Prediction optimization control method and system for joint debugging of bed temperature and bed pressure of circulating fluidized bed boiler

Abstract

The invention discloses a predictive optimization control method and system for joint debugging of bed temperature and bed pressure of a circulating fluidized bed boiler, and relates to the field of industrial automatic control, and the method comprises the steps: screening out key operation parameters affecting the bed temperature and the bed pressure from a historical data set according to the operation condition and the actual operation demand of the boiler, determining a modeling feature set influencing the bed temperature and the bed pressure; based on the boiler bed temperature and bed pressure prediction model, obtaining a boiler bed temperature and bed pressure prediction value under each load; determining an optimization target and constraint conditions to obtain a boiler bed temperature and a bed pressure target value under each load; according to the boiler bed temperature and bed pressure prediction values and the corresponding target values under each load, the error and the error change rate of each prediction value and the corresponding target value are determined; therefore, the parameter adjustment amount of the PID controller can be determined. And further determining the control quantity output of the rotating speed of the slag cooler or the frequency of the primary fan. According to the invention, the response capability of the system to emergencies can be improved, and the flexibility and accuracy of regulation and control are enhanced.

Description

Technical Field

[0001] The present application relates to the field of industrial automatic control technology, and in particular to a predictive optimization control method and system for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler. Background Art

[0002] Boilers are important equipment for thermal energy production and are widely used in the fields of electricity, heating and industrial production. Bed temperature and bed pressure are key parameters affecting the operating performance of boilers, and have a direct impact on their efficiency, stability and environmental impact. Bed temperature is directly related to the combustion efficiency and emission control of fuel. Appropriate bed temperature can not only increase the combustion reaction rate and reduce fuel consumption, but also effectively reduce nitrogen oxides (NO x ) emissions, thereby reducing environmental pollution. At the same time, bed pressure affects the airflow distribution and boiler load regulation. Reasonable bed pressure can optimize the uniform combustion of fuel, improve heat transfer efficiency, and ensure equipment safety.

[0003] In the prior art, boiler control mainly relies on several common methods, including proportional-integral-differential control (PID) and model predictive control (MPC). PID control is widely used in various boiler systems due to its simple structure and easy implementation, especially in situations where stability and response speed requirements are relatively low. This method optimizes the control output by adjusting the proportional, integral and differential parameters to achieve the ideal bed temperature and bed pressure control effect. In contrast, model predictive control (MPC) uses the dynamic model of the system for prediction and optimization, can handle multivariable control problems, and is more adaptable. MPC predicts future system behavior and optimizes control based on these predictions, which is particularly suitable for complex boiler systems, where the regulation of bed temperature and bed pressure is affected by multiple factors. In recent years, the application of deep learning technology has brought new opportunities for boiler control. By building an efficient prediction model, the bed temperature and bed pressure of the boiler can be estimated more accurately. These advanced deep learning models not only improve the accuracy of predictions, but also can handle the complexity of nonlinear and dynamic changes.

[0004] However, there are the following disadvantages in the prior art:

[0005] 1. Existing control methods are difficult to cope with the nonlinear characteristics of boiler operation. In actual operation, the bed temperature and bed pressure of the boiler will be affected by many factors, such as fuel characteristics, load changes and environmental conditions. The nonlinear relationship between these factors makes it difficult for PID and MPC control methods to maintain a stable control effect, which easily leads to response lag or oscillation.

[0006] 2. Most existing control methods adopt a single-object control strategy, which cannot achieve coordinated regulation of bed temperature and bed pressure. Single-object control methods often ignore the mutual influence between bed temperature and bed pressure, and cannot achieve the best balance between the two. Adjusting the bed temperature alone may cause changes in bed pressure, while ignoring the adjustment of bed pressure may affect the stability of bed temperature. This limitation of lack of coordinated control makes it difficult for the boiler to maintain an efficient and stable operating state when facing complex changes in operating conditions.

[0007] 3. Existing control technology is highly dependent on models. The accuracy of model parameters directly affects the control effect. In actual applications, model parameters may fluctuate with time and changes in operating conditions, increasing the complexity of control.

[0008] 4. Existing methods use deep learning methods to predict boiler bed temperature and bed pressure, but these methods often only stay at a single prediction stage and do not apply the prediction results to actual optimization control. The lack of an overall comprehensive control system makes it impossible to adjust operating parameters according to real-time prediction data in actual operation. Summary of the invention

[0009] The purpose of this application is to provide a predictive optimization control method and system for the joint regulation of bed temperature and bed pressure of a circulating fluidized bed boiler, which can improve the system's response capability to emergencies, enhance the flexibility and accuracy of regulation, and reduce the complexity of control.

[0010] To achieve the above objectives, this application provides the following solutions:

[0011] In a first aspect, the present application provides a predictive optimization control method for coordinating bed temperature and bed pressure of a circulating fluidized bed boiler, the predictive optimization control method for coordinating bed temperature and bed pressure of a circulating fluidized bed boiler comprising:

[0012] Obtain historical data sets for boiler combustion systems.

[0013] According to the boiler operating conditions and actual operating requirements, key operating parameters affecting bed temperature and bed pressure are screened out from the historical data set; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load.

[0014] According to the key operating parameters, a modeling feature set affecting the bed temperature and the bed pressure is determined.

[0015] The modeling feature set is input into the boiler bed temperature and bed pressure prediction model to obtain the predicted values ​​of the boiler bed temperature and bed pressure under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on the LSTM network and the KAN network.

