fused feature vector retains the coupling relationship between multiple variables and is used to more comprehensively represent the state of the gas environment in the furnace.

[0185] The dynamic time series layer uses a long short-term memory network to generate dynamic feature vectors based on historical time series data. LSTM can capture the time correlation of the furnace state and is particularly suitable for processing systems with time dependence.

[0186] The fully connected layer performs deep nonlinear mapping based on the intermediate feature vectors (static features after fusion) and dynamic feature vectors (time series features) to generate prediction values. The fully connected layer contains multiple layers of fully connected networks (FCN) to further process the features.

[0187] The output layer generates the final predicted value based on the output of the fully connected layer: the first temperature information T1 and the second temperature information T2. ​​The output layer can use an activation function (such as ReLU or linear function) to ensure the rationality of the predicted value.

[0188] In summary, this embodiment can accurately predict the temperature in the furnace and improve the accuracy of the control system through the deep analysis of multidimensional features and time series by neural networks. The use of LSTM modules to capture time series features enhances the system's dynamic adaptability to changes in furnace temperature. Combining the four characteristics of temperature, pressure, gas composition and control variables helps to more comprehensively understand the state of the furnace and improve the reliability of prediction. Accurately predict and control the temperature in the furnace, optimize the combustion process, reduce energy waste, and improve system efficiency. Through efficient prediction and control, reduce the impact of temperature fluctuations on furnace operation and extend the service life of the equipment.

[0189] In some examples, the vector fusion operation is a feature weighted fusion operation, the long short-term memory network is a bidirectional long short-term memory network, and each fully connected network of the fully connected layer includes a Dropout module and a BatchNormalization module.

[0190] For example, feature weighted fusion is a fusion method that enhances the contribution of important features and weakens the influence of irrelevant features by assigning weights to different feature branches. Compared with simple splicing or averaging, weighted fusion can more effectively utilize the contribution of different features. Feature weights are adaptively learned based on training data to ensure that the most important feature branches receive higher weights in different tasks, thereby improving the model's predictive ability.

[0191] Bidirectional LSTM can capture both the forward and backward dependencies of a time series. For example, when predicting the temperature of a furnace, you can use the law that the current state depends on history, or you can combine future trends for prediction. Bidirectional LSTM consists of two LSTM modules: one that processes the forward information of the time series (from the past to the future), and one that processes the reverse information of the time series (from the future to the past). The output is the concatenation of dynamic feature vectors in two directions.

[0192] The fully connected layer enhances the stability and generalization ability of the network through the Dropout and Batch Normalization modules. The Dropout module is used to randomly discard a certain proportion of neurons (e.g. 50%) and their connections. This reduces the overfitting problem and improves the generalization ability of the model. During training, the output of some neurons is randomly set to zero; during the inference phase, all neurons are retained.

[0193] The Batch Normalization module is used to standardize the output of each layer of neurons so that its mean is 0 and its variance is 1. It accelerates network convergence and reduces the gradient vanishing problem. Dropout reduces overfitting, and Batch Normalization improves training stability. The combination of the two enables the fully connected layer to learn complex feature mappings more efficiently.

[0194] In some examples, Figure 6 Another specific step flow diagram of a positive pressure biomass gasification furnace and a coal-fired boiler coupled power generation method provided in an embodiment of the present application, the method further includes:

[0195] S310, obtaining temperature difference information between the first temperature information and the second temperature information;

[0196] S320. When the temperature difference information exceeds the preset temperature difference, the flow rate of the gas delivery pipeline is adjusted again according to the first furnace gas environment information and the second furnace gas environment information.

[0197] For example, combined with the previous temperature prediction method based on neural network, the temperature information of the two furnaces is obtained in real time to calculate the temperature difference. When the temperature difference exceeds the set threshold, it is considered that there may be a problem with the flow configuration of the current gas delivery pipeline, and the flow needs to be readjusted according to the gas environment information to ensure that the temperature of the two furnaces is within a reasonable range.

