Auxiliary control system and method for oxygen content of boiler for deep peak regulation of unit

Abstract

The invention relates to a boiler oxygen amount auxiliary control system and method for deep peak regulation of a unit, and the system achieves the precise regulation of the oxygen concentration of a boiler through combining a plurality of advanced technologies, including Kalman filtering, LSTM prediction, fuzzy logic control and closed-loop feedback control. Kalman filtering is used for fusing oxygen concentration data from multiple sensors, eliminating noise interference and providing more accurate concentration estimation. The LSTM model predicts the oxygen demand change trend based on boiler load, environmental parameters and other input data. The fuzzy logic control is used for dynamically adjusting the air flow according to data such as oxygen concentration and CO concentration monitored in real time. And the closed-loop feedback control system is used for setting the error between the target oxygen concentration and the actual concentration. According to the oxygen amount auxiliary control system, the combustion efficiency of the boiler is improved, stable operation of the boiler can be kept under the condition of load fluctuation or external environment change, energy consumption and pollutant emission are reduced, and remarkable economic benefits and environmental benefits are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy and environmental engineering, and particularly to an oxygen amount auxiliary control system and method for a unit deep peak shaving boiler. Background Art

[0002] With the continuous growth of energy demand and the increasingly strict environmental protection regulations, traditional energy equipment such as coal-fired boilers still occupies an important position in the industrial and power industries. As the core equipment of the coal-fired power generation and heating system, the operating efficiency, stability and environmental protection emission level of the boiler directly affect the energy utilization effect and environmental quality. However, the combustion process of the boiler is often affected by various factors, such as load changes, fuel property fluctuations, external climate condition changes, etc. These factors often lead to the instability of the boiler combustion process, and then cause problems such as oxygen concentration fluctuations, incomplete combustion, and increased pollutant emissions.

[0003] The stability and efficiency of the combustion process largely depend on the precise control of the oxygen supply amount. The oxygen concentration in the boiler combustion process needs to be maintained within a suitable range to ensure complete combustion of the fuel, thereby improving the energy utilization efficiency and reducing the emissions of harmful gases (such as CO, NOx, SOx, etc.). However, due to the following several factors, traditional oxygen amount control methods face great challenges:

[0004] The change of the boiler load will directly affect the oxygen demand for combustion. When the load increases, the oxygen supply amount needs to increase to ensure complete combustion; when the load decreases, excessive oxygen will lead to energy waste and increased emissions. Traditional manual or simple automatic adjustment methods often cannot accurately respond to the rapid fluctuations of the load, resulting in unstable combustion; oxygen sensors are often used in modern boilers to monitor the oxygen concentration, but due to the measurement error of the sensor itself, environmental interference and data noise, the feedback of a single sensor may not be accurate enough to timely reflect the actual change of the oxygen demand of the boiler, resulting in oxygen adjustment errors, and then affecting the combustion efficiency and emission level; the boiler combustion environment is affected by external factors such as temperature, humidity, and air pressure, resulting in fluctuations in the oxygen demand under the same load. Traditional control methods often ignore these dynamic changes and are difficult to achieve efficient and stable combustion regulation; while maintaining the stability of the oxygen concentration, it is necessary to avoid energy waste and harmful gas emissions caused by over-oxidation. Excessive oxygen not only wastes energy, but also may lead to the generation of harmful substances such as nitrogen oxides (NOx). How to ensure complete combustion while avoiding excessive oxygen supply has always been a difficult point in boiler control.

[0005] The current oxygen amount control methods for boilers mainly rely on the following several ways:

[0006] PID controllers are widely used in industrial processes, but their control effects in complex systems are usually not satisfactory. For the control of boiler oxygen concentration, PID control cannot adapt to complex factors such as boiler load fluctuations and environmental changes, which easily leads to inaccurate control, large fluctuations in oxygen concentration, and inability to achieve the stability of the combustion process. Many traditional boiler systems still rely on a single oxygen sensor for feedback regulation. This method often cannot accurately capture the changes in oxygen concentration because the oxygen sensor is easily affected by factors such as temperature and pressure, resulting in inaccurate measurement results. In some old boiler systems, the regulation of oxygen quantity often relies on manual experience or simple automation systems, lacking real-time response to factors such as load changes and environmental changes, resulting in low combustion efficiency and excessive pollutant emissions. Many existing boiler control systems adopt a fixed oxygen concentration control mode. Even when the boiler load changes or other conditions change, the system cannot dynamically adjust the control strategy, resulting in the combustion process being unable to adapt to the real-time changes in load and environment.

