A multi-purpose modular operation optimization system for municipal solid waste incineration process

Abstract

The application provides a kind of urban solid waste incineration process multipurpose modular operation optimization system, comprising: multi-modal historical data synchronous drive module, MSWI process virtual control object module, MSWI process loop control module, MSWI process monitoring module, operation parameter auxiliary decision module, data acquisition forward isolation module, operation parameter reverse transmission module, MSWI process single target / multi-target operation optimization module, difficult-to-measure process parameter soft measurement module, multi-modal data-driven process parameter prediction module, visually driven combustion state identification module, flame combustion line quantification module and multi-modal data acquisition module.The urban solid waste incineration process multipurpose modular operation optimization system provided by the application can realize laboratory operation optimization algorithm verification, laboratory process parameter modeling algorithm simulation real-time verification, industrial field data acquisition and process parameter modeling, industrial field auxiliary decision and operation optimization and other functions, and is convenient to use.

Description

Technical Field

[0001] This invention relates to the field of urban solid waste incineration technology, and in particular to a multi-purpose modular operation optimization system for urban solid waste incineration processes. Background Technology

[0002] MSW incineration (MSWI) technology, characterized by harmlessness, volume reduction, and resource recovery, is a typical process industry process. While maintaining its own energy needs, it can provide various forms of energy, including electricity and heat, while also posing a relatively low risk of environmental pollution emissions. Studies have shown that MSWI can achieve reduction rates of 70%, 90%, and energy recovery rates of 19%, respectively, and its potential economic and environmental value has been recognized by developing countries.

[0003] MSWI processes play a crucial role in low-carbon, environmental protection, and sustainable energy. Currently, domestic incineration plants mainly employ a manual control model where domain experts (knowledge workers) manually implement operational rules based on their understanding of mechanistic and experiential changes in operating conditions to meet diverse scenario requirements. This model, while possessing intelligent autonomous behavior, suffers from limitations in expert capacity, experience disparities, and control subjectivity, making it difficult to guarantee the stability of continuous operation and impacting the pollution and carbon reduction of MSWI power plants. Therefore, it is essential to independently develop intelligent operation optimization technologies tailored to the characteristics of MSWI processes in my country. The safety considerations of MSWI power plants and the closed nature of their distributed control systems (DCS) make interaction with external algorithms difficult; non-internal systems cannot directly acquire data or write parameters to the existing MSWI control system. These limitations make it difficult to achieve online verification of modeling, control, and optimization algorithms researched in the laboratory for MSWI processes. Therefore, an operation optimization algorithm verification system is an indispensable tool for the practical application of relevant theoretical and technological research conducted in the laboratory. Furthermore, a primary condition for algorithm development in the laboratory is the ability to acquire multimodal data from the industrial site in real time, including video and process data. It is also necessary to consider how the operational optimization algorithm, after offline validation, can gradually gain acceptance from experts in the industrial field for practical application. Simultaneously, the applicability and portability of the laboratory hardware and software for the operational optimization algorithm in the industrial field must be considered. Therefore, designing a multi-purpose modular operational optimization system for urban solid waste incineration processes is essential. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-purpose modular operation optimization system for urban solid waste incineration processes, which can realize functions such as laboratory operation optimization algorithm verification, laboratory process parameter modeling algorithm real-time verification, industrial field data acquisition and process parameter modeling, and industrial field auxiliary decision-making and operation optimization, and is easy to use.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A multi-purpose modular operation optimization system for urban solid waste incineration processes includes: a multimodal historical data synchronization and driving module, a MSWI process virtual control object module, an MSWI process loop control module, an MSWI process monitoring module, an operation parameter auxiliary decision-making module, a data acquisition forward isolation module, an operation parameter reverse transmission module, an MSWI process single-objective / multi-objective operation optimization module, a difficult-to-measure process parameter soft measurement module, a multimodal data-driven process parameter prediction module, a vision-driven combustion state recognition module, a flame combustion line quantification module, and a multimodal data acquisition module. The multimodal historical data synchronization and driving module is connected to the multimodal data acquisition module, and the multimodal data acquisition module is connected to the flame combustion line quantification module, the vision-driven combustion state recognition module, the multimodal data-driven process parameter prediction module, the difficult-to-measure process parameter soft measurement module, and the MSWI process single-objective / multi-objective operation optimization module. The flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, and MSWI process single-objective / multi-objective operation optimization module are connected to the operation parameter auxiliary decision-making module and the operation parameter reverse transmission module. The operation parameter reverse transmission module is connected to the operation parameter auxiliary decision-making module. The operation parameter auxiliary decision-making module is connected to the MSWI process monitoring module and the field process monitoring system. The field process monitoring system is connected to the data acquisition forward isolation module. The data acquisition forward isolation module is connected to the multimodal data acquisition module. The MSWI process monitoring module is connected to the MSWI process loop control module. The MSWI process loop control module is connected to the MSWI process virtual control object module. The field process monitoring system is connected to the field loop control system. The field loop control system is connected to the actuators and instruments.