[0198] The temperature difference can reflect whether the operating status between the two furnaces is balanced. A large temperature difference may indicate that the flow distribution of the gas pipeline is unreasonable, resulting in overheating of one furnace or insufficient gas supply to the other furnace. When the temperature difference exceeds the threshold, the flow of the gas pipeline is readjusted according to the real-time gas environment information of the two furnaces.

[0199] Dynamically monitor and adjust the gas flow rate to keep the temperature difference between the gasifier and the boiler within a reasonable range to avoid overheating or overcooling. Ensure that both the gasifier and the boiler operate at the optimal temperature to improve combustion efficiency and energy utilization. Through real-time adjustment, reduce the operating risks caused by temperature fluctuations, protect the equipment and extend its service life. Accurately adjust the gas flow rate to avoid energy waste caused by excessive or insufficient flow.

[0200] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method, characterized in that: include: Acquire first furnace environment information in the positive pressure biomass gasification furnace, wherein the first furnace environment information includes first pressure information, first temperature information and first gas composition information; Acquire second furnace environment information in the coal-fired boiler, wherein the second furnace environment information includes second pressure information, second temperature information, and second gas composition information; According to the first furnace gas environment information and the second furnace gas environment information, the gas delivery pipeline flow is adjusted, wherein the gas delivery pipeline is a gas delivery pipeline connecting the positive pressure biomass gasification furnace and the coal-fired boiler, and a gas temporary storage tank is also provided in the gas delivery pipeline; After the flow rate adjustment operation of the gas delivery pipeline is completed for a preset time, adjusting the first temperature information and the second temperature information according to the first furnace gas environment information and the second furnace gas environment information; The adjusting the flow rate of the gas delivery pipeline according to the first furnace gas environment information and the second furnace gas environment information comprises: According to the first furnace gas environment information and the second furnace gas environment information, a reinforcement learning algorithm is used to adjust the flow rate of the gas delivery pipeline; The gas delivery pipeline flow rate includes a first gas flow rate corresponding to the positive pressure biomass gasification furnace and a second gas flow rate corresponding to the coal-fired boiler; The method of adjusting the flow rate of the gas delivery pipeline by using a reinforcement learning algorithm according to the first furnace gas environment information and the second furnace gas environment information includes: Construct a state vector according to the first furnace gas environment information and the second furnace gas environment information : ； is the first pressure information, is the first temperature information, is the first gas composition information, is the second pressure information, is the second temperature information, The second gas composition information; According to the first gas flow and the second gas flow rate Constructing pipeline flow regulation actions : ； in, is the first gas flow rate, is the second gas flow rate; Define the reward function based on the control objective : ； in, is the target value of the first pressure information or the target value of the second pressure information, is the target value of the first temperature information or the target value of the second temperature information, is the target value of the first gas composition information or the target value of the second gas composition information, Coupling_Error is the coupling error between furnaces, is the first weight information corresponding to the target value of the first pressure information or the target value of the second pressure information, is a second weight coefficient corresponding to the target value of the first temperature information or the target value of the second temperature information, is the weight coefficient corresponding to the target value of the first gas composition information or the target value of the second gas composition information, is the weight coefficient corresponding to the coupling error; Constructing a coupling model of the positive pressure biomass gasification furnace and the coal-fired boiler: ； in, is the value of the first temperature information at the next moment, is the value of the first pressure information at the next moment, is the value of the first gas composition information at the next moment, is the value of the second temperature information at the next moment, is the value of the second pressure information at the next moment, is the value of the second gas composition information at the next moment, is the current value of the first temperature information, is the current value of the first pressure information, is the current value of the first gas composition information, is the current value of the second temperature information, is the current value of the second pressure information, The current value of the second gas composition information, is the first gas flow rate, is the second gas flow rate, is random noise; The coupling model and historical experimental data are reinforced by a reinforcement learning algorithm to improve the reward function Corresponding cumulative rewards and obtaining the optimized strategy network; The flow rate of the gas delivery pipeline is adjusted through the optimization strategy network according to the first furnace gas environment information and the second furnace gas environment information.

2. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 1 is characterized in that: The reinforcement learning algorithm includes a deep deterministic policy gradient algorithm.

3. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 2 is characterized in that: The coupling model and historical experimental data are reinforced by a reinforcement learning algorithm to improve the reward function The corresponding cumulative rewards and the optimized strategy network are obtained, including: Creating a policy network and a Q network according to the coupling model; Create an experience replay pool based on the historical experimental data ; According to the current status , generating actions through the policy network ; Execute an action , record the reward function and the next state ; Will Stored in the experience replay pool D; Calculate target value : ； in, is the reward function at time t, is the discount factor, For Q network in the next state and the action output by the policy network The predicted cumulative reward value under is the weight of the Q network; Calculate the loss function of the Q network based on the target value y , to update the Q network parameters : ； in, are the parameters of the Q network, Represents experience pool The expected value is calculated for the samples in is the target value With the Q network in its current state and the action output by the policy network The difference between the predicted cumulative rewards under , is the weight of the Q network; Calculate the gradient of the policy network to maximize the Q value and update the policy network parameters : ； in, is the gradient of the policy network, Indicates a random sampling state from the experience replay pool D And the expected value of the gradient of these samples, Indicates the output value of the Q network action Find the partial derivative of, Represents the policy network output parameter The partial derivative of are the parameters of the policy network; Update the target Q network according to the preset update coefficient and target strategy network , to obtain the optimization strategy network: ； ； Among them, the target network It is the main Q network The delayed version of the target policy network Main Strategy Network A delayed version of are the parameters of the target Q network, The parameters of the target policy network, are the parameters of the main network, are the parameters of the main strategy network, is the soft update coefficient, Represents an update operation.

4. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 1 is characterized in that: The gas delivery pipeline flow rate includes the inlet flow rate of the positive pressure biomass gasification furnace and the outlet flow rate of the coal-fired boiler. The method further comprises: When a first difference between the inlet flow rate of the positive pressure biomass gasification furnace and the outlet flow rate of the coal-fired boiler is greater than a preset difference, obtaining excess flow difference information between the first difference and the preset difference; The gas storage flow corresponding to the gas temporary storage tank is controlled based on the excess flow difference information.

5. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 1 is characterized in that: The adjusting the first temperature information and the second temperature information according to the first furnace gas environment information and the second furnace gas environment information includes: The first temperature information and the second temperature information are predicted and controlled according to the first furnace gas environment information, the second furnace gas environment information and a target neural network model.

6. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 5 is characterized in that: The target neural network model includes an input layer, a feature extraction layer, an intermediate fusion layer, a dynamic time series layer, a fully connected layer and an output layer; The feature extraction layer includes a temperature feature branch, a pressure feature branch, a gas composition variable branch and a control variable branch; The intermediate fusion layer is used to perform a vector fusion operation on the feature vectors extracted from the temperature feature branch, the pressure feature branch, the gas composition variable branch and the control variable branch to form an intermediate feature vector; The dynamic time series layer generates dynamic feature vectors based on the long short-term memory network and the historical time series; The fully connected layer includes a multi-layer fully connected network, and the fully connected layer is used to predict and generate the first temperature information and the second temperature information according to the intermediate feature vector and the dynamic feature vector; The output layer is used to output the first temperature information and the second temperature information.

7. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 6 is characterized in that: The vector fusion operation is a feature weighted fusion operation, the long short-term memory network is a bidirectional long short-term memory network, and each fully connected network of the fully connected layer includes a Dropout module and a Batch Normalization module.

8. The positive pressure biomass gasification furnace and coal-fired boiler coupling power generation method according to claim 7 is characterized in that: The method further comprises: Acquire temperature difference information between the first temperature information and the second temperature information; When the temperature difference information exceeds the preset temperature difference, the flow rate of the gas delivery pipeline is adjusted again according to the first furnace gas environment information and the second furnace gas environment information.

Images