[0007] Therefore, the existing technology has great limitations when facing the complex relationships among boiler system load fluctuations, environmental changes, and the stability between oxygen supply and combustion.

[0008] In order to improve the combustion stability of boilers, enhance energy efficiency, and reduce emissions, there is an urgent need for a new oxygen quantity regulation control system that can handle various complex factors such as load fluctuations, sensor errors, and environmental changes. This system should have the following key characteristics:

[0009] It can accurately predict the oxygen demand according to the real-time load and operating conditions of the boiler, avoiding excessive or insufficient oxygen supply; by fusing the data of multiple sensors, it eliminates the influence of single-sensor errors on oxygen concentration measurement, ensuring the accuracy and stability of oxygen concentration data; it can dynamically adjust according to load changes, environmental condition changes, and other operating parameters, and optimize oxygen supply in real time to ensure combustion stability and energy efficiency; it uses modern artificial intelligence technologies such as deep learning and fuzzy control to enhance the adaptive ability of the system, realizing fully automatic and intelligent oxygen quantity regulation, avoiding manual intervention, and improving control accuracy.

[0010] Aiming at the deficiencies in the existing technology, the present invention proposes an oxygen quantity auxiliary control system for a unit's deep peak shaving boiler based on multi-sensor data fusion, load prediction, and fuzzy control. This system optimizes the oxygen supply quantity by real-time monitoring of the boiler's oxygen concentration and load changes, combining environmental data, sensor feedback, and deep learning models, ensuring the stability and efficiency of the boiler combustion process under various working conditions. Through an intelligent and automated control system, it not only improves the combustion efficiency but also effectively reduces harmful gas emissions, meeting the increasingly stringent environmental protection requirements. Summary of the Invention

[0011] For the above purposes, the present invention provides an oxygen amount auxiliary control system and method for a unit's deep peak shaving boiler, including:

[0012] An oxygen concentration monitoring module, configured to monitor the oxygen concentration in the boiler's flue gas and output real-time data;

[0013] A data fusion module, configured to receive data from multiple oxygen sensors and perform data fusion through the Kalman filter algorithm to obtain an accurate estimated value of the oxygen concentration;

[0014] A load prediction module, configured to use historical load data and environmental data to predict the change of the boiler load based on the LSTM model, and further predict the corresponding oxygen demand;

[0015] A fuzzy logic control module, configured to generate a corresponding control output according to the difference between the oxygen concentration and the target concentration, and adjust the air flow;

[0016] A closed-loop feedback control module, configured to receive oxygen concentration deviation information and adjust the air flow in real time according to the PID control algorithm to ensure that the oxygen concentration is stable within the target range.

[0017] Further, the formula of the Kalman filter algorithm is as follows:

[0018]

[0019] K k =P k|k-1 H T (HP k|k-1 H T +R) -1

[0020]

[0021] Wherein, is the prediction of the current state, A is the state transition matrix, is the estimated state at the previous moment, B is the control input matrix, u k is the control input; based on the estimated state at the previous moment (the estimated value of the oxygen concentration), the current state is predicted using the state transition matrix A; the control input matrix B combines the external variable of the boiler's load to adjust the prediction of the current state, K k is the Kalman gain, P k|k-1 is the prediction error covariance, H is the measurement matrix, z k is the actual measurement value, R is the covariance of the measurement noise; calculate the Kalman gain K k , which determines how to update the prediction value based on the sensor data at the current moment. The larger the Kalman gain, the greater the weight of the sensor data; the smaller the gain, the greater the weight of the prediction data.

[0022] Furthermore, the formula of the LSTM model includes:

[0023] Forget gate:

[0024] f t = σ(W f · [h t-1 , x t + b f );

[0025] Input gate:

[0026] i t = σ(W i · [h t-1 , x t + b i );

[0027] Candidate cell state calculation unit:

[0028]

[0029] Update unit state:

[0030]

[0031] Output gate:

[0032] o t = σ(W o · [h t-1 , x t + b o )

[0033] h t = o t · tanh(C t )

[0034] where f t is the output of the forget gate, i t is the output of the input gate, is the candidate cell state, C t is the cell state, o t is the output of the output gate, and h t is the current output of the LSTM network.

