Method, system, device and medium for controlling waste incineration with large proportion of sludge blending combustion

Abstract

The application relates to the technical field of solid waste incineration treatment and intelligent incineration control, and discloses a garbage incineration control method, system, equipment and medium with a large proportion of sludge mixed burning. The method comprises the following steps: collecting multi-source sensing monitoring data in real time under the garbage incineration working condition; performing edge pre-processing, extracting core process characteristics, and completing real-time anomaly identification and hierarchical early warning; predicting the average low calorific value of the garbage and the sludge entering the furnace based on a time sequence prediction model, and determining the maximum safe mixing and burning proportion of the sludge; under the proportion and the preset safe operation constraint, the comprehensive objective function constructed is co-optimized, the optimal incineration control parameter combination is solved, and the actuator is driven to adjust. The application realizes safe mixing and burning of a large proportion of sludge, and improves the combustion stability and the collaborative emission reduction capacity of pollutants.

Description

Technical Field

[0001] This application relates to the field of solid waste incineration treatment and intelligent incineration control technology, specifically to a waste incineration control method, system, equipment and medium for large-proportion co-firing of sludge. Background Technology

[0002] With the acceleration of urbanization, the generation of municipal solid waste and sewage sludge continues to increase, making the efficient and environmentally friendly disposal of these two types of solid waste a major issue in the environmental field. Utilizing existing municipal solid waste incineration facilities to co-process municipal sewage sludge can avoid the high investment required for building independent sludge treatment facilities, while fully utilizing the existing flue gas purification systems and waste heat recovery facilities of the incineration plant, representing an important technological direction for sludge treatment. However, current technologies still have many shortcomings in the large-scale co-firing of sludge.

[0003] The existing implementation schemes mainly include the following two types: (a) Conventional manual experience-based co-firing + single-loop PID (proportional-integral-derivative) incineration control scheme.

[0004] The main technical aspects of this solution are: The sludge co-firing ratio is fixed and depends on the experience of the boiler operator. Usually, the co-firing ratio is less than 15%, which makes it impossible to achieve stable and safe high-ratio sludge co-firing treatment.

[0005] The incineration control is mainly based on a single-loop PID controller, which only independently adjusts the furnace temperature, total air volume and grate speed. It lacks coordination and linkage between various parameters, making it difficult to achieve multiple objectives such as thermal efficiency, steam production and pollutant emissions, and often resulting in an imbalance of indicators.

[0006] The air distribution is mainly based on the total air volume adjustment, without the ability to finely distribute the air in segments, which can easily lead to problems such as large fluctuations in furnace temperature, incomplete combustion, flameout, and coking.

[0007] It lacks online calorific value prediction capabilities, visual flame recognition methods, and a coupled control mechanism between co-firing ratio, grate movement, and segmented air volume, resulting in poor operational stability.

[0008] Abnormal operating conditions rely on manual judgment and intervention, resulting in a significant delay in response and making it difficult to adapt to rapid and large fluctuations in the calorific value of waste and sludge.

[0009] (II) Scheme for sludge drying followed by co-firing + conventional distributed control system (DCS) incineration control.

[0010] The main technical aspects of this solution are: First, the sludge is thermally dried to reduce its moisture content. Then, the dried sludge is mixed with the garbage and fed into the furnace. The drying system has high investment costs, high energy consumption, and poor overall economic efficiency.

[0011] While basic interlocking control is achieved through a distributed control system with load stabilization as the primary control objective, the dynamic matching problem of combustion under high-proportion co-firing conditions remains unresolved.

[0012] It focuses on end-of-pipe flue gas treatment, lacks intelligent prediction and feedforward control capabilities, has insufficient optimization and control of the combustion process in the furnace, is weakly adaptable to fluctuations in the calorific value of waste and sludge, and has failed to achieve multi-objective synergistic optimization of combustion efficiency, steam production and pollutant emissions.

[0013] Therefore, there is an urgent need to propose an intelligent incineration control scheme that can achieve a large proportion of sludge co-firing in waste incineration, in order to break through the bottleneck of the traditional limited co-firing ratio, solve the problems of combustion instability and excessive pollutant emissions under the large proportion co-firing condition, and achieve stable, efficient and clean co-firing disposal. Summary of the Invention

[0014] To address the aforementioned issues, this application provides a waste incineration control method, system, equipment, and medium for high-proportion sludge co-firing. It collects incineration operating data in real time through multimodal full-domain perception, performs data preprocessing and feature extraction locally via edge computing, dynamically determines the maximum safe sludge co-firing ratio based on a time-series prediction model, and uses a multi-objective optimization algorithm to solve for the optimal combination of incineration control parameters, driving closed-loop adjustment of the on-site actuators. This achieves stable combustion and compliant pollutant emissions under conditions where the sludge co-firing ratio is ≥40%.

[0015] The technical solution adopted in this application is as follows: Firstly, this application provides a method for controlling waste incineration with a high proportion of sludge co-firing, including: Real-time acquisition of multi-source sensor monitoring data under waste incineration conditions; The collected multi-source sensor monitoring data is preprocessed at the edge end to extract the core process features corresponding to the waste incineration conditions, and real-time anomaly identification and graded early warning of waste incineration conditions are completed based on the core process features. Based on the trained time-series prediction model, the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace are predicted in future periods according to the core process characteristics, pre-processed multi-source sensor monitoring data and auxiliary input parameters. Based on the predicted average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, and the preset safe operation constraints, the maximum safe co-firing ratio of sludge is determined. A comprehensive objective function is constructed. Under the constraints of the maximum safe co-firing ratio of sludge and the preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function and solve for the optimal combination of incineration control parameters. Based on the optimal combination of incineration control parameters, the various actuators on site are driven to adjust the waste incineration conditions.

[0016] Secondly, this application also provides a waste incineration control system for a large proportion of sludge co-firing, comprising: The sensing layer is used to collect multi-source sensor monitoring data in real time under waste incineration conditions. The edge processing layer is used to preprocess the collected multi-source sensor monitoring data at the edge, extract the core process features corresponding to the waste incineration conditions, and complete the real-time anomaly identification and graded early warning of the waste incineration conditions based on the core process features. The intelligent decision-making layer is used to predict the average lower heating value of waste and sludge entering the furnace in future periods based on a trained time-series prediction model, core process characteristics, pre-processed multi-source sensor monitoring data, and auxiliary input parameters. Based on the predicted average lower heating values ​​of the waste and sludge, and preset safe operation constraints, it determines the maximum safe co-firing ratio of sludge. A comprehensive objective function is constructed, and under the constraints of the maximum safe co-firing ratio and preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function to find the optimal combination of incineration control parameters. The execution control layer is used to drive the various actuators on site to adjust the waste incineration conditions based on the optimal combination of incineration control parameters.

[0017] Thirdly, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described waste incineration control method for high-proportion co-firing of sludge.

[0018] Fourthly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described waste incineration control method for high-proportion co-firing of sludge.

[0019] The above-mentioned technical solution adopted in this application can achieve the following beneficial effects: This application uses a time-series prediction model to dynamically predict the average lower heating value of waste and sludge fed into the furnace, and combines preset safe operation constraints to determine the maximum safe co-combustion ratio of sludge. This breaks through the limitations of traditional fixed-ratio co-combustion, maximizes the amount of sludge disposed of while ensuring safe combustion, and significantly improves the sludge co-processing capacity of the incineration plant.