[0016] Based on the boiler bed temperature and bed pressure prediction model, the optimization objectives and constraints are determined; the optimization objectives include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints.

[0017] Taking the lowest nitrogen oxide emission and slag carbon content as the optimization objectives, and taking the primary fan constraint, slag cooler constraint and various related operating parameter constraints as the constraints, the target values ​​of boiler bed temperature and bed pressure under various loads are obtained.

[0018] According to the predicted values ​​of boiler bed temperature and bed pressure under various loads and the target values ​​of boiler bed temperature and bed pressure under various loads, the error and error change rate between each predicted value and the corresponding target value are determined.

[0019] According to the error and error change rate between each predicted value and the corresponding target value, a fuzzy PID control method is adopted to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: a proportional coefficient adjustment amount, an integral coefficient adjustment amount and a differential coefficient adjustment amount.

[0020] According to the parameter adjustment amount of the PID controller, the control amount output of the slag cooler speed or the primary fan frequency is determined.

[0021] In a second aspect, the present application provides a predictive optimization control system for the hot bed pressure joint regulation of a circulating fluidized bed boiler, the predictive optimization control system for the hot bed pressure joint regulation of a circulating fluidized bed boiler comprising:

[0022] The historical data set acquisition module is used to acquire the historical data set of the boiler combustion system.

[0023] The screening module is used to screen out key operating parameters affecting bed temperature and bed pressure from the historical data set according to the boiler operating conditions and actual operating requirements; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load.

[0024] The modeling feature set determination module is used to determine the modeling feature set that affects the bed temperature and bed pressure according to the key operating parameters.

[0025] The prediction value acquisition module is used to input the modeling feature set into the boiler bed temperature and bed pressure prediction model to obtain the boiler bed temperature and bed pressure prediction values ​​under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on the LSTM (Long Short-Term Memory) network and the KAN (Kolmogorov–Arnold Networks) network.

[0026] The module for determining optimization objectives and constraints is used to determine optimization objectives and constraints based on the boiler bed temperature and bed pressure prediction model; the optimization objectives include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints.

[0027] The target value acquisition module is used to obtain the target values ​​of boiler bed temperature and bed pressure under various loads with the lowest nitrogen oxide emissions and slag carbon content as the optimization targets, and with primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints.

[0028] The error and error change rate determination module is used to determine the error and error change rate of each predicted value and the corresponding target value based on the predicted values ​​of boiler bed temperature and bed pressure under various loads and the target values ​​of boiler bed temperature and bed pressure under various loads.

[0029] The parameter adjustment amount determination module is used to determine the parameter adjustment amount of the PID controller using a fuzzy PID control method according to the error and error change rate between each predicted value and the corresponding target value; the parameter adjustment amount of the PID controller includes: a proportional coefficient adjustment amount, an integral coefficient adjustment amount and a differential coefficient adjustment amount.

[0030] The control quantity output determination module is used to determine the control quantity output of the slag cooler speed or the primary fan frequency according to the parameter adjustment amount of the PID controller.

[0031] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0032] The present application provides a predictive optimization control method and system for the joint adjustment of bed temperature and bed pressure of a circulating fluidized bed boiler. First, a historical data set of a boiler combustion system is obtained; according to the boiler operating conditions and actual operating requirements, key operating parameters affecting bed temperature and bed pressure are screened out from the historical data set; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load; according to the key operating parameters, a modeling feature set affecting bed temperature and bed pressure is determined; the modeling feature set is input into a boiler bed temperature and bed pressure prediction model to obtain predicted values ​​of boiler bed temperature and bed pressure under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on an LSTM network and a KAN network; the LSTM+KAN network is used to establish a boiler bed temperature and bed pressure prediction model, which can achieve accurate estimation of parameter changes under complex conditions. Secondly, based on the boiler bed temperature and bed pressure prediction model, the optimization objectives and constraints are determined; the optimization objectives include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints; taking the minimum nitrogen oxide emissions and slag carbon content as the optimization objectives, and taking the primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints, the target values ​​of boiler bed temperature and bed pressure under various loads are obtained; combined with the boiler bed temperature and bed pressure prediction model, taking the minimum nitrogen oxide emissions and slag carbon content as the optimization objectives, the optimal boiler bed temperature and bed pressure target values ​​for each load can be solved. Then, according to the predicted values ​​of boiler bed temperature and bed pressure under each load and the target values ​​of boiler bed temperature and bed pressure under each load, determine the error and error change rate between each predicted value and the corresponding target value; according to the error and error change rate between each predicted value and the corresponding target value, adopt the fuzzy PID control method to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: proportional coefficient adjustment amount, integral coefficient adjustment amount and differential coefficient adjustment amount; according to the parameter adjustment amount of the PID controller, determine the control quantity output of the slag cooler speed or the primary fan frequency. The present application adopts a collaborative control strategy, based on the difference between the predicted value and the target value and its rate of change, and adopts a fuzzy PID control method to adjust the adjustment amount of the slag cooler and the primary air in real time, so as to realize the collaborative control of the bed temperature and bed pressure. The present application combines the comprehensive system design of prediction, optimization and control to realize the dynamic adjustment of the boiler bed temperature and bed pressure, and improve the stability and response speed of the boiler system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 This is an application environment diagram of a predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler in one embodiment of the present application.