[0035] Furthermore, the working process of the fuzzy logic control module includes:

[0036] Fuzzify the input oxygen concentration to obtain the membership function value;

[0037] Based on the fuzzy rule base, perform fuzzy inference to obtain the control output;

[0038] The air flow regulation amount is calculated by a defuzzification method.

[0039] Furthermore, the defuzzification method adopts the centroid method, and its formula is:

[0040]

[0041] where y i is the output value of the fuzzy rule, and μ i is the corresponding membership degree.

[0042] Furthermore, the closed-loop feedback control module uses the PID control algorithm, and the control formula is:

[0043]

[0044] where u k is the control input (air flow regulation amount) at the current moment, C k is the oxygen concentration at the current moment, C target is the set target oxygen concentration, K p and K d are the proportional, integral, and differential gains respectively, and Δt is the time step.

[0045] Furthermore, the oxygen concentration monitoring module includes multiple oxygen sensors, and the data fusion module uses Kalman filtering to fuse the data of each sensor, thereby reducing the influence of sensor noise on the measurement result.

[0046] Furthermore, the system also includes a load regulation module for adjusting the combustion intensity of the burner according to the predicted load change, thereby optimizing the air flow and combustion efficiency.

[0047] Furthermore, the fuzzy logic control module adopts a multi-layer fuzzy rule base to make detailed control decisions according to the deviation between the real-time oxygen concentration and the target concentration.

[0048] Furthermore, the system automatically adjusts the air flow by real-time monitoring of the oxygen concentration of the boiler's exhaust gas, load change, and environmental parameters, ensuring stable combustion of the boiler under various loads and environmental changes.

[0049] A control method for the oxygen-assisted control system of a unit's deep peak shaving boiler includes the following steps:

[0050] Step 1: Real-time monitor the oxygen concentration of the boiler through the oxygen concentration monitoring module;

[0051] Step 2: Use the Kalman filtering algorithm through the data fusion module to fuse the oxygen concentration data from different sensors to obtain an accurate oxygen concentration estimate value;

[0052] Step 3: The load prediction module uses the LSTM model to predict the boiler load and predicts the oxygen demand according to the load change;

[0053] Step 4: The fuzzy logic control module generates an air flow regulation signal and obtains the final control output through the defuzzification process;

[0054] Step 5: The closed-loop feedback control module adjusts the air flow according to the PID control algorithm to ensure that the oxygen concentration is stable within the target value range.

[0055] Advantages of the present invention: By introducing the oxygen content auxiliary control system for the unit's deep peak shaving boiler, the present invention uses the Kalman filter algorithm to fuse multi-sensor data, the LSTM model to predict the boiler load change, the fuzzy logic control to adjust the oxygen concentration, and the closed-loop feedback mechanism to finely adjust the air flow, realizing the precise control of the boiler oxygen content and the optimization of combustion stability. This system can monitor and adjust the oxygen concentration in real time, reduce problems such as low combustion efficiency and excessive emissions caused by too much or too little oxygen, and effectively improve the economy and environmental protection of boiler operation. At the same time, through load prediction and dynamic adjustment, the lag of traditional control methods is avoided, ensuring that the boiler can maintain an efficient and stable operation state under different load conditions, thereby reducing energy consumption, reducing pollutant emissions, improving energy utilization efficiency, and having significant environmental and economic benefits. Brief Description of the Drawings

[0056] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only those of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0057] Figure 1 It is a schematic diagram of the framework of the oxygen content auxiliary control system for the boiler in the embodiment of the present invention;

[0058] Figure 2 It is a schematic diagram of the flow of the Kalman filter module in the embodiment of the present invention;

[0059] Figure 3 It is a schematic diagram of the flow of the LSTM model load and oxygen demand prediction in the embodiment of the present invention;

[0060] Figure 4 It is a schematic diagram of the fuzzy control system module in the embodiment of the present invention;

[0061] Figure 5 It is a schematic diagram of the working process of the PID closed-loop controller in the embodiment of the present invention. Detailed Embodiment

[0062] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0063] It should be noted that in the specification, when referring to "an embodiment", "embodiment", "exemplary embodiment", "some embodiments", etc., it indicates that the described embodiment may include specific features, structures or characteristics, but not necessarily every embodiment includes such specific features, structures or characteristics. Additionally, when combining embodiments to describe specific features, structures or characteristics, implementing such features, structures or characteristics in combination with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the relevant art.