[0020] Under the dual constraints of the maximum safe sludge co-firing ratio and the preset safe operation constraints, this application performs synergistic optimization on a comprehensive objective function that includes at least the actual co-firing ratio, furnace temperature, pollutant emissions, and auxiliary equipment power consumption. By incorporating multiple objectives into the same optimization framework, it effectively suppresses large fluctuations in furnace temperature, avoids flameout and coking, and ensures that flue gas pollutants meet emission standards.

[0021] This application completes data preprocessing, feature extraction, and real-time anomaly identification and hierarchical early warning at the edge, shortening the identification and response time of abnormal operating conditions, avoiding data transmission delay and packet loss risks, and providing a safe and reliable data foundation for intelligent decision-making.

[0022] This application uses a time-series prediction model to make forward-looking predictions of the calorific value of materials to be fed into the furnace in the future, forming a closed-loop control architecture of "prediction feedforward - optimization decision - execution feedback", which significantly improves the ability to adapt to large fluctuations in the calorific value of waste and sludge.

[0023] This application features a fully automated closed-loop operation from data acquisition to execution and adjustment, which significantly improves the system's automation rate, reduces reliance on manual intervention and the labor intensity of operators, and enhances the intelligence level of incineration plant operation. Attached Figure Description

[0024] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A schematic diagram of a waste incineration control system for high-proportion co-firing of sludge according to an embodiment of this application is shown. Figure 2 A schematic flowchart of a waste incineration control method with a high proportion of sludge co-firing according to an embodiment of this application is shown. Figure 3 A schematic diagram of the structure of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Figure 1 A schematic diagram of a waste incineration control system for high-proportion co-firing of sludge according to an embodiment of this application is shown. (Refer to...) Figure 1 As shown, this embodiment provides a waste incineration control system 100 for high-proportion co-firing of sludge, comprising: The sensing layer 110 is used to collect multi-source sensor monitoring data in real time under the waste incineration operation. The edge processing layer 120 is used to perform edge preprocessing on the collected multi-source sensor monitoring data, extract the core process features corresponding to the waste incineration conditions, and complete real-time anomaly identification and graded early warning of waste incineration conditions based on the core process features. The intelligent decision layer 130 is used to predict the average lower heating value of waste and sludge entering the furnace in future periods based on a trained time-series prediction model, core process characteristics, pre-processed multi-source sensor monitoring data, and auxiliary input parameters. Based on the predicted average lower heating values ​​of the waste and sludge, and preset safe operation constraints, it determines the maximum safe co-firing ratio of sludge. A comprehensive objective function is constructed, and under the constraints of the maximum safe co-firing ratio and preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function to solve for the optimal combination of incineration control parameters. The control layer 140 is used to drive the various actuators on site to adjust the waste incineration conditions based on the optimal combination of incineration control parameters.

[0027] In this embodiment, the system adopts a distributed hierarchical architecture, which is divided into a perception layer 110, an edge processing layer 120, an intelligent decision-making layer 130, and an execution and control layer 140 from top to bottom. It is also equipped with a standardized industrial communication protocol for seamless integration with the incineration plant's distributed control system. The system achieves closed-loop data transmission, real-time command issuance, and dynamic adaptive control of operating conditions throughout the entire process, requiring virtually no human intervention under non-extreme abnormal operating conditions.

[0028] Based on the above system, Figure 2 A schematic flow diagram of a waste incineration control method for high-proportion co-firing of sludge according to an embodiment of this application is shown. (Refer to...) Figure 2 As shown, this embodiment provides a waste incineration control method with a large proportion of sludge co-firing, including steps S210 to S260.

[0029] Step S210: Collect multi-source sensor monitoring data in real time under waste incineration conditions.

[0030] The perception layer 110 achieves real-time acquisition of multi-source sensing monitoring data through multi-source sensing devices. These devices are deployed at various key points in the waste incinerator, covering four core monitoring dimensions: the properties of the material entering the furnace, the combustion conditions in the furnace, flue gas emission parameters, and the equipment operating status. The perception layer 110 employs collaborative sensing by multi-source sensing devices to achieve comprehensive, real-time, and accurate acquisition of parameters throughout the entire incineration process, providing raw data support for subsequent edge processing and intelligent decision-making.

[0031] In some optional implementations, step S210 involves real-time acquisition of multi-source sensor monitoring data under waste incineration conditions, including: acquiring waste material distribution data via lidar to determine the weight of waste in the waste bin and the hourly feed rate of waste into the furnace; acquiring flame images inside the furnace via a flame vision camera to determine the average flame brightness and flame area; acquiring grate section temperature, furnace flame temperature, furnace outlet temperature, flue gas temperature, and material bed surface temperature via temperature sensors; acquiring furnace negative pressure, primary air pressure, secondary air pressure, flue gas pressure, and pressure difference between each air chamber via pressure sensors; acquiring the opening degree of each section of air damper via damper opening sensors; and acquiring O2 concentration, CO concentration, and NO concentration in the flue gas via a flue gas analyzer. x Concentrations of SO2, HCl, dust, CO2, and dioxin precursors are measured; flue gas flow rate and air flow rate of each air chamber are obtained through flow sensors; the moisture content of waste and sludge is obtained through moisture content detection devices; and the pusher stroke, grate running speed, primary fan frequency, secondary fan frequency, and hourly sludge feed rate are read through the distributed control system.

[0032] The location and function of each sensor are as follows.

[0033] LiDAR: Deployed at a suitable location above the waste storage area to collect data on the distribution of waste levels, and to calculate the weight of waste in the waste storage area and the hourly feed rate of waste into the furnace based on changes in the waste level.

[0034] Flame vision camera: placed at the fire observation hole or observation port in the incinerator furnace to collect real-time image data of flame shape, brightness and combustion stability in the furnace, and extract the average flame brightness and flame area through image analysis.

[0035] Temperature sensors include furnace thermocouples, infrared thermometers, and flue gas temperature sensors. Furnace thermocouples are positioned at key cross-sections and heights in the furnace grate and flue, collecting multi-point temperature data such as grate segment temperature and furnace outlet temperature. Infrared thermometers are positioned at the furnace observation holes or observation ports above the grate, for non-contact, real-time acquisition of grate bed surface temperature and furnace flame temperature data. Flue gas temperature sensors are positioned at the economizer outlet or chimney inlet to collect exhaust gas temperature.

[0036] Pressure sensors are placed in the furnace, primary air duct, secondary air duct, waste heat boiler, and various air chambers to monitor key pressure parameters such as furnace negative pressure, primary / secondary air pressure, flue gas pressure, and pressure difference between air chambers.

[0037] Damper opening sensor: Located at the actuator of each segment damper, used to collect the current opening angle data of each damper.

[0038] Flue gas analyzer: Located at economizer outlet and chimney inlet, etc., for online monitoring of O2, CO, and NO in flue gas. x Concentrations of key components such as SO2, HCl, CO2, dioxin precursors, and soot.

[0039] Flow sensors: installed in primary air ducts, secondary air ducts, each air chamber, and flue gas related pipes to accurately measure the air supply flow and flue gas flow data of each air chamber.

[0040] Moisture content detection device: Located at the waste storage and sludge feed inlet, used for online or semi-online detection of the moisture content of the waste and sludge entering the furnace.