[0035] Figure 2 A flow chart of a predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler provided in one embodiment of the present application.

[0036] Figure 3 A schematic diagram of a collaborative control strategy provided in one embodiment of the present application.

[0037] Figure 4 A schematic diagram of the structure of a boiler bed temperature and bed pressure prediction model provided in one embodiment of the present application.

[0038] Figure 5 A schematic diagram of the fuzzy PID reasoning process provided in one embodiment of the present application.

[0039] Figure 6 A schematic diagram of the functional modules of a predictive optimization control system for bed temperature and bed pressure co-regulation of a circulating fluidized bed boiler provided in one embodiment of the present application. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0041] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0042] Existing technologies mostly use PID and MPC methods to control boiler bed temperature and bed pressure. Traditional PID control methods have difficulty in dealing with the nonlinear characteristics of boiler operation, and are prone to response lag or oscillation. In addition, most existing technologies use a single control strategy, which is unable to achieve joint regulation of bed temperature and bed pressure, and ignores the mutual influence between the two. The MPC method is also highly dependent on the model. The accuracy of the model parameters directly affects the control effect, and these parameters may fluctuate with time and conditions in actual applications. In addition, the existing control methods are not flexible enough when responding to emergencies or disturbances, and it is difficult to quickly adjust the control strategy, which reduces the safety and economy of the boiler system.

[0043] In response to the above problems, the present application provides a predictive optimization control method and system for the joint adjustment of bed temperature and bed pressure of a circulating fluidized bed boiler. First, in view of the nonlinear characteristics and dynamic changes in boiler operation, a boiler bed temperature and bed pressure prediction model is established by adopting deep learning technology, combined with LSTM and KAN networks, to achieve accurate prediction of the boiler system state. Secondly, combined with the boiler bed temperature and bed pressure prediction model, with the lowest nitrogen oxide emissions and the lowest slag carbon content as the optimization goals, considering multiple equipment states and related parameters as constraints, a multi-objective optimization algorithm is used to solve the optimal bed temperature and bed pressure target values ​​under each load. Finally, a collaborative control strategy is designed, and a fuzzy PID control method is adopted. The difference between the predicted value and the target value and its rate of change are used as inputs. A fuzzy rule base is constructed by analyzing the historical operation data of the boiler, and the control adjustment amount of the slag cooler and the primary fan is dynamically adjusted based on different loads. The system's ability to respond to emergencies is improved, and the flexibility and accuracy of regulation are enhanced.

[0044] The predictive optimization control of the circulating fluidized bed boiler bed temperature and bed pressure joint regulation provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the acquired historical data set of the boiler combustion system to the server 104. After the server 104 receives the historical data set of the boiler combustion system, for the historical data set of the boiler combustion system, the server 104 screens out the key operating parameters that affect the bed temperature and bed pressure from the historical data set according to the boiler operating conditions and actual operating requirements; the key operating parameters include: bed temperature, bed pressure, coal feed, air volume and boiler load; according to the key operating parameters, determine the modeling feature set that affects the bed temperature and bed pressure; input the modeling feature set into the boiler bed temperature and bed pressure prediction model to obtain the predicted values ​​of the boiler bed temperature and bed pressure under each load; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on the LSTM network and the KAN network; based on the boiler bed temperature and bed pressure prediction model, determine the optimization objectives and constraints; the optimization objectives include: nitrogen oxide emissions and slag content The carbon content is the lowest; the constraint conditions include: primary fan constraint, slag cooler constraint and various related operating parameter constraints as constraint conditions; with the lowest nitrogen oxide emission and slag carbon content as the optimization target, with the primary fan constraint, slag cooler constraint and various related operating parameter constraints as constraint conditions, the target values ​​of boiler bed temperature and bed pressure under each load are obtained; according to the predicted values ​​of boiler bed temperature and bed pressure under each load and the target values ​​of boiler bed temperature and bed pressure under each load, the error and error change rate between each predicted value and the corresponding target value are determined; according to the error and error change rate between each predicted value and the corresponding target value, the fuzzy PID control method is used to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: proportional coefficient adjustment amount, integral coefficient adjustment amount and differential coefficient adjustment amount; according to the parameter adjustment amount of the PID controller, the control amount output of the slag cooler speed or the primary fan frequency is determined. The server 104 can feed back the obtained control amount output of the slag cooler speed or the primary fan frequency to the terminal 102. In addition, in some embodiments, the predictive optimization control method for the bed temperature and bed pressure joint regulation of the circulating fluidized bed boiler can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly perform predictive optimization control on the historical data set of the boiler combustion system, or the server 104 can obtain the historical data set of the boiler combustion system from the data storage system and perform predictive optimization control on the historical data set of the boiler combustion system.

[0045] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, and tablet computers. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0046] In an exemplary embodiment, Figure 2 and Figure 3 As shown, a predictive optimization control method for the bed temperature and bed pressure joint adjustment of a circulating fluidized bed boiler is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the method, which includes the following steps S1 to S9. Among them:

[0047] S1: Obtain the historical data set of the boiler combustion system.

[0048] S2: According to the boiler operating conditions and actual operating requirements, the key operating parameters affecting the bed temperature and bed pressure are screened out from the historical data set; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load.

[0049] S3: Determine a modeling feature set that affects bed temperature and bed pressure based on the key operating parameters.

[0050] S4: Input the modeling feature set into the boiler bed temperature and bed pressure prediction model to obtain the boiler bed temperature and bed pressure prediction values ​​under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on LSTM network and KAN network.