[0064] Generally, terms can be understood at least in part from their use in context. For example, at least in part depending on the context, the term "one or more" as used herein can be used to describe any feature, structure or characteristic in a singular sense, or can be used to describe a combination of features, structures or characteristics in a plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but rather, at least in part depending on the context, can allow for the existence of other factors that may not be explicitly described.

[0065] Embodiment 1

[0066] The present invention provides a unit deep peak shaving boiler oxygen content auxiliary control system and method, which solves the problems of hysteresis, inaccurate oxygen content and combustion fluctuations during deep peak shaving existing in traditional boiler oxygen content control. The following are the specific implementation manners of this system:

[0067] Refer to Figure 1 as shown

[0068] 1. System hardware architecture design

[0069] This system includes the following key hardware components:

[0070] Sensor array: Installed at multiple key positions of the boiler for real-time monitoring of the boiler operation status, specifically including:

[0071] Oxygen sensor: Installed at the boiler flue gas outlet for real-time monitoring of the oxygen concentration in the flue gas.

[0072] Carbon monoxide (CO) sensor: Installed in the boiler flue for real-time monitoring of the carbon monoxide concentration to judge whether the combustion is complete.

[0073] Nitrogen Oxide (NOx) Sensor: Installed in the boiler flue, it monitors the NOx concentration in real time and evaluates the pollutant emissions during combustion.

[0074] Temperature and Humidity Sensor: Installed in the boiler combustion area and flue, it monitors the temperature and humidity changes inside and outside the boiler in real time.

[0075] Data Acquisition Module: Used to receive the signals output by each sensor and transmit the data to the edge computing module through industrial Ethernet or other communication methods.

[0076] Edge Computing Module: One edge computing unit is configured for each boiler combustion unit, which is responsible for receiving sensor data, performing data preprocessing, filtering, preliminary analysis, and local control decisions.

[0077] Central Control System: The central control system receives the data transmitted by the edge computing module, performs global optimization, and generates final control instructions. This system uses a deep learning model to predict the oxygen demand and adjusts the air supply volume using a fuzzy logic control algorithm.

[0078] Actuator: According to the instructions of the central control system, it adjusts the air flow rate and fuel supply volume. The actuator includes the valves for adjusting the fan and the fuel supply system (gas burner, liquid fuel injector).

[0079] 2. Multi-Sensor Data Fusion and Calibration

[0080] Refer to Figure 2 as shown

[0081] To improve the accuracy of oxygen concentration measurement, the following scheme is adopted in this system for data fusion and real-time calibration:

[0082] Sensor Signal Acquisition and Preprocessing: Oxygen, CO, NOx, and temperature and humidity sensors collect data in real time and transmit it to the edge computing module. The edge computing module filters the noise of these data and eliminates invalid signals.

[0083] Data Fusion: The Kalman filter algorithm is used to fuse the sensor data. By integrating the inputs of the oxygen sensor, CO sensor, and temperature and humidity sensor, a more accurate oxygen concentration value is calculated. The Kalman filter eliminates measurement errors by combining system state prediction and measurement data, improving the accuracy of the oxygen concentration;

[0084] During the multi-sensor data fusion process, the outputs of the sensors are affected by noise and interference. Therefore, the Kalman filter algorithm is used to fuse the data of multiple sensors to obtain a more accurate oxygen concentration. The basic formula of the Kalman filter is as follows:

[0085] Kalman Filter Formula:

[0086] 1. Prediction update equation:

[0087]

[0088] Wherein, is the prediction of the current state, A is the state transition matrix, is the estimated state at the previous moment, B is the control input matrix, u k is the control input; based on the estimated state at the previous moment (the estimated value of oxygen concentration), the current state is predicted using the state transition matrix A; the control input matrix B combines the external variable of the boiler load to adjust the prediction of the current state.

[0089] 2. Measurement update equation:

[0090] K k = P k|k-1 H T (HP k|k-1 H T + R) -1

[0091]

[0092] Wherein, K k is the Kalman gain, P k|k-1 is the prediction error covariance, H is the measurement matrix, z k is the actual measurement value, R is the covariance of the measurement noise; calculate the Kalman gain K k , which determines how to update the prediction value based on the sensor data at the current moment. The larger the Kalman gain, the greater the weight of the sensor data; the smaller the gain, the greater the weight of the prediction data.

[0093] H represents the measurement matrix of the sensor, describing how the sensor maps the state value to the measurement value. The measurement noise covariance R represents the uncertainty of the measurement data; the prediction value is corrected through the Kalman gain K k to make it closer to the real measurement data z k .