[0041] In addition, a two-way data channel is established with the incineration plant’s existing distributed control system through a standardized industrial communication protocol to read equipment operation values ​​such as pusher stroke, grate running speed, primary air fan frequency, secondary air fan frequency, and hourly sludge feeding rate.

[0042] The multi-source sensor monitoring data collected by each sensing device includes the following parameters.

[0043] Material attributes entering the furnace: Weight of waste in the waste storage area (tons), moisture content of waste (%) Hourly feed rate of waste into the furnace (tons / hour), hourly sludge feed rate into the furnace (tons / hour), sludge moisture content (%) Furnace combustion conditions: grate section temperature (Celsius, of which) (Indicates the serial number of each measuring point along the length of the grate) and the furnace flame temperature. (Celsius), furnace outlet temperature (degrees Celsius), furnace negative pressure (Pascal), primary wind pressure (kPa), secondary wind pressure (kPa), average flame brightness (candela / square meter), flame area (Pixel area), material layer surface temperature (Celsius), pressure difference between each air chamber (Pascal, of which) (Indicates the sequence number of the air chamber section along the length of the grate) and the air flow rate of each air chamber. (cubic meters / hour); Flue gas emission parameters: O2 concentration (%), CO concentration (mg / standard cubic meter), NO xconcentration (mg / standard cubic meter), SO2 concentration (mg / standard cubic meter), HCl concentration (mg / standard cubic meter), smoke and dust concentration (mg / standard cubic meter), flue gas flow rate (Standard cubic meters / hour), flue gas temperature (degrees Celsius), CO2 concentration (%), dioxin precursor concentration (Ng / Standard Cubic Meter), Flue Gas Pressure (kPa); Equipment operating status: Pusher stroke (mm), grate running speed (meters / minute), primary air fan frequency (Hertz), secondary fan frequency (Hertz), damper opening (%).

[0044] Step S220: Perform edge preprocessing on the collected multi-source sensor monitoring data, extract the core process features corresponding to the waste incineration conditions, and complete real-time anomaly identification and graded early warning of waste incineration conditions based on the core process features.

[0045] Edge processing layer 120 serves as the system's on-site data processing hub and interface, handling real-time on-site computation, multi-source data governance, and plant-wide system collaboration. It is the core module for achieving seamless integration of intelligent control with the existing production management system. This layer adopts an industrial-grade, highly reliable hardware architecture, adapting to the harsh operating conditions of waste incineration plants, including high temperatures, high dust levels, and strong electromagnetic interference, ensuring long-term stable operation.

[0046] At the hardware level: an industrial-grade edge computing gateway is selected, equipped with a high-performance processing chip, which has low latency, high computing power and strong anti-interference characteristics. It is directly deployed in the on-site control cabinet of the incineration plant, which avoids the problems of latency, packet loss and signal interference caused by long-distance transmission of raw sensor data from the physical level, and realizes on-site data processing in real time, which greatly alleviates the computing power pressure of the back-end intelligent decision layer 130.

[0047] System integration: Establish a two-way communication channel with the existing distributed control system of the waste incineration plant. System integration can be completed without the need to install an additional independent communication module, simplifying on-site wiring and equipment integration complexity, and improving system compatibility and ease of operation and maintenance.

[0048] Built-in functional modules: This layer has four core processing modules built-in: data preprocessing, feature extraction, anomaly detection, and DCS communication interaction. These modules work together in a closed loop.

[0049] In some optional implementations, step S220 involves edge-end preprocessing of the collected multi-source sensor monitoring data to extract core process features corresponding to the waste incineration conditions, and real-time anomaly identification and graded early warning of the waste incineration conditions based on these core process features. This includes: performing edge-end time synchronization processing, filtering and noise reduction processing, and outlier removal processing on the collected multi-source sensor monitoring data to form standardized time-series operating condition data; extracting material calorific value fluctuation characteristics, furnace temperature field distribution characteristics, flame visual morphology characteristics, flue gas pollutant dynamic characteristics, and material layer thickness uniformity characteristics as core process features from the standardized time-series operating condition data; wherein, material calorific value fluctuation characteristics include material calorific value fluctuation characteristic values; furnace temperature field distribution characteristics include the average temperature and temperature standard deviation of measuring points in different areas of the furnace; flame visual morphology characteristics include flame area ratio and relative brightness statistics; and flue gas pollutant dynamic characteristics include CO concentration change rate and NO... x The concentration is converted; the uniformity of the material layer thickness includes the uniformity characteristic value of the material layer thickness; the extracted core process characteristics are compared with the preset safe operation threshold system. When the core process characteristics are detected to exceed the safe operation threshold range or deviate from the historical stable operating condition characteristic range, the waste incineration condition is identified as an abnormal condition; according to the degree of deviation of the abnormal condition, a corresponding level of early warning signal is generated and the graded early warning is pushed out.

[0050] The data preprocessing module receives multi-source sensor monitoring data transmitted from the sensing layer 110. Taking into account the differences in signal frequency, dimension, and acquisition accuracy of different types of sensing devices, it performs targeted standardization processing such as precise alignment of data time synchronization, noise filtering optimization, and outlier removal and cleaning. It unifies the data format and acquisition timestamp to form standardized time-series operating condition data, ensuring the accuracy, consistency, and effectiveness of the input data from the source.

[0051] Based on standardized time-series operating data, the feature extraction module combines the characteristics of waste incineration operating conditions with intelligent control requirements to accurately extract the following core process features.

[0052] The calorific value fluctuation characteristics of the material are expressed by the following formula (1): , formula (1); in, Indicates the current moment. Indicates a time window. express The instantaneous composite calorific value at a given moment. express The instantaneous composite calorific value at a given moment. express The characteristic value of material calorific value fluctuation at any given time.

[0053] in, , express The hourly feed rate of waste into the furnace. This indicates the lowest heating value of the waste as determined by the most recent laboratory calibration. express The hourly feed rate of sludge into the furnace. This indicates the lowest heating value of the sludge as determined in the most recent laboratory calibration. The algorithm is similar and will not be described in detail again.

[0054] Furnace temperature field distribution characteristics: including the average temperature of measuring points in different areas of the furnace. and temperature standard deviation This can be expressed by the following formulas (2) and (3): , formula (2); , formula (3); in, This indicates the total number of thermocouple measuring points arranged along the length of the grate. express The time along the length of the grate Temperature at each zone measurement point.

[0055] Visual morphological characteristics of flames: including the proportion of flame area With relative brightness statistics This can be expressed by the following formulas (4) and (5): , formula (4); , formula (5); in, express The area of ​​the flame at any given moment. This represents the pixel area of ​​a predefined region of interest in the image. This represents the average brightness value of the flame area under the reference operating conditions.

[0056] Dynamic characteristics of flue gas pollutants, including the rate of change of CO concentration and NO x Conversion concentration This can be expressed by the following formulas (6) and (7): , formula (6); , formula (7); in, express CO concentration at time 10:00 express CO concentration at time 10:00 This indicates the baseline oxygen content stipulated by environmental emission standards; for municipal solid waste incinerators, 11% can be used. express NO at any moment x concentration.

[0057] The uniformity of the material layer thickness is represented by the following formula (8): , formula (8); in, This indicates the total number of air chamber zones along the length of the grate. express The first moment The thickness of the material layer in each zone, express The average thickness of the material layer in each zone at any given time. express The uniformity characteristic value of the material layer thickness at any given time.