[0051] S5: Based on the boiler bed temperature and bed pressure prediction model, determine the optimization objectives and constraints; the optimization objectives include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints.

[0052] S6: Taking the lowest nitrogen oxide emission and slag carbon content as the optimization objectives, and taking the primary fan constraint, slag cooler constraint and various related operating parameter constraints as constraints, the target values ​​of boiler bed temperature and bed pressure under various loads are obtained.

[0053] S7: According to the predicted values ​​of the boiler bed temperature and bed pressure under each load and the target values ​​of the boiler bed temperature and bed pressure under each load, the error and error change rate of each predicted value and the corresponding target value are determined.

[0054] S8: According to the error and error change rate between each predicted value and the corresponding target value, a fuzzy PID control method is used to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: a proportional coefficient adjustment amount, an integral coefficient adjustment amount and a differential coefficient adjustment amount.

[0055] S9: Determine the control output of the slag cooler speed or the primary fan frequency according to the parameter adjustment amount of the PID controller.

[0056] By implementing the above steps S1 to S9, the present application has the following beneficial effects:

[0057] 1. Deep learning prediction model: The LSTM+KAN network is used to establish a boiler bed temperature and bed pressure prediction model, which can achieve accurate estimation of parameter changes under complex working conditions.

[0058] 2. Energy efficiency-carbon economy multi-objective optimization: Combined with the bed temperature and bed pressure prediction model, a multi-objective optimization algorithm is adopted to minimize nitrogen oxide emissions and slag carbon content, and the optimal boiler bed temperature and bed pressure target values ​​for each load are solved.

[0059] 3. Collaborative control strategy: Based on the difference between the predicted value and the target value and their rate of change, the fuzzy PID control method is used to adjust the adjustment amount of the slag cooler and primary air in real time to achieve collaborative control of bed temperature and bed pressure.

[0060] 4. Integrated control system: The integrated system design combining prediction, optimization and control can realize dynamic adjustment of boiler bed temperature and bed pressure, and improve the stability and response speed of the boiler system.

[0061] As an optional implementation, in step S3, specifically including:

[0062] S31: Detect missing values ​​and abnormal values ​​according to the key operating parameters.

[0063] S32: Using an interpolation method or a mean method to process the missing values ​​and abnormal values ​​to obtain processed operating parameters.

[0064] S33: performing correlation analysis on the processed operating parameters, and taking the parameters with correlation greater than a correlation threshold and having a control association as modeling features that affect the bed temperature and bed pressure.

[0065] S34: The modeling features are subjected to feature generation to obtain time features and coupling features, which can improve the accuracy of the prediction model; the time features include: decomposing the modeling feature timestamp into corresponding time, and using sine and cosine functions to convert periodic features and time difference features of bed temperature and bed pressure; the coupling features are based on control experience and mechanism analysis to identify the mutual influence mechanism between bed temperature and bed pressure, and generate new features through mathematical operations.

[0066] 1) Time feature: decompose the modeling feature timestamp into year, month, day, hour, etc., and use sine and cosine functions to convert periodic features. At the same time, the time difference of bed temperature and bed pressure is calculated as the modeling feature by performing a time difference generated by a sliding window on the bed temperature and bed pressure.

[0067] 2) Coupling characteristics, based on control experience and mechanism analysis, identify the mutual influence mechanism between bed temperature and bed pressure, and generate new features through mathematical operations. For example, temperature-pressure ratio (bed temperature / bed pressure): by calculating the ratio of bed temperature to bed pressure, the stability and efficiency of the system under different working conditions are revealed. Temperature-pressure difference (bed temperature-bed pressure): calculating the difference between the current bed temperature and bed pressure can help capture the dynamic changes of the system. When the temperature-pressure difference changes significantly, it may indicate a change in equipment status or a risk of failure. Rate of change: Introducing the rate of change of bed temperature and bed pressure can more sensitively capture dynamic changes in a short period of time.

[0068] S35: Determine a modeling feature set that affects bed temperature and bed pressure according to the modeling features, time features, and coupling features.

[0069] Interaction features: Consider the product of bed temperature and bed pressure, capture the nonlinear relationship between the two, and model complex system behaviors. Integrate all acquired and generated features into a feature set, including modeling features, time features, coupling features, etc., and perform normalization to form the feature matrix that is finally input into the model.

[0070] In another exemplary embodiment of the present application, a composite model of LSTM and KAN is constructed based on the above-mentioned modeling feature set for accurate prediction of boiler bed temperature and bed pressure. The model captures the long-term and short-term dependencies in time series data through the LSTM network, and the KAN network weights the key features.

[0071] (1) Network structure design, using a composite model structure of LSTM and KAN in parallel. The LSTM network adopts a two-layer structure and introduces residual connections between layers. The KAN network contains two modules. Through special residual connections, the output of the first KAN is concatenated with the output of the first layer LSTM as the input of the second layer LTSM. The output features of the final LSTM and KAN are concatenated and processed by the linear layer to output the final bed temperature and bed pressure prediction values. The overall network diagram is shown in the figure below. Figure 4 shown.

[0072] (2) LSTM. In order to improve the model performance, a two-layer LSTM network is designed, and residual connections are introduced between layers. The specific design is as follows:

[0073] Number of layers: The LSTM network adopts a two-layer structure to better extract features.