[0094] The updated state estimate is the final estimated value of the current oxygen concentration.

[0095] Real-time correction: The system uses a model-based correction mechanism to compare the fused oxygen concentration with the system prediction value. If there is a large deviation, the system will automatically correct the sensor output to ensure that the oxygen concentration data accurately reflects the combustion state of the boiler.

[0096] 3. Load prediction and oxygen demand prediction based on deep learning

[0097] Refer toFigure 3 shown

[0098] To achieve precise oxygen content regulation, based on historical data and real-time data, this system uses a long short-term memory network to predict the load and oxygen demand:

[0099] Data preparation: The system collects historical boiler operation data, including boiler load fluctuations, fuel consumption, and flue gas oxygen concentration data. The training data set includes different operation modes of the boiler and external environment changes (climate, humidity).

[0100] Model training: The load data is trained through an LSTM network to learn the temporal pattern of boiler load changes. During the training process, the LSTM model can capture the impact of load changes on oxygen demand.

[0101] Real-time prediction: During the operation of the system, the LSTM model predicts the change in oxygen demand in the future for a period of time based on real-time load data. The prediction result provides accurate oxygen demand prediction information for the central control system.

[0102] The deep learning model (LSTM) is used to predict the impact of boiler load changes on oxygen demand. The LSTM model models the relationship between boiler load and oxygen demand through long time series data.

[0103] The deep learning model (LSTM) is used to predict the impact of boiler load changes on oxygen demand. The LSTM model models the relationship between boiler load and oxygen demand through long time series data.

[0104] Calculation process of LSTM:

[0105] The calculation formula of the LSTM model is as follows, mainly including the operations of the input gate, forget gate, and output gate:

[0106] Forget gate:

[0107] f t =σ(W f ·[h t-1 , x t +b f )

[0108] where f t is the output of the forget gate, σ is the sigmoid activation function, W f is the weight matrix, b f is the bias term, h t-1 is the output of the previous moment, and x t is the input of the current moment; Input: the output h t-1 of the previous moment, and the input x t(Boiler load, temperature and humidity). Output: Output f of the forget gate t , obtained through the sigmoid activation function σ, representing the proportion of information to be retained at the current time step. Calculate the forget gate f t Controls the information to be "forgotten" or "retained" at the current time step. If f t is large, it means that more information at that time step needs to be retained; if f t is small, more information will be forgotten.

[0109] Input gate:

[0110] i t = σ(W i ·[h t-1 , x t + b i )

[0111] where i t is the output of the input gate, W i is the weight matrix, b i is the bias term; calculate the input gate i t , determining whether the input information at the current time step needs to be stored in the "cell state".

[0112] Candidate cell state calculation unit:

[0113]

[0114] where is the candidate unit, W C is the weight matrix, b C is the bias term, and tanh is the hyperbolic tangent activation function; calculate the candidate cell state It will together with the outputs of the forget gate and the input gate determine the update of the cell state at the current time step.

[0115] Update unit state:

[0116]

[0117] where C t is the unit state at the current time step, C t-1 is the unit state at the previous time step; through the forget gate f t and the input gate i t control the cell state C t at the current time step; C t represents the "memory" part of the LSTM network, determining how to calculate the oxygen demand based on historical data and current input.

[0118] Output gate:

[0119] ot = σ(W o · [h t-1 , x t + b o )

[0120] h t = o t · tanh(C t )

[0121] where o t is the output of the output gate, W o is the weight matrix, b o is the bias term, and h t is the output at the current time step; the output h t at the current time step is calculated through the output gate o t and the cell state C t ; the output h t is the prediction result of the boiler oxygen demand.

[0122] 4. Fuzzy Logic Control and Hierarchical Control Mechanism

[0123] Refer to Figure 4 shown

[0124] This system uses fuzzy logic control (FLC) and a hierarchical control mechanism to finely adjust the oxygen supply to cope with complex and uncertain combustion conditions.

[0125] Fuzzy rule design: Based on the real-time data of the oxygen concentration, CO concentration, and NOx concentration in the boiler flue gas, a fuzzy control rule base is established, including the following core rules:

[0126] When the oxygen concentration is too high, reduce the air volume.

[0127] When the oxygen concentration is too low, increase the air volume.

[0128] When the CO concentration is too high and the oxygen concentration is low, increase the air flow rate to promote complete combustion.

[0129] When the NOx concentration is too high, reduce the air flow rate and reduce the supply of excess oxygen.