[0058] in, , , This represents the calibration coefficient, obtained through on-site cold calibration tests or by fitting historical operating data, and is related to fuel bulk density and particulate properties. express The first moment Pressure difference in each air chamber express The moment is sent in The air supply flow rate of each air chamber.

[0059] The anomaly detection module incorporates a pre-set safety threshold system and an intelligent operating condition discrimination model, dynamically comparing the characteristics of each core process with the safe operating threshold range and the historical stable operating condition characteristic range in real time. The safe operating threshold range is determined by the incinerator design parameters and environmental emission standards, while the historical stable operating condition characteristic range is obtained through statistical modeling of stable operating data from at least 12 months prior to system commissioning, with the historical mean of each core process characteristic ±2 standard deviations serving as the normal fluctuation range.

[0060] This module can quickly and accurately identify various dangerous and abnormal operating conditions, such as furnace overheating, accidental flameout, sudden rises or falls in material calorific value, abnormal material layer thickness, and excessive emissions of pollutants in flue gas. It generates graded early warning signals based on the type and degree of abnormality. The early warning levels include at least the following two levels: Level 1 Warning (Alert Level): The core process characteristic parameters exceed the historical stable operating condition characteristic range but have not yet reached the boundary of the safe operation threshold range, indicating that the operating condition has an abnormal trend but is still within the safe boundary. The system maintains the intelligent decision layer 130 automatic operation and simultaneously pushes the warning information to the distributed control system operation station to remind the operators to pay attention. Level 2 Early Warning (Intervention Level): When the core process characteristic parameters reach the boundary of the safe operation threshold, it indicates that the operating condition is approaching the safety boundary. The system automatically cuts off the control of the intelligent decision layer 130 over the actuators and switches the control to the manual control mode of the distributed control system through the DCS communication interaction module, so that the operator can take full manual control to ensure the safety of the incineration system.

[0061] The DCS communication module serves as the core of data interaction between the system and the existing distributed control system. It relies on standardized industrial communication protocols to achieve stable two-way communication, undertaking the dual functions of data reception and command transmission. Simultaneously, it supports one-click switching between manual and automatic operation modes. Under intervention-level early warning conditions, it can directly switch to the centralized manual control mode of the distributed control system, retaining the traditional control fallback capability and comprehensively ensuring the system's compatibility, flexibility, and security.

[0062] Step S230: Based on the trained time-series prediction model, and according to the core process characteristics, preprocessed multi-source sensor monitoring data, and auxiliary input parameters, predict the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods.

[0063] The intelligent decision-making layer 130 is the core control layer of the system. It adopts a high-performance industrial control computer, receives standardized time-series operating condition data and core process characteristics transmitted from the edge processing layer 120, and combines them with auxiliary input parameters. Based on industrial big data and intelligent optimization algorithms, it realizes accurate judgment of incineration conditions, parameter optimization, and command generation. This layer includes three core algorithm model modules: calorific value prediction module, co-firing ratio verification module, and multi-objective optimization algorithm module.

[0064] In some optional implementations, step S230, based on the trained time-series prediction model, predicts the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods according to core process characteristics, preprocessed multi-source sensor monitoring data, and auxiliary input parameters. This includes: inputting the material calorific value fluctuation characteristics, standardized time-series operating condition data, and auxiliary input parameters into the trained time-series prediction model, and predicting the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods through the trained time-series prediction model; wherein, the auxiliary input parameters include real-time sludge blending amount, historical calorific value moving average, and seasonal operating condition quantitative indicators.

[0065] The calorific value prediction module uses a trained calorific value prediction model, which is built on a long short-term memory neural network (LSTM) to predict the overall low calorific value of the materials fed into the furnace in advance, so as to avoid combustion instability caused by sudden changes in calorific value.

[0066] In this embodiment, the calorific value prediction model adopts a two-layer LSTM structure, and the specific parameter configuration is as follows: Model Structure: The first LSTM layer has 128 units, and the second LSTM layer has 64 units, both using the Tanh activation function. A Dropout layer is introduced between the two LSTM layers and between the LSTM layer and the output layer, with a Dropout rate of 0.2. The output layer is a fully connected layer with a 2-dimensional output, displaying the average lower heating value of the waste and sludge fed into the furnace over the next 30 minutes, along with a 95% confidence interval obtained based on MC-Dropout sampling or quantile regression. This information is used to quantify prediction uncertainty and provide data support for subsequent co-firing ratio determination and multi-objective optimization.

[0067] Input features: The input time step is set to 60 (i.e., time series data from the past 60 minutes). The feature vector for each time step includes: material calorific value fluctuation characteristics, standardized time series operating condition data, and auxiliary input parameters. All input features are Z-score standardized before being fed into the network, and the standardized parameters (mean and standard deviation) are taken from the training set.

[0068] Auxiliary input parameters include real-time sludge blending amount, historical calorific value moving average, and seasonal operating condition quantitative indicators.

[0069] The real-time sludge blending amount is expressed by the following formula (9): , formula (9).

[0070] Historical heat value moving average Based on recent sets (e.g., 7 sets) of laboratory-calibrated calorific value data (including and ) Moving average calculation.

[0071] Seasonal operating condition quantitative indicators The Gregorian calendar month is used as a quantitative representation of seasonal working conditions.

[0072] Training dataset: Collect at least 12 months of continuous historical data, with a sampling period of 1 minute. The data has been cleaned as follows: Data from the start-up and shutdown phases (within 1 hour after ignition and within 1 hour before shutdown) are excluded. Remove missing segments caused by sensor malfunction or communication interruption (if the missing segment is continuous for more than 5 minutes, the entire segment will be removed). The retained valid data are randomly divided into training set, validation set and test set in a ratio of 8:1:1, ensuring that the three sets cover different seasons and load conditions.

[0073] Training process: The loss function is mean squared error (MSE), the optimizer is Adam, the initial learning rate is set to 0.001, the batch size is set to 64, and the maximum number of training epochs is set to 200. An early stopping strategy is employed: training terminates when the validation set loss no longer decreases within 20 consecutive epochs, and the model weights with the minimum validation set loss are rolled back. To prevent overfitting, in addition to Dropout, L2 regularization is used (weight decay coefficient set to 1e-5).

[0074] Online update mechanism: After the model is deployed, incremental fine-tuning is performed every 7 days. The fine-tuning is triggered by: obtaining at least 50 sets of valid laboratory calorific value calibration data within the last 7 days, and ensuring that the input features of each set of data do not significantly deviate from the training set distribution (verified by KL divergence test, with a threshold of 0.1). Fine-tuning uses a small learning rate (1e-4) and the current model weights as initial values, trains for only 5 epochs, and monitors performance on the validation set. If the validation set MSE increases by more than 3% after fine-tuning, it automatically rolls back to the previous version. All version histories are retained up to at least the most recent 3 versions to ensure rollback capability.

[0075] Step S240: Determine the maximum safe co-firing ratio of sludge based on the predicted average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, and the preset safe operation constraints.

[0076] The blending ratio determination module employs a blending ratio determination model, combining calorific value prediction results, furnace target load, and environmental emission requirements to dynamically determine the maximum safe blending ratio of sludge. This blending ratio determination model is constructed as a mechanistic model constrained by energy balance and environmental emissions, used to dynamically determine the maximum safe blending ratio of sludge, overcoming the limitations of traditional fixed-ratio blending and adapting to the blending requirements of large-scale sludge production. The system uses a binary search method to solve this blending ratio determination model in real time.