[0074] Residual connection: Residual connections are introduced between each layer to solve the gradient vanishing problem in deep networks, thereby improving the training efficiency and accuracy of the model.

[0075] (3) KAN: The adopted KAN model innovatively transforms the paradigm of “fixed activation function + learnable weight” in neural networks into the form of “learnable activation function + function summation”.

[0076]

[0077] Among them, x n For n-dimensional input vector, the internal function φ q,p The domain is usually [0, 1], and its range is the set of real numbers. External function Φ q The domain and range are

[0078] In another exemplary embodiment of the present application, based on the above-mentioned boiler bed temperature and bed pressure prediction model, with the lowest nitrogen oxide emissions and slag carbon content as optimization goals, the equipment status of the slag cooler and primary fan and the boundaries of related operating parameters are considered as constraints to solve the optimal boiler bed temperature and bed pressure target values ​​under various loads.

[0079] (1) Optimization target, set the minimum nitrogen oxide emission f1 = NO X (T,P), minimize the carbon content of slag f2 = C ash (T,P), establish the energy efficiency-carbon economy optimization target min y=F(x)=(min f1,min f2), where: T is the predicted bed temperature and P is the predicted bed pressure.

[0080] (2) Constraints: In order to ensure the safety and economy of the boiler during optimized operation, the following constraints are set, including but not limited to the boundaries of equipment status and related operating parameters.

[0081] The expression of the primary fan constraint is:

[0082] state i ∈{0,1} (2);

[0083]

[0084] Among them, state i is the start-up state of the i-th primary fan, is the frequency of the i-th primary fan, is the minimum frequency of the i-th primary fan, is the maximum frequency of the i-th primary fan, is the maximum rate of change of the primary fan frequency, and N is the number of primary fans.

[0085] The expression of the slag cooler constraint is:

[0086] state j ∈{0,1} (5);

[0087]

[0088] Among them, state j is the opening state of the jth slag cooler, is the speed of the jth slag cooler, is the minimum speed of the jth slag cooler, is the maximum speed of the jth slag cooler, and M is the number of slag coolers.

[0089] The expressions of the constraints of the relevant operating parameters are:

[0090] g1=P min ≤P o ≤P max (7);

[0091] g2=T min ≤T0≤T max (8);

[0092] Among them, P o is the reference bed pressure, P min is the minimum reference bed pressure, P max is the maximum value of the reference bed pressure, g1 is the constraint of the reference bed pressure, T o is the reference bed temperature, T min is the minimum reference bed temperature, T max is the maximum value of the reference bed temperature, and g2 is the constraint of the reference bed temperature.

[0093] It should be noted that various relevant operating parameter constraints include, but are not limited to, temperature and pressure constraints. For example, the operating range of bed temperature and bed pressure is determined by equipment design parameters.

[0094] (3) Mapping relationship: Through the boiler bed temperature and bed pressure prediction model established above, a mapping relationship between the bed temperature and bed pressure and the two optimization objectives is established.

[0095] NOx emissions model:

[0096]

[0097] Among them, NO Xis the nitrogen oxide emission, K1 is the first coefficient obtained by fitting the historical data, K2 is the second coefficient obtained by fitting the historical data, T is the predicted bed temperature, P is the predicted bed pressure, T o is the reference bed temperature, P o is the reference bed pressure, a is the first empirical coefficient, b is the second empirical coefficient, c is the third empirical coefficient, and e is the fifth empirical coefficient.

[0098] Slag carbon content model:

[0099]

[0100] Among them, C ash is the carbon content of the slag, K3 is the third coefficient obtained by fitting the historical data, K4 is the fourth coefficient obtained by fitting the historical data, T is the predicted bed temperature, P is the predicted bed pressure, T o is the reference bed temperature, P o is the reference bed pressure, d is the fourth empirical coefficient, e is the fifth empirical coefficient, and φ is the phase offset.

[0101] (4)Multi-objective optimization algorithm.

[0102] The multi-objective optimization algorithm NSGA-II is used to optimize the objective function under given constraints. The optimization process includes the following steps:

[0103] 1. Initialize the population: Randomly generate the boiler bed temperature, bed pressure, and the opening status and frequency / speed of each primary fan and slag cooler under various load conditions.

[0104] 2. Evaluation goal: Calculate the sum of each individual according to the mapping formula.

[0105] 3. Selection and crossover: Select individuals with high fitness for crossover to generate new individuals.

[0106] 4. Mutation: Make small changes to the bed temperature, bed pressure and equipment status of new individuals to increase diversity.

[0107] 5. Iterative update: Repeat the evaluation, selection, crossover and mutation process until the convergence conditions are met.

[0108] Through the above steps, the optimal boiler bed temperature and bed pressure target values ​​under various loads are obtained.

[0109] In another exemplary embodiment of the present application, in step S8, it specifically includes:

[0110] S81: According to the fuzzy membership function, the error and the error change rate are mapped into a fuzzy set.

[0111] S82: According to the fuzzy set and the preset fuzzy control rule base, using Mamdani fuzzy reasoning or Sugeno fuzzy reasoning, the increment of the PID controller parameters is calculated; the increment of the PID controller parameters includes: proportional coefficient increment, integral coefficient increment and differential coefficient increment.

[0112] S83: Defuzzify the increment of the PID controller parameter using the center of gravity method or the maximum membership method to obtain the parameter adjustment amount of the PID controller.