[0130] Hierarchical control: Oxygen control is divided into two levels:

[0131] Primary control: According to the load changes of the boiler and environmental factors (external temperature and humidity), adjust the overall air flow rate of the boiler. The primary control sets the preliminary air supply of the boiler based on the oxygen demand predicted by the LSTM.

[0132] Secondary control: Based on the real-time monitoring data inside the boiler (oxygen, CO, NOx concentrations), fine-tune each combustion unit to ensure that the oxygen supply for each unit meets the demand.

[0133] Fuzzy inference system: The fuzzy controller adjusts the air flow in real time according to the real-time data, precisely controls the oxygen supply of the boiler, and ensures the stability of the combustion process.

[0134] In the regulation of oxygen concentration, fuzzy logic control (FLC) is adopted to handle the complexity of boiler oxygen concentration regulation. The basic process of fuzzy control is as follows:

[0135] Fuzzy inference process:

[0136] 1. Fuzzify the input: According to the real-time monitoring data (oxygen concentration, CO concentration, NOx concentration), convert the input data into fuzzy sets through the fuzzification function. For example, the fuzzy sets of oxygen concentration can be "low", "normal", and "high".

[0137]

[0138] Among them, μ low (x) is the membership function of "low" oxygen concentration, μ high (x) is the membership function of "high" oxygen concentration, x low and x high are the low and high thresholds of oxygen concentration, Δx is the interval width; convert the actual oxygen concentration into a fuzzy value through the membership function for inference with the fuzzy rule base. According to different oxygen concentrations, obtain the membership degrees of "low" and "high" oxygen concentrations respectively.

[0139] 2. Fuzzy rule base: Establish a fuzzy rule base according to the fuzzy input. For example:

[0140] If the oxygen concentration is "low", then increase the air flow.

[0141] If the oxygen concentration is "high", then decrease the air flow.

[0142] Fuzzy inference: According to the fuzzy rule base, combine the input values for inference to obtain the fuzzy control output.

[0143] Defuzzification: Convert the fuzzy output into the actual control value (air flow). Defuzzification uses the centroid method:

[0144]

[0145] Among them, y i is the output value of the fuzzy rule, μ iis the corresponding membership degree; according to the outputs and membership degrees of all fuzzy rules, calculate the weighted average value to obtain the final control output (air flow regulation amount).

[0146] The fuzzy logic controller dynamically adjusts the air flow according to the real-time oxygen concentration, CO concentration, and NOx concentration. When the oxygen concentration is low, increase the air flow to promote complete combustion; when the oxygen concentration is high, reduce the air flow to avoid over-oxidation.

[0147] 5. Edge Computing and Closed-Loop Feedback Optimization

[0148] Refer to Figure 5 as shown

[0149] The system ensures real-time response and continuous optimization of performance through edge computing and closed-loop feedback optimization mechanisms:

[0150] Edge computing node: Each combustion unit is equipped with an edge computing node, which processes data in real time on-site, conducts preliminary oxygen demand prediction and air flow regulation. This node makes an immediate judgment on the boiler state and feeds back the decision result to the central control system.

[0151] Central control system optimization: The preliminary decision results of the edge computing module are transmitted to the central control system, and the central system makes further optimization decisions based on global data. The central system continuously adjusts the oxygen supply strategy through reinforcement learning algorithms to optimize combustion efficiency and reduce emissions.

[0152] Closed-loop feedback mechanism: During the operation of the boiler, the system monitors the oxygen concentration, CO concentration, and NOx concentration data in real time. If a deviation is found, the system will immediately adjust the control strategy for dynamic correction to ensure that the boiler always maintains the best combustion state.

[0153] The closed-loop feedback control system continuously adjusts the air flow through the closed-loop feedback mechanism to ensure combustion stability. Assume that the control target is to reach the set value C of the oxygen concentration target , and the feedback control calculation formula of the system is as follows:

[0154] Feedback control formula:

[0155]

[0156] where, u k is the control input (air flow regulation amount) at the current moment, C k is the oxygen concentration at the current moment, C target is the set target oxygen concentration, K p and K dare the proportional, integral, and derivative gains respectively, and Δt is the time step; by calculating the error (the difference between the target oxygen concentration and the current concentration) and combining the proportional, integral, and derivative terms, the system adjusts the air flow in real time to ensure that the oxygen concentration remains near the target value. This closed-loop control mechanism ensures the stable operation of the boiler under load fluctuations and external environmental changes; through the proportional term K p , the integral term K i , and the derivative term K d calculate the control input. According to the error and its rate of change, adjust the air flow in real time to keep the oxygen concentration stable.