[0077] In some optional implementations, step S240 involves determining the maximum safe co-firing ratio of sludge based on the predicted average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, and preset safe operation constraints. This includes setting preset safe operation constraints such as the total energy input of the materials fed into the furnace not being lower than the target load requirement, the average furnace temperature being within the set temperature range, the concentration of pollutants in the flue gas being lower than the emission limit, and the flue gas residence time not being less than the set duration. Under the condition that all preset safe operation constraints are met, the maximum safe co-firing ratio of sludge is determined by iteratively solving using a binary search method.

[0078] The objective function of the blending ratio determination model is: max ,in, This indicates the sludge co-firing ratio being solved. This represents the hourly feed rate of sludge into the furnace. This represents the hourly feed rate of waste into the furnace. , This indicates the total amount of material fed into the furnace per hour.

[0079] The preset safe operation constraints are as follows.

[0080] Energy conservation constraint: ;in, This represents the predicted average lower heating value of the waste entering the furnace. This represents the predicted average lower heating value of the sludge fed into the furnace. This indicates the target load demand.

[0081] NO x Emission constraints: (mg / standard cubic meter); among which, Indicates corresponding to NO x Conversion concentration.

[0082] It can be estimated using the following formula (10): , formula (10); in, Indicates NO x The sensitivity coefficient of the concentration to the sludge co-firing ratio was obtained by regression analysis of historical operating data.

[0083] CO emission constraints: (mg / standard cubic meter); among which, Indicates corresponding to CO concentration.

[0084] It can be estimated using the following formula (11): , formula (11); in, The sensitivity coefficient of CO concentration to the sludge co-firing ratio is obtained through regression analysis of historical operating data.

[0085] Furnace temperature constraints: (degrees Celsius); among which, Indicates corresponding to The average temperature of measuring points in different areas of the furnace.

[0086] It can be estimated using the following formula (12): , formula (12); in, The sensitivity coefficient representing the average temperature at different measuring points in the furnace to the sludge co-firing ratio is obtained through regression analysis of historical operating data.

[0087] Flue gas residence time constraints: (seconds); where, Indicates corresponding to The residence time of the flue gas.

[0088] It can be estimated using the following formulas (13) and (14): , formula (13); , formula (14); in, Indicates the effective volume of the furnace. Indicates corresponding to flue gas flow rate express The amount of smoke at any given time The sensitivity coefficient of flue gas flow rate to the sludge co-firing ratio is obtained through regression analysis of historical operating data.

[0089] Under the premise of ensuring all the above-mentioned preset safe operation constraints, the binary search method is used to output the results. Maximum value.

[0090] During the search for candidate sludge co-firing ratios, the estimations of various parameters by formulas (10) to (14) are all based on a first-order linear approximation near the current operating point, and are only applied to candidate values. Relative to the current actual value The deviation is sufficient in accuracy as long as it does not exceed a set threshold. Therefore, the search boundary of the binary search is limited to the current doping ratio. Centered on a preset deviation threshold The range is defined by a radius, while ensuring that it is neither lower than 0% nor higher than the upper limit of the sludge co-firing ratio allowed by the incinerator design. For example, when When taking 5 percentage points, if the current blending ratio is 30%, the search range is 25% to 35%. This boundary limit ensures that the search process always stays within a locally effective linear approximation range, avoiding estimation distortion due to excessive deviation of candidate values.

[0091] In addition, each sensitivity coefficient ( , , , The initial value of the extrapolation estimate can be obtained by regression from the step test data before the system is put into operation. After the system is put into operation, it is updated again based on the new data according to the preset cycle (such as quarterly) to ensure the long-term accuracy of the extrapolation estimate.

[0092] Step S250: Construct a comprehensive objective function. Under the constraints of the maximum safe co-firing ratio of sludge and the preset safe operation constraints, use an intelligent iterative algorithm to collaboratively optimize the comprehensive objective function and solve for the optimal combination of incineration control parameters.

[0093] The multi-objective optimization algorithm module employs the particle swarm optimization (PSO) algorithm to minimize the negative value of the sludge co-firing ratio, minimize the deviation between the furnace temperature and the target set temperature, minimize the flue gas CO concentration, and minimize the flue gas NO concentration. x The concentration and minimizing the power consumption of auxiliary equipment are used as the comprehensive objective function. The optimal combination of control parameters is solved by combining real-time operating data.

[0094] In some optional implementations, step S250 involves constructing a comprehensive objective function. Under the constraints of the maximum safe sludge co-firing ratio and preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function to solve for the optimal combination of incineration control parameters. These parameters include: the determined maximum safe sludge co-firing ratio, the average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, the target load requirement, the average temperature of measuring points in different areas of the furnace, CO concentration, O2 concentration, and NO concentration. x The converted concentration, flame area ratio, relative brightness statistics, grate operating speed, hourly feed rate of waste, hourly feed rate of sludge, waste moisture content, and sludge moisture content are input into a BP neural network. The BP neural network predicts the initial values ​​of the combination of incineration control parameters. These parameters include the pusher operating frequency, primary air fan frequency, secondary air fan frequency, grate operating speed, and the opening degree of each damper. The initial value output by the BP neural network is used as the population center to initialize the search population for the particle swarm optimization algorithm. A comprehensive objective function is constructed, which includes the following optimization objectives: minimizing the negative value of the sludge co-firing ratio, minimizing the deviation between the furnace temperature and the target set temperature, minimizing the CO concentration, and minimizing the NO concentration. x The optimal incineration control parameter combination is obtained by using the concentration and minimizing the power consumption of auxiliary equipment, taking the preset safe operation constraints as the boundary of the operating conditions and the maximum safe co-firing ratio of sludge as the upper limit constraint, minimizing the comprehensive objective function value as the optimization direction, and iteratively optimizing through the particle swarm algorithm.

[0095] First, a backpropagation (BP) neural network is introduced as the initial value generator for the particle swarm optimization algorithm. The BP neural network uses a 3-layer feedforward neural network, with the input layer containing 15-dimensional parameters (including...). , , , , , , , , , , , , , , The hidden layer has 32 neurons, and the ReLU activation function is used to enhance nonlinear fitting ability. The output layer contains a 9-dimensional combination of incineration control parameters. ;in, Indicates the operating frequency of the feeder. Indicates the frequency of the primary fan. Indicates the frequency of the secondary air fan. Indicates the grate operating speed. This indicates the opening degree of each damper.

[0096] The "optimal operating condition" sample was selected from historical operating data of no less than 12 months. The selection criteria were: simultaneously meeting the following conditions: furnace temperature stable at 900~1050℃, CO concentration ≤80mg / Nm³, and NO... x A 30-minute continuous run was conducted with a concentration ≤350 mg / Nm³. Samples were uniformly drawn from this run (one sample point every 5 minutes) to obtain at least 10,000 training samples. The samples were then divided into training, validation, and test sets in a 6:2:2 ratio.

[0097] The loss function used is mean squared error (MSE), the optimizer is Adam, the initial learning rate is set to 0.01, the batch size is set to 128, and the maximum number of training epochs is set to 500. A learning rate decay strategy is adopted (decaying to 0.1 times the original value every 100 epochs). Early stopping is used on the validation set (the number of early stopping epochs is set to 50). To prevent overfitting, a Dropout layer (Dropout rate 0.1) is added after the hidden layers, and L2 regularization is used (weight decay coefficient 1e-4). After training, performance is evaluated on the test set. If the MSE on the test set exceeds 1.2 times the MSE on the training set, the hyperparameters are readjusted and training is performed.