[0113] In order to achieve precise control of the slag cooler and primary fan, this application adopts a fuzzy PID controller, taking the difference between the predicted values ​​of boiler bed temperature and bed pressure under various loads and the target values ​​of boiler bed temperature and bed pressure under various loads and their change rates as inputs, and dynamically adjusting the three parameters K of the PID controller. p (proportional coefficient), K i (integral coefficient) and K d (differential coefficient).

[0114] Using fuzzy control method, the error and error change rate are quantified into several fuzzy sets, such as negative large (NB), negative small (NS), zero (ZE), positive small (PS), positive large (PB). Control rules are established based on these fuzzy sets.

[0115] According to the control experience, fuzzy control rules are designed. The control target is to adjust K p , K i , and K d Make the system output close to the set target value. For example, when the error is large, K should be increased. p to quickly correct the deviation; when the error is close to zero and changes slowly, K should be reduced d To avoid over-regulation of the system; when the error persists, it is necessary to increase K i to eliminate steady-state errors.

[0116] In the control process, the increment ΔK of the PID coefficient is dynamically adjusted according to the fuzzy control rule output. p , ΔK i and ΔK d The adjusted PID coefficient is:

[0117]

[0118] Among them, K p (t) is the adjusted proportional coefficient, K i (t) is the adjusted integral coefficient, K d (t) is the adjusted differential coefficient, is the initial proportionality factor, is the initial integration coefficient, is the initial differential coefficient, ΔK p is the increment of the proportional coefficient, ΔK i is the increment of the integral coefficient, ΔK d is the increment of the differential coefficient, e(t) is the error between the predicted value and the corresponding target value, and Δe(t) is the rate of change of the error between the predicted value and the corresponding target value.

[0119] PID control formula, the control output of the slag cooler speed or primary fan frequency is realized by the following formula:

[0120]

[0121] Among them, u(t) is the control output of the slag cooler speed or primary fan frequency, τ is the integral variable, and e(τ) is the integrand.

[0122] like Figure 5 As shown in the figure, the fuzzy reasoning process mainly includes three steps: fuzzification, fuzzy rules and defuzzification:

[0123] 1. Fuzzification: The error between the predicted value and the corresponding target value and the error change rate between the predicted value and the corresponding target value are mapped into fuzzy sets through fuzzy membership functions.

[0124] 2. Fuzzy rules: According to the preset fuzzy control rule base, use Mamdani fuzzy reasoning or Sugeno fuzzy reasoning to calculate the increment ΔK of PID parameters p , ΔK i = and ΔK d .

[0125] 3. Defuzzification: Use the centroid method or maximum membership method to defuzzify the inference results and obtain the final PID parameter adjustment amount.

[0126] Update the rotation speed of the slag cooler and the frequency of the primary fan according to the adjustment amount to ensure stable control of the boiler bed temperature and bed pressure.

[0127] The present application also provides an application scenario, which applies the above-mentioned prediction optimization control method for the bed temperature and bed pressure joint regulation of the circulating fluidized bed boiler. Specifically: The prediction optimization control method for the bed temperature and bed pressure joint regulation of the circulating fluidized bed boiler provided in this embodiment can be applied in the boiler bed temperature and bed pressure joint regulation control scenario. The boiler bed temperature and bed pressure joint regulation control scenario includes: feature processing link, prediction model establishment link, energy efficiency-carbon economy multi-objective optimization link and collaborative control strategy link. Obtain a historical data set of the boiler combustion system; select key operating parameters that affect bed temperature and bed pressure from the historical data set according to the boiler operating conditions and actual operating requirements; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load; determine a modeling feature set that affects bed temperature and bed pressure according to the key operating parameters; input the modeling feature set into the boiler bed temperature and bed pressure prediction model to obtain the predicted values ​​of boiler bed temperature and bed pressure under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on LSTM network and KAN network; determine optimization objectives and constraints based on the boiler bed temperature and bed pressure prediction model; the optimization objectives include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operation constraints. The operation parameter constraint is taken as the constraint condition; the nitrogen oxide emission and the minimum carbon content of the slag are taken as the optimization objectives, and the primary fan constraint, the slag cooler constraint and the various related operation parameter constraints are taken as the constraint conditions to obtain the target values ​​of the boiler bed temperature and bed pressure under various loads; according to the predicted values ​​of the boiler bed temperature and bed pressure under various loads and the target values ​​of the boiler bed temperature and bed pressure under various loads, the error and the error change rate between each predicted value and the corresponding target value are determined; according to the error and the error change rate between each predicted value and the corresponding target value, the fuzzy PID control method is used to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: proportional coefficient adjustment amount, integral coefficient adjustment amount and differential coefficient adjustment amount; according to the parameter adjustment amount of the PID controller, the control amount output of the slag cooler speed or the primary fan frequency can be determined.

[0128] Based on the same inventive concept, the embodiment of the present application also provides a circulating fluidized bed boiler bed temperature bed pressure joint adjustment predictive optimization control system for realizing the above-mentioned circulating fluidized bed boiler bed temperature bed pressure joint adjustment predictive optimization control method. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the embodiments of the predictive optimization control system for the bed temperature bed of the circulating fluidized bed boiler provided below can be referred to the limitations of the predictive optimization control method for the bed temperature bed of the circulating fluidized bed boiler bed pressure joint adjustment above, and will not be repeated here.