[0007] The multimodal historical data synchronization driver module is used to provide multimodal data sources for the MSWI process;

[0008] The MSWI process virtual control object module is used to simulate the MSWI process built in the laboratory;

[0009] The MSWI process loop control module is used to realize loop control of the virtual MSWI process;

[0010] The MSWI process monitoring module is used to monitor the virtual MSWI process;

[0011] The operating parameter auxiliary decision-making module is used to acquire operating parameters and perform comparative analysis and decision-making.

[0012] The data acquisition forward isolation module is used to acquire data from all process variables in the monitoring module of the virtual MSWI process through physical isolation.

[0013] The reverse transmission module for operating parameters is used to transmit the optimized values ​​of operating parameters obtained from the MSWI process single-objective / multi-objective operation optimization module, as well as the operating parameters from the flame combustion line quantification module, the vision-driven combustion state recognition module, the multimodal data-driven process parameter prediction module, and the operating parameter detection values ​​from the difficult-to-measure process parameter soft measurement module in a physically isolated manner.

[0014] The MSWI process single-objective / multi-objective operation optimization module is used to optimize the MSWI process operation parameters based on multimodal data and a soft measurement model for difficult-to-measure parameters;

[0015] The soft measurement module for difficult-to-detect process parameters is used to perform soft measurement modeling for parameters that are difficult to detect based on multimodal data and production reports.

[0016] The multimodal data-driven process parameter prediction module is used to realize single-step / multi-step prediction of process parameters such as furnace temperature, flue gas oxygen content, and steam flow rate based on multimodal data.

[0017] The vision-driven combustion state recognition module is used to recognize the combustion state in the furnace using a domain expert-like recognition mechanism for the MSWI process.

[0018] The flame combustion line quantization module is used to quantify the flame combustion line using a domain expert identification mechanism for the MSWI process.

[0019] The multimodal data acquisition module is used to acquire simulated and actual MSWI processes with multimodal data, including left grate flame video, right grate flame video, and historical process data, as well as process and input various production reports related to product quality, environmental indicators, and economic indicators generated in the actual process.

[0020] Optionally, the multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter auxiliary decision-making module, MSWI process monitoring module, MSWI process loop control module, and MSWI process virtual control object module together constitute a laboratory operation optimization algorithm verification subsystem. The multimodal data acquisition module acquires synchronized left and right grate flame image data and historical process data from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, and multimodal data-driven process parameter prediction module... The difficult-to-measure process parameter soft measurement module and its processing module obtain various types of operating parameters. These operating parameters are transmitted to the MSWI process monitoring module via OPC and then downloaded to the MSWI process loop control module based on a real PLC / DCS device. The control quantity is then transmitted as an analog output to the virtual actuator of the MSWI process virtual control object module. The output of the virtual actuator acts on the virtual object to generate the controlled variable output, which is transmitted as an analog input to the MSWI process loop control module via a virtual instrument device. It is then transmitted to the MSWI process monitoring module, and finally to the multimodal data acquisition module via OPC. Finally, it is fed back to the difficult-to-measure process parameter soft measurement module, completing the verification of the operation optimization algorithm for laboratory use.

[0021] Optionally, the multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together constitute a laboratory process parameter modeling algorithm simulated real-time verification subsystem. The multimodal data acquisition module acquires synchronized left and right grate flame image data and historical process data from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module realize the process parameter prediction, combustion state recognition, combustion line quantization, and process parameter soft measurement results simulated in real-time multimodal data synchronization and release in an industrial setting.

[0022] Optionally, the data forward acquisition isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together form an industrial field data acquisition and process parameter modeling subsystem. The industrial field data is transmitted to the forward server via OPC, and then transmitted to the multimodal data acquisition module after physical isolation forward acquisition. The flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module realize the process parameter prediction, combustion state recognition, combustion line quantification, and process parameter soft measurement results based on the real-time multimodal data of the industrial field.

[0023] Optionally, the data forward acquisition isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter reverse transmission module, and operation parameter auxiliary decision-making module together constitute an industrial field auxiliary decision-making and operation optimization subsystem. Process data is transmitted to the forward server of the data acquisition forward isolation module via OPC. After physical isolation and forward acquisition, it is transmitted to the multimodal data acquisition module. After processing by the flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, and MSWI process single-objective / multi-objective operation optimization module, various types of key operation parameter predicted values ​​and soft measurement values, as well as optimized operation parameter values, are obtained. After being transmitted to the reverse server of the operating parameter reverse transmission module, the parameters are physically isolated and then transmitted back to the operating parameter reverse receiving server. After auxiliary decision analysis by the operating parameter auxiliary decision module, the optimized operating parameters are transmitted to the field monitoring system and PLC / DCS system according to the safety requirements of the MSWI plant using OPC protocol or OCR recognition. The control quantity is then transmitted to the actual MSWI process actuator in the form of analog output, acting on the actual object consisting of solid waste storage and transportation, solid waste combustion, waste heat exchange, flue gas purification and flue gas emission stages. The analog input is then collected by the instrument device and sent to the PLC / DCS system and the field monitoring system. It is then transmitted to the data acquisition forward isolation module and the multimodal data acquisition module via OPC, and then fed back to the MSWI process single-objective / multi-objective operation optimization module to complete the implementation of the operation optimization algorithm for the actual industrial site.