[0098] After the system is put into operation, the prediction accuracy of the current BP model is automatically evaluated every week. If the deviation (relative error) between the model output and the actual optimal control parameters exceeds 10% for three consecutive days, online fine-tuning is triggered. During fine-tuning, samples (no less than 200 groups) that meet the "optimal operating condition" screening criteria within the last 30 days are extracted and incrementally trained using a small learning rate (1e-3), with no more than 20 epochs of training. Version management is also performed after fine-tuning, and if the accuracy of the validation set decreases after fine-tuning, it is automatically rolled back.

[0099] Within each control cycle (e.g., every minute), the BP neural network first rapidly predicts a set of optimal initial values ​​for the combination of incineration control parameters based on the current input feature vector. Particle swarm optimization algorithm with The population (population size 30 particles) is initialized within a ±20% neighborhood of the centered particle, and then iterative optimization is performed, with a maximum of 50 iterations (in actual operation, convergence is usually achieved within 20 iterations). Compared with random initialization, this strategy improves the convergence speed by about 60% (experimental statistics), ensuring real-time performance.

[0100] Secondly, a comprehensive objective function is constructed to evaluate the overall performance of each candidate combination of incineration control parameters: ,in, Indicates the combination of incineration control parameters The proportion of sludge co-firing under different conditions. Indicates the combination of incineration control parameters Furnace temperature under certain conditions Indicates the combination of incineration control parameters CO concentration under the following conditions Indicates the combination of incineration control parameters NO under certain circumstances x Conversion concentration, Indicates the combination of incineration control parameters The power consumption of auxiliary equipment under certain conditions. Since the sub-objectives have different dimensions and distinct physical meanings, direct comparison and weighting are not possible. Therefore, this embodiment performs Min-Max normalization on each sub-objective, mapping them uniformly to the [0,1] interval to eliminate the influence of dimensions. After normalization, a single scalar fitness function is constructed using a linear weighted summation method. The particle swarm optimization algorithm iterates in the direction of minimizing this single scalar fitness function.

[0101] Finally, the particle swarm optimization algorithm satisfies Based on the premise of not exceeding the maximum safe co-firing ratio of sludge and the preset safe operation constraints, the goal is to minimize the comprehensive objective function value. To find the optimal direction, iterative optimization is performed to solve for the optimal combination of incineration control parameters. .

[0102] Step S260: Based on the optimal combination of incineration control parameters, drive each actuator on site to complete the adjustment of the waste incineration operating conditions.

[0103] The execution control layer 140 is the terminal execution layer of the system. It directly receives the optimal combination of combustion control parameters issued by the intelligent decision layer 130 and transmits it to the field distributed control system. This enables precise control of the entire process of incinerator feeding, air distribution, and grate operation, ensuring that intelligent decision commands are implemented, responding quickly to changes in operating conditions, and maintaining stable combustion in the furnace.

[0104] In some optional implementations, step S260, based on the optimal combination of incineration control parameters, drives each on-site actuator to complete the adjustment of the waste incineration operating conditions, including: sending the solved optimal combination of incineration control parameters to the distributed control system, and driving each on-site actuator to perform the adjustment of the incineration operating conditions through the distributed control system; wherein, the on-site actuators include a pusher, a primary air fan, a secondary air fan, a damper actuator, and a grate drive mechanism; the process of adjusting the incineration operating conditions is as follows: the pusher adjusts the feeding rate according to the optimal pusher operating frequency, the primary air fan adjusts the total combustion air volume according to the optimal primary air fan frequency, the secondary air fan adjusts the disturbance air volume in the upper part of the furnace according to the optimal secondary air fan frequency, the damper actuator independently adjusts the air intake volume of each air chamber along the length of the grate according to the optimal opening of each section of the damper, and the grate drive mechanism adjusts the grate travel speed according to the optimal grate running speed; during the adjustment of the incineration operating conditions, multi-source sensor monitoring data is collected in real time as feedback, and the operating parameters of each on-site actuator are corrected in a closed loop.

[0105] The execution control layer 140 specifically drives the following actuators through a distributed control system.

[0106] The feeder adjusts the feeding rate according to the optimal feeder operating frequency. The primary air fan adjusts the total combustion air volume according to the optimal primary air fan frequency; The secondary air fan adjusts the disturbance air volume in the upper part of the furnace according to the optimal secondary air fan frequency; The damper actuator can independently adjust the air intake of each air chamber along the length of the grate according to the optimal opening of each damper section. The grate drive mechanism adjusts the grate travel speed according to the optimal grate operating speed.

[0107] During the incineration operation adjustment process, multi-source sensor monitoring data is collected in real time as feedback to perform closed-loop correction of the operating parameters of each field actuator. Flame vision and furnace temperature are fed back in real time, and the air distribution and grate speed are corrected at set time intervals to keep the furnace temperature stable in the range of 850~1100 degrees Celsius, with fluctuations controlled within ±50℃.

[0108] The system monitors the operating conditions in real time. When abnormal trends such as weak local combustion or a slight increase in CO are detected, it automatically implements intervention measures (such as increasing the frequency of secondary air and appropriately reducing the grate speed) to restore stable combustion within a set time, preventing flameout, coking, and excessive emissions.

[0109] The following is a specific implementation method: for a mechanical grate-type municipal solid waste incinerator with a daily processing capacity of 750 tons, the system provided in this application is used to achieve safe co-firing of sludge at a large proportion, stable combustion, and pollutant emissions that meet standards.

[0110] I. Implementation conditions and equipment foundation.

[0111] Incinerator type: Mechanical grate furnace; Rated daily processing capacity: 750 tons; Rated evaporation capacity: approximately 82 tons / hour; Materials to be co-processed: municipal solid waste and sewage sludge; Controlled objects: pusher, primary air fan, secondary air fan, grate drive mechanism, damper actuator.

[0112] All of the aforementioned actuators are connected to the existing distributed control system of the incineration plant. This system communicates bidirectionally with the distributed control system through a standardized industrial communication protocol to enable the issuance of control commands and the reading of operating status.

[0113] II. Sensor Deployment and Parameter Acquisition LiDAR: Deployed above the waste storage area to collect real-time data on waste distribution and calculate the weight of waste in the waste storage area and the hourly feed rate of waste into the furnace; Flame vision camera: Installed at the fire observation hole in the furnace, with a cooling and purging device, used to acquire real-time images of the flame inside the furnace and extract the average flame brightness and flame area characteristics; Multi-point thermocouples: Arranged along the length of the grate in different sections of the furnace to collect the temperature of the grate sections and construct the temperature field distribution of the furnace. Air pressure transmitter and flow meter: installed in primary air duct, secondary air duct and each air chamber, used to collect air pressure, air volume and air chamber pressure difference in real time; Flue gas monitoring system (CEMS): Located at the economizer outlet and chimney inlet, used for online monitoring of O2, CO, and NO. x Concentrations of flue gas components such as SO2, HCl, and particulate matter.

[0114] The initial operating data are as follows: Current incinerator load: approximately 90% of rated load; The current actual sludge co-firing ratio is approximately 30%.

[0115] III. Intelligent Control Execution Steps.

[0116] Step 1: Data preprocessing and feature extraction.