[0157] 6. System Implementation and Commissioning

[0158] Hardware Installation and Wiring: Install oxygen sensors, carbon monoxide sensors, NOx sensors, and temperature and humidity sensor devices at key positions in the boiler and flue. All sensor signals are transmitted to the edge computing node through the data acquisition module.

[0159] Data Communication and Model Deployment: The edge computing node and the central control system transmit data through industrial Ethernet or a dedicated communication link. The deep learning model (LSTM) and the fuzzy controller are deployed in the central control system and optimized regularly.

[0160] System Commissioning and Optimization: During system commissioning, adjust the fuzzy control rules and LSTM model parameters according to the actual operating conditions of the boiler. By testing the boiler's response under different loads and different environmental conditions, ensure that the system can adaptively adjust the oxygen supply.

[0161] 7. System Operation and Feedback Optimization

[0162] Real-time Monitoring and Evaluation: During system operation, real-time monitor key parameters such as the oxygen concentration, CO concentration, NOx concentration, and combustion efficiency of the boiler. The system continuously optimizes the control strategy based on this data to ensure the accuracy of oxygen regulation.

[0163] Feedback Optimization: Based on the long-term operation data of the boiler, the central control system continuously optimizes the control strategy through reinforcement learning algorithms to further improve the accuracy of oxygen control and combustion efficiency.

[0164] This implementation method combines technologies such as precise design of the hardware architecture, multi-sensor data fusion, deep learning prediction, fuzzy logic control, edge computing, and feedback optimization to solve the problems of hysteresis and inaccurate regulation existing in traditional boiler oxygen control. The system can adjust the oxygen supply in real time and precisely under deep peak shaving conditions, ensure the stability of boiler combustion, improve combustion efficiency, and reduce pollutant emissions.

[0165] In the foregoing technical solutions, multiple key hardware devices are involved, such as oxygen sensors, temperature and humidity sensors, controllers, data acquisition devices, etc. The selection of these hardware devices is crucial for the stability and accuracy of the system. The following hardware models and brands involved are based on common market options:

[0166] 1. Oxygen concentration sensor

[0167] Figaro TGS-813: Suitable for gas monitoring, capable of detecting the oxygen concentration in the air and providing good stability.

[0168] 2. Carbon dioxide (CO2) and carbon monoxide (CO) sensors

[0169] Alphasense CO-A1: Carbon monoxide sensor.

[0170] Honeywell CO2 Sensor (CM-001): Carbon dioxide sensor.

[0171] 3. Temperature and humidity sensor:

[0172] Sensirion SHT35: High-precision temperature and humidity sensor.

[0173] 4. PLC controller:

[0174] Siemens Siemens S7-1200.

[0175] 5. Data acquisition system

[0176] National Instruments NI cDAQ-9178: Modular data acquisition system.

[0177] 6. Human-machine interface (HMI)

[0178] Siemens TP700 Comfort.

[0179] 7. Computing device and software platform

[0180] Computing devices and software platforms for supporting complex calculations such as LSTM models, Kalman filtering algorithms, and fuzzy control.

[0181] Intel Xeon series processors.

[0182] NVIDIA Tesla V100.

[0183] MATLAB&Simulink MATLAB: For model development, data analysis, and algorithm simulation.

[0184] Python + TensorFlow (open source)

[0185] Use the Python programming language and the TensorFlow framework to train and infer the LSTM model.

[0186] 8. Network device

[0187] Cisco Catalyst 9000 series switches: industrial-grade network devices.

[0188] This invention covers any alternatives, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To enable the public to have a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments of this invention. However, those skilled in the art can fully understand this invention without these detailed descriptions. Additionally, well-known methods, processes, procedures, components, and circuits are not described in detail to avoid unnecessary confusion to the essence of this invention.

[0189] The above are only the preferred embodiments of this invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of this invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this invention.