[0129] In an exemplary embodiment, Figure 6As shown, a prediction optimization control system for the hot bed pressure joint adjustment of a circulating fluidized bed boiler is provided, and the prediction optimization control system for the hot bed pressure joint adjustment of a circulating fluidized bed boiler comprises:

[0130] The historical data set acquisition module M1 is used to acquire the historical data set of the boiler combustion system.

[0131] The screening module M2 is used to screen out key operating parameters affecting bed temperature and bed pressure from the historical data set according to the boiler operating conditions and actual operating requirements; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load.

[0132] The modeling feature set determination module M3 is used to determine the modeling feature set that affects the bed temperature and bed pressure according to the key operating parameters.

[0133] The prediction value acquisition module M4 is used to input the modeling feature set into the boiler bed temperature and bed pressure prediction model to obtain the boiler bed temperature and bed pressure prediction values ​​under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on the LSTM network and the KAN network.

[0134] The optimization target and constraint condition determination module M5 is used to determine the optimization target and constraint conditions based on the boiler bed temperature and bed pressure prediction model; the optimization targets include: minimum nitrogen oxide emissions and slag carbon content; the constraints include: primary fan constraints, slag cooler constraints and various related operating parameter constraints as constraints.

[0135] The target value acquisition module M6 is used to obtain the target values ​​of boiler bed temperature and bed pressure under various loads with the lowest nitrogen oxide emission and slag carbon content as the optimization targets, and with the primary fan constraint, slag cooler constraint and various related operating parameter constraints as the constraints.

[0136] The error and error change rate determination module M7 is used to determine the error and error change rate of each predicted value and the corresponding target value based on the predicted values ​​of boiler bed temperature and bed pressure under various loads and the target values ​​of boiler bed temperature and bed pressure under various loads.

[0137] The parameter adjustment determination module M8 is used to determine the parameter adjustment of the PID controller using a fuzzy PID control method based on the error and error change rate between each predicted value and the corresponding target value; the parameter adjustment of the PID controller includes: a proportional coefficient adjustment, an integral coefficient adjustment and a differential coefficient adjustment.

[0138] The control quantity output determination module M9 is used to determine the control quantity output of the slag cooler speed or the primary fan frequency according to the parameter adjustment amount of the PID controller.

[0139] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0140] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A predictive optimization control method for the bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler, characterized in that: The predictive optimization control method for the combined regulation of bed temperature and bed pressure of the circulating fluidized bed boiler comprises: Obtain historical data sets for boiler combustion systems; According to the boiler operating conditions and actual operating requirements, the key operating parameters affecting the bed temperature and bed pressure are screened out from the historical data set; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load; Determining a modeling feature set that affects bed temperature and bed pressure based on the key operating parameters; The modeling feature set is input into a boiler bed temperature and bed pressure prediction model to obtain the predicted values ​​of the boiler bed temperature and bed pressure under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on an LSTM network and a KAN network; Based on the boiler bed temperature and bed pressure prediction model, the optimization target and constraint conditions are determined; the optimization target includes: the lowest nitrogen oxide emission and slag carbon content; the constraint conditions include: primary fan constraint, slag cooler constraint and various related operating parameter constraints as constraint conditions; Taking the lowest nitrogen oxide emission and slag carbon content as the optimization target, and taking the primary fan constraint, slag cooler constraint and various related operating parameter constraints as the constraint conditions, the target values ​​of boiler bed temperature and bed pressure under various loads are obtained; According to the predicted values ​​of boiler bed temperature and bed pressure under various loads and the target values ​​of boiler bed temperature and bed pressure under various loads, the error and error change rate of each predicted value and the corresponding target value are determined; According to the error and error change rate between each predicted value and the corresponding target value, a fuzzy PID control method is used to determine the parameter adjustment amount of the PID controller; the parameter adjustment amount of the PID controller includes: a proportional coefficient adjustment amount, an integral coefficient adjustment amount and a differential coefficient adjustment amount; According to the parameter adjustment amount of the PID controller, the control amount output of the slag cooler speed or the primary fan frequency is determined.

2. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: According to the key operating parameters, a modeling feature set affecting bed temperature and bed pressure is determined, specifically including: Detect missing values ​​and outliers based on the key operating parameters; The missing values ​​and abnormal values ​​are processed by using an interpolation method or a mean method to obtain processed operating parameters; Performing correlation analysis on the processed operating parameters, and taking parameters with correlation greater than a correlation threshold and having a control association as modeling features affecting bed temperature and bed pressure; The modeling feature is subjected to feature generation to obtain a time feature and a coupling feature; the time feature includes: decomposing the modeling feature timestamp into corresponding time, and using sine and cosine functions to convert the periodic feature and the time difference feature of the bed temperature and bed pressure; the coupling feature is based on the control experience and mechanism analysis, identifies the mutual influence mechanism between the bed temperature and bed pressure, and generates a new feature through mathematical operation; A modeling feature set affecting the bed temperature and bed pressure is determined according to the modeling features, time features and coupling features.

3. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: The expression of the primary fan constraint is: state i ∈{0,1}; Among them, state i is the start-up state of the i-th primary fan, is the frequency of the i-th primary fan, is the minimum frequency of the i-th primary fan, is the maximum frequency of the i-th primary fan, is the maximum rate of change of the primary fan frequency, and N is the number of primary fans.