[0117] The collected data is time-aligned, filtered, and outlier removed to extract key features.

[0118] Material calorific value fluctuation characteristics: Material calorific value fluctuation characteristic value +0.02MJ / kg·min ( minute); Furnace temperature field distribution characteristics: The average temperature at measuring points in different areas of the furnace is 920℃, and the standard deviation is 18℃; Flame visual morphological characteristics: Flame area accounts for 65%, and relative brightness statistic is 0.92; Dynamic characteristics of flue gas pollutants: CO concentration change rate +1.2 mg / Nm³ 3 ·min,NO x Conversion concentration 320 mg / Nm 3 ; Material layer thickness uniformity characteristic: Material layer thickness uniformity characteristic value 0.15.

[0119] Step 2: LSTM-based calorific value prediction.

[0120] Based on historical 72-hour data, the LSTM model predicts that the average lower heating value of waste entering the furnace in the next 20 minutes will be 6200 kJ / kg, and the average lower heating value of sludge entering the furnace will be 3800 kJ / kg.

[0121] Step 3: Determine the maximum safe co-firing ratio of sludge.

[0122] Maximum safe co-firing ratio of sludge: 41% (by mass).

[0123] The 750t / d grate furnace corresponds to the following input rates: municipal solid waste 21.6t / h, sludge 14.9t / h, and total fuel 36.5t / h.

[0124] Step 4: BP-PSO multi-objective coupling optimization.

[0125] 1. Feeding system: Municipal solid waste feeding rate: 21.6 t / h; Sludge feed rate: 14.9 t / h; The feeder's operating speed is 14.6 m / h. 2. Primary air system: Frequency: 35Hz; The air chamber ratio along the grate is 40%:25%:25%:10%; 3. Secondary air system: Frequency: 10Hz; 4. Grate system: Drying grate circulation interval: 70s; Combustion grate circulation interval: 45s; Burnout grate circulation interval: 45s.

[0126] Step 5: Execution of feedforward-feedback closed-loop control.

[0127] 1. The command is issued to the distributed control system, and the actuators operate according to the optimal parameters; 2. Real-time feedback of flame vision and furnace temperature, with adjustments made to air distribution and grate speed at regular intervals (e.g., 60 seconds); 3. The furnace temperature is stable at 900~1020℃, with a fluctuation range of ≤±50℃.

[0128] Step 6: Identification and rapid intervention of abnormal operating conditions.

[0129] Real-time system monitoring: Localized combustion was weak, resulting in a slight increase in CO concentration.

[0130] Automatic intervention: Increase the secondary air frequency to 25Hz; The grate speed was reduced to 14.4 m / h; It can restore stable combustion within a certain period of time, without flameout, coking, or excessive emissions.

[0131] Step 7: Model Iteration and Self-Learning The actual operating data is transmitted back every hour to update the weights of the LSTM and BP neural networks, continuously improving performance.

[0132] IV. Technical effects achieved in this embodiment.

[0133] 1. Sludge stabilization and co-firing ratio: 41%, meeting the design requirement of ≥40%; 2. Furnace temperature: stable at 900~1040℃, with fluctuations ≤±50℃; 3. Flue gas emissions: CO concentration is reduced by 41% compared to traditional methods; NO... x The equivalent concentration decreased by 12% to 14%; 4. Thermal efficiency improvement: 4%; 5. Automation rate: over 98%, no manual operation required under normal working conditions.

[0134] It should be noted that the aforementioned waste incineration control system 100 with a large proportion of sludge co-firing can realize all of the above-mentioned waste incineration control methods with a large proportion of sludge co-firing, which will not be elaborated further.

[0135] Figure 3 This invention illustrates a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 3As shown, the electronic device includes a processor, internal memory, a network interface, and a non-volatile storage medium connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external devices via a network connection. When executed by the processor, the computer program implements the functions or steps of a waste incineration control method that involves co-firing a large proportion of sludge.

[0136] In one embodiment, the electronic device provided in this application includes a memory and a processor. The memory stores a database and a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of a waste incineration control method that involves co-firing a large proportion of sludge.

[0137] The above is as stated in this application. Figure 1 The method implemented by the waste incineration control system for high-proportion co-firing of sludge disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The steps of the method disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0138] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of a waste incineration control method that involves co-firing a large proportion of sludge.

[0139] It should be noted that the functions or steps that the above-mentioned electronic devices or computer-readable storage media can achieve can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0140] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchlink, DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0142] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0143] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for controlling waste incineration with a high proportion of sludge co-firing, characterized in that, include: Real-time acquisition of multi-source sensor monitoring data under waste incineration conditions; The collected multi-source sensor monitoring data is preprocessed at the edge end to extract the core process features corresponding to the waste incineration conditions, and real-time anomaly identification and graded early warning of waste incineration conditions are completed based on the core process features. Based on the trained time-series prediction model, the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace are predicted in future periods according to the core process characteristics, pre-processed multi-source sensor monitoring data and auxiliary input parameters. Based on the predicted average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, and the preset safe operation constraints, the maximum safe co-firing ratio of sludge is determined. A comprehensive objective function is constructed. Under the constraints of the maximum safe co-firing ratio of sludge and the preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function and solve for the optimal combination of incineration control parameters. Based on the optimal combination of incineration control parameters, drive each actuator on site to complete the adjustment of waste incineration conditions; The process involves edge-end preprocessing of the collected multi-source sensor monitoring data to extract core process features corresponding to the waste incineration operation, and real-time anomaly identification and graded early warning based on these core process features. This includes: The collected multi-source sensor monitoring data is processed by edge-end time synchronization, filtering and noise reduction, and outlier removal to form standardized time-series operating condition data. Core process features were extracted from standardized time-series operating data, including material calorific value fluctuation characteristics, furnace temperature field distribution characteristics, flame visual morphology characteristics, flue gas pollutant dynamic characteristics, and material bed thickness uniformity characteristics. Specifically, material calorific value fluctuation characteristics include the material calorific value fluctuation feature value; furnace temperature field distribution characteristics include the mean temperature and temperature standard deviation of measuring points in different areas of the furnace; flame visual morphology characteristics include the flame area ratio and relative brightness statistics; and flue gas pollutant dynamic characteristics include the CO concentration change rate and NO... x Conversion concentration; uniformity characteristics of material layer thickness, including uniformity characteristic values ​​of material layer thickness; The extracted core process features are compared with the preset safe operation threshold system. When the core process features are detected to exceed the safe operation threshold range or deviate from the historical stable operating condition feature range, the waste incineration operating condition is identified as an abnormal operating condition. Based on the degree of deviation from the abnormal operating conditions, corresponding warning signals are generated and graded warnings are pushed out. The pre-trained time-series prediction model, based on core process characteristics, pre-processed multi-source sensor monitoring data, and auxiliary input parameters, predicts the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods, including: The material calorific value fluctuation characteristics, standardized time-series operating condition data, and auxiliary input parameters are input into the trained time-series prediction model. The trained time-series prediction model is used to predict the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods. Among them, the auxiliary input parameters include real-time sludge blending amount, historical calorific value moving average, and seasonal operating condition quantitative indicators.