Claims

1. An oxygen content auxiliary control system and method for a unit's deep peak shaving boiler, characterized in that, Including: An oxygen concentration monitoring module for monitoring the oxygen concentration in the boiler flue gas and outputting real-time data; A data fusion module for receiving data from multiple oxygen sensors and performing data fusion through the Kalman filter algorithm to obtain an accurate estimated value of the oxygen concentration; A load prediction module for predicting the boiler load change based on the historical load data and environmental data using the LSTM model, and then predicting the corresponding oxygen demand; A fuzzy logic control module for generating a corresponding control output according to the difference between the oxygen concentration and the target concentration to adjust the air flow; A closed-loop feedback control module for receiving oxygen concentration deviation information and adjusting the air flow in real time according to the PID control algorithm to ensure that the oxygen concentration is stable within the target range.

2. The oxygen content auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that The formula of the Kalman filter algorithm is as follows: K k = P k|k-1 H T (HP k|k-1 H T + R) -1 Among them, is the prediction of the current state, A is the state transition matrix, is the estimated state at the previous moment, B is the control input matrix, and u k is the control input; based on the estimated state at the previous moment (the estimated value of oxygen concentration), the current state is predicted using the state transition matrix A; the control input matrix B combines the external variable of the boiler load to adjust the prediction of the current state, and K k is the Kalman gain, and P k|k-1 is the prediction error covariance, H is the measurement matrix, and z k is the actual measurement value, and R is the covariance of the measurement noise; the Kalman gain K k is calculated, which determines how to update the prediction value based on sensor data at the current moment; the larger the Kalman gain, the greater the weight of the sensor data; the smaller the gain, the greater the weight of the prediction data.

3. An oxygen content auxiliary control system and method for a unit deep peak shaving boiler according to claim 1, characterized in that, The formula of the LSTM model includes: Forgotten gate: f t = σ(W f · [h t-1 , x t + b f ); Input gate: i t = σ(W i · [h t-1 , x t + b i ); Candidate cell state calculation unit: Update unit state: Output gate: o t = σ(W o · [h t-1 , x t + b o ) h t = o t ·tanh(C t ) where f t is the output of the forget gate, i t is the output of the input gate, is the candidate cell state, C t is the cell state, o t is the output of the output gate, h t is the current output of the LSTM network.

4. An oxygen content auxiliary control system and method for a unit deep peak shaving boiler according to claim 1, characterized in that, The working process of the fuzzy logic control module includes: Fuzzifying the input oxygen concentration to obtain the membership function value; Based on the fuzzy rule base, performing fuzzy inference to obtain the control output; Calculating the air flow adjustment amount through the defuzzification method.

5. An oxygen content auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that The defuzzification method adopts the centroid method, and its formula is: where y i is the output value of the fuzzy rule, and μ i is the corresponding membership degree.

6. An oxygen content auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that, The closed-loop feedback control module uses the PID control algorithm, and the control formula is: where, u k is the control input (air flow rate adjustment) at the current moment, C k is the oxygen concentration at the current moment, C target is the set target oxygen concentration, K p and K d are the proportional, integral, and derivative gains respectively, and Δt is the time step.

7. An oxygen amount auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that The oxygen concentration monitoring module includes multiple oxygen sensors, and the data fusion module uses the Kalman filter to fuse the data of each sensor, thereby reducing the influence of sensor noise on the measurement result.

8. An oxygen content auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that, The system further includes a load adjustment module for adjusting the combustion intensity of the burner according to the predicted load change, thereby optimizing the air flow and combustion efficiency.

9. The oxygen content auxiliary control system and method for a unit's deep peak shaving boiler according to claim 1, characterized in that, The fuzzy logic control module adopts a multi-layer fuzzy rule base to make detailed control decisions according to the deviation between the real-time oxygen concentration and the target concentration. The system automatically adjusts the air flow by real-time monitoring the oxygen concentration, load change, and environmental parameters of the boiler flue gas, ensuring stable combustion of the boiler under various loads and environmental changes.

10. A control method for the oxygen content auxiliary control system of the unit deep peak shaving boiler according to any one of claims 1 to 9, characterized in that, Including the following steps: Step 1: Real-time monitor the oxygen concentration of the boiler through the oxygen concentration monitoring module; Step 2: Use the Kalman filter algorithm through the data fusion module to fuse the oxygen concentration data from different sensors to obtain an accurate estimated value of the oxygen concentration; Step 3: Predict the boiler load using the LSTM model through the load prediction module and predict the oxygen demand according to the load change; Step 4: Generate an air flow adjustment signal through the fuzzy logic control module and obtain the final control output through the defuzzification process; Step 5: Adjust the air flow according to the PID control algorithm through the closed-loop feedback control module to ensure that the oxygen concentration is stable within the target value range.

Images