4. The predictive optimization control method for the bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: The expression of the slag cooler constraint is: state j ∈{0,1}; Among them, state j is the opening state of the jth slag cooler, is the speed of the jth slag cooler, is the minimum speed of the jth slag cooler, is the maximum speed of the jth slag cooler, and M is the number of slag coolers.

5. The predictive optimization control method for the combined adjustment of bed temperature and bed pressure of a circulating fluidized bed boiler according to claim 1, characterized in that: The expressions of the constraints of the relevant operating parameters are: g1=P min ≤P o ≤P max ; g2=T min ≤T o ≤T max ; Among them, P o is the reference bed pressure, P min is the minimum reference bed pressure, P max is the maximum value of the reference bed pressure, g1 is the constraint of the reference bed pressure, T o is the reference bed temperature, T min is the minimum reference bed temperature, T max is the maximum value of the reference bed temperature, and g2 is the constraint of the reference bed temperature.

6. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: The calculation formula for the nitrogen oxide emissions is: Among them, NO X is the nitrogen oxide emission, K1 is the first coefficient obtained by fitting the historical data, K2 is the second coefficient obtained by fitting the historical data, T is the predicted bed temperature, P is the predicted bed pressure, T o is the reference bed temperature, P o is the reference bed pressure, a is the first empirical coefficient, b is the second empirical coefficient, c is the third empirical coefficient, and e is the fifth empirical coefficient.

7. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: The calculation formula of the carbon content of the slag is: Among them, C ash is the carbon content of the slag, K3 is the third coefficient obtained by fitting the historical data, K4 is the fourth coefficient obtained by fitting the historical data, T is the predicted bed temperature, P is the predicted bed pressure, T o is the reference bed temperature, P o is the reference bed pressure, d is the fourth empirical coefficient, e is the fifth empirical coefficient, and φ is the phase shift.

8. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: According to the error and error change rate between each predicted value and the corresponding target value, the fuzzy PID control method is used to determine the parameter adjustment amount of the PID controller, including: According to the fuzzy membership function, the error and error change rate are mapped into fuzzy sets; According to the fuzzy set and the preset fuzzy control rule base, the increment of the PID controller parameters is calculated using Mamdani fuzzy reasoning or Sugeno fuzzy reasoning; the increment of the PID controller parameters includes: a proportional coefficient increment, an integral coefficient increment and a differential coefficient increment; The increment of the PID controller parameter is defuzzified using the center of gravity method or the maximum membership method to obtain the parameter adjustment amount of the PID controller.

9. The predictive optimization control method for bed temperature and bed pressure joint regulation of a circulating fluidized bed boiler according to claim 1, characterized in that: The expression of the control quantity output of the slag cooler speed or primary fan frequency is: Where u(t) is the control output of the slag cooler speed or primary fan frequency, K p (t) is the adjusted proportional coefficient, K i (t) is the adjusted integral coefficient, K d (t) is the adjusted differential coefficient, is the initial proportionality factor, is the initial integration coefficient, is the initial differential coefficient, ΔK p is the increment of the proportional coefficient, ΔK i is the increment of the integral coefficient, ΔK d is the increment of the differential coefficient, e(t) is the error between the predicted value and the corresponding target value, Δe(t) is the rate of change of the error between the predicted value and the corresponding target value, τ is the integral variable, and e(τ) is the integrand.

10. A predictive optimization control system for the combined regulation of bed temperature and bed pressure of a circulating fluidized bed boiler, characterized in that: The predictive optimization control system for the circulating fluidized bed boiler bed hot bed pressure joint regulation includes: A historical data set acquisition module is used to acquire historical data sets of a boiler combustion system; A screening module is used to screen out key operating parameters that affect bed temperature and bed pressure from the historical data set according to the boiler operating conditions and actual operating requirements; the key operating parameters include: bed temperature, bed pressure, coal feed rate, air volume and boiler load; A modeling feature set determination module, used to determine a modeling feature set affecting bed temperature and bed pressure according to the key operating parameters; A prediction value acquisition module, used for inputting the modeling feature set into a boiler bed temperature and bed pressure prediction model to obtain the boiler bed temperature and bed pressure prediction values ​​under various loads; the boiler bed temperature and bed pressure prediction model is a composite model constructed based on an LSTM network and a KAN network; An optimization target and constraint condition determination module is used to determine the optimization target and constraint condition based on the boiler bed temperature and bed pressure prediction model; the optimization target includes: the lowest nitrogen oxide emission and slag carbon content; the constraint condition includes: primary fan constraint, slag cooler constraint and various related operating parameter constraints as constraint conditions; The target value acquisition module is used to obtain the target values ​​of boiler bed temperature and bed pressure under various loads with the lowest nitrogen oxide emission and slag carbon content as the optimization targets, and with the primary fan constraint, slag cooler constraint and various related operating parameter constraints as the constraint conditions; An error and error change rate determination module is used to determine the error and error change rate of each predicted value and the corresponding target value based on the predicted values ​​of boiler bed temperature and bed pressure under each load and the target values ​​of boiler bed temperature and bed pressure under each load; A parameter adjustment amount determination module is used to determine the parameter adjustment amount of the PID controller using a fuzzy PID control method according to the error and error change rate between each predicted value and the corresponding target value; the parameter adjustment amount of the PID controller includes: a proportional coefficient adjustment amount, an integral coefficient adjustment amount and a differential coefficient adjustment amount; The control quantity output determination module is used to determine the control quantity output of the slag cooler speed or the primary fan frequency according to the parameter adjustment amount of the PID controller.

Images