2. The waste incineration control method with a large proportion of sludge co-firing as described in claim 1, characterized in that, The real-time acquisition of multi-source sensor monitoring data under waste incineration conditions includes: Data on the distribution of waste material is collected by lidar to determine the weight of waste in the waste storage and the hourly feed rate of waste into the furnace. Flame images inside the furnace are captured using a flame vision camera to determine the average flame brightness and flame area; Temperature sensors are used to obtain the grate section temperature, furnace flame temperature, furnace outlet temperature, flue gas temperature, and material bed surface temperature. The furnace negative pressure, primary air pressure, secondary air pressure, flue gas pressure, and pressure difference between each air chamber are obtained through pressure sensors. The opening degree of each section of the damper is obtained by the damper opening degree sensor; The concentrations of O2, CO, and NO in the flue gas are obtained using a flue gas analyzer. x Concentrations of SO2, HCl, particulate matter, CO2, and dioxin precursors; The flue gas flow rate and the air supply flow rate of each air chamber are obtained through flow sensors; The moisture content of waste and sludge is obtained using a moisture content detection device; The distributed control system reads the pusher stroke, grate speed, primary fan frequency, secondary fan frequency, and hourly sludge feed rate into the furnace.

3. The waste incineration control method with a large proportion of sludge co-firing as described in claim 2, characterized in that, The determination of the maximum safe co-firing ratio of sludge based on the predicted average lower calorific value of the waste fed into the furnace, the average lower calorific value of the sludge fed into the furnace, and preset safe operation constraints includes: The preset safe operation constraints are: the total energy input of the material entering the furnace is not lower than the target load requirement, the average temperature of the furnace is within the set temperature range, the concentration of pollutants in the flue gas is lower than the emission limit, and the residence time of the flue gas is not less than the set duration. Under the condition that all preset safe operation constraints are met, the maximum safe sludge co-firing ratio is determined by iterative solution using the bisection search method.

4. The waste incineration control method with a large proportion of sludge co-firing as described in claim 3, characterized in that, The construction of the comprehensive objective function, under the constraints of the maximum safe co-firing ratio of sludge and preset safe operation constraints, employs an intelligent iterative algorithm to collaboratively optimize the comprehensive objective function and solve for the optimal combination of incineration control parameters, including: The determined maximum safe co-firing ratio of sludge, the average lower heating value of the waste fed into the furnace, the average lower heating value of the sludge fed into the furnace, the target load requirement, the average temperature of measuring points in different areas of the furnace, CO concentration, O2 concentration, and NO concentration will be used as parameters. x The converted concentration, flame area ratio, relative brightness statistics, grate operating speed, hourly feed rate of waste, hourly feed rate of sludge, waste moisture content, and sludge moisture content are input into a BP neural network. The BP neural network then predicts the initial values ​​of the combination of incineration control parameters. The combination of incineration control parameters includes the pusher operating frequency, primary air fan frequency, secondary air fan frequency, grate operating speed, and the opening degree of each damper. The initial value of the BP neural network output is used to initialize the search population of the particle swarm optimization algorithm. Construct a comprehensive objective function, which includes the following optimization objectives: minimizing the negative value of the sludge co-firing ratio, minimizing the deviation between the furnace temperature and the target set temperature, minimizing the CO concentration, and minimizing NO. x Concentration and minimizing auxiliary equipment power consumption; Using preset safe operation constraints as the working condition boundary and the maximum safe sludge co-firing ratio as the upper limit constraint, the optimization direction is to minimize the comprehensive objective function value. The particle swarm optimization algorithm is used to iteratively optimize and output the optimal combination of incineration control parameters.

5. The waste incineration control method with a large proportion of sludge co-firing as described in claim 1, characterized in that, The process of adjusting the waste incineration operating conditions by driving various actuators on-site based on the optimal combination of incineration control parameters includes: The optimal combination of combustion control parameters obtained from the solution is sent to the distributed control system, which drives the various actuators on site to adjust the combustion conditions. The on-site actuators include the pusher, primary air fan, secondary air fan, damper actuator, and grate drive mechanism. The process of adjusting the incineration conditions is as follows: the feeder adjusts the feeding rate according to the optimal feeder operating frequency; the primary air fan adjusts the total combustion air volume according to the optimal primary air fan frequency; the secondary air fan adjusts the disturbance air volume in the upper part of the furnace according to the optimal secondary air fan frequency; the damper actuator independently adjusts the air volume of each air chamber along the length of the grate according to the optimal opening of each damper; and the grate drive mechanism adjusts the grate travel speed according to the optimal grate running speed. During the incineration operation adjustment process, multi-source sensor monitoring data is collected in real time as feedback to perform closed-loop correction of the operating parameters of each on-site actuator.

6. A waste incineration control system for high-proportion co-firing of sludge, characterized in that, include: The sensing layer is used to collect multi-source sensor monitoring data in real time under waste incineration conditions. The edge processing layer is used to preprocess the collected multi-source sensor monitoring data at the edge, extract the core process features corresponding to the waste incineration conditions, and complete the real-time anomaly identification and graded early warning of the waste incineration conditions based on the core process features. The intelligent decision-making layer is used to predict the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods based on the trained time series prediction model, core process characteristics, pre-processed multi-source sensor monitoring data and auxiliary input parameters; and to determine the maximum safe co-firing ratio of sludge based on the predicted average lower heating value of waste entering the furnace, the average lower heating value of sludge entering the furnace and preset safe operation constraints. A comprehensive objective function is constructed. Under the constraints of the maximum safe co-firing ratio of sludge and the preset safe operation constraints, an intelligent iterative algorithm is used to collaboratively optimize the comprehensive objective function and solve for the optimal combination of incineration control parameters. The execution control layer is used to drive the various actuators on site to adjust the waste incineration conditions based on the optimal combination of incineration control parameters. The edge processing layer is used to perform time synchronization processing, filtering and noise reduction processing, and outlier removal processing on the collected multi-source sensor monitoring data at the edge end, forming standardized time-series operating condition data. From the standardized time-series operating condition data, the core process features are extracted as follows: material calorific value fluctuation characteristics, furnace temperature field distribution characteristics, flame visual morphology characteristics, flue gas pollutant dynamic characteristics, and material layer thickness uniformity characteristics. Among these, the material calorific value fluctuation characteristics include the material calorific value fluctuation feature value; the furnace temperature field distribution characteristics include the temperature mean and temperature standard deviation of measuring points in different areas of the furnace; the flame visual morphology characteristics include the flame area ratio and relative brightness statistics; and the flue gas pollutant dynamic characteristics include the CO concentration change rate and NO... x Conversion concentration; uniformity characteristics of material layer thickness include uniformity characteristic values ​​of material layer thickness; compare the extracted core process characteristics with the preset safe operation threshold system, and when the core process characteristics are detected to exceed the safe operation threshold range or deviate from the historical stable operating condition characteristic range, the waste incineration operating condition is identified as an abnormal operating condition; generate corresponding level early warning signals according to the degree of deviation of the abnormal operating condition and complete the graded early warning push. The intelligent decision-making layer is used to input the material calorific value fluctuation characteristics, standardized time-series operating condition data, and auxiliary input parameters into the trained time-series prediction model. The trained time-series prediction model is used to predict the average lower heating value of waste entering the furnace and the average lower heating value of sludge entering the furnace in future periods. The auxiliary input parameters include real-time sludge blending amount, historical calorific value moving average, and seasonal operating condition quantitative indicators.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the waste incineration control method with a high proportion of sludge co-firing as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the waste incineration control method with a high proportion of sludge co-firing as described in any one of claims 1 to 5.

Images