[0024] According to specific embodiments provided by the present invention, the following technical effects are disclosed: The multi-purpose modular operation optimization system for urban solid waste incineration processes provided by the present invention includes a multi-modal historical data synchronization driving module, a MSWI process virtual control object module, an MSWI process loop control module, an MSWI process monitoring module, an operation parameter auxiliary decision-making module, a data acquisition forward isolation module, an operation parameter reverse transmission module, an MSWI process single-objective / multi-objective operation optimization module, a difficult-to-measure process parameter soft measurement module, a multi-modal data-driven process parameter prediction module, a vision-driven combustion state recognition module, a flame combustion line quantification module, and a multi-modal data acquisition module. This system can be simultaneously applied to real-world applications... This system, designed for both laboratories and industrial sites, avoids the drawbacks of traditional simulation systems that are either overly complex or overly simplistic. It can be modularly built according to specific needs and effectively meets the requirements for isolated data acquisition and software porting in industrial settings. The system can effectively integrate multimodal data synchronous prediction and operational optimization control. It takes into account the security requirements of different MSWI enterprises and allows for the selection of OPC servers or OCR recognition methods for transmitting operating parameters. It has strong practicality and adaptability. The system can realize functions such as laboratory operation optimization algorithm verification, laboratory process parameter modeling algorithm simulation real-time verification, industrial site data acquisition and process parameter modeling, and industrial site auxiliary decision-making and operation optimization, making it easy to use. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A process flow diagram of urban solid waste incineration;

[0027] Figure 2 This is a schematic diagram of the multi-purpose modular operation optimization system for urban solid waste incineration process according to an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] The purpose of this invention is to provide a multi-purpose modular operation optimization system for urban solid waste incineration processes, which can realize functions such as laboratory operation optimization algorithm verification, laboratory process parameter modeling algorithm real-time verification, industrial field data acquisition and process parameter modeling, and industrial field auxiliary decision-making and operation optimization, and is easy to use.

[0030] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0031] The process flow of urban solid waste incineration is as follows: Figure 1 As shown.

[0032] like Figure 2 As shown in the embodiment of the present invention, the multi-purpose modular operation optimization system for urban solid waste incineration processes includes: a multimodal historical data synchronization and driving module, an MSWI process virtual control object module, an MSWI process loop control module, an MSWI process monitoring module, an operation parameter auxiliary decision-making module, a data acquisition forward isolation module, an operation parameter reverse transmission module, an MSWI process single-objective / multi-objective operation optimization module, a difficult-to-measure process parameter soft measurement module, a multimodal data-driven process parameter prediction module, a vision-driven combustion state recognition module, a flame combustion line quantification module, and a multimodal data acquisition module. The multimodal historical data synchronization and driving module is connected to the multimodal data acquisition module, and the multimodal data acquisition module is connected to the flame combustion line quantification module, the vision-driven combustion state recognition module, the multimodal data-driven process parameter prediction module, the difficult-to-measure process parameter soft measurement module, and the MSWI process single-objective / multi-objective operation optimization module. The module comprises a flame combustion line quantification module, a vision-driven combustion state recognition module, a multimodal data-driven process parameter prediction module, a difficult-to-measure process parameter soft measurement module, and an MSWI process single-objective / multi-objective operation optimization module. These modules are connected to the operation parameter auxiliary decision-making module and the operation parameter reverse transmission module. The operation parameter reverse transmission module is connected to the operation parameter auxiliary decision-making module. The operation parameter auxiliary decision-making module is connected to the MSWI process monitoring module and the field process monitoring system. The field process monitoring system is connected to the data acquisition forward isolation module. The data acquisition forward isolation module is connected to the multimodal data acquisition module. The MSWI process monitoring module is connected to the MSWI process loop control module. The MSWI process loop control module is connected to the MSWI process virtual control object module. The field process monitoring system is connected to the field loop control system. The field loop control system is connected to the actuators and instruments.

[0033] exist Figure 2In the diagram, dashed lines connecting the modules represent connection methods unique to the laboratory operation optimization algorithm verification subsystem, solid lines represent connection methods common to all four subsystems, and thick boxes represent the composition and related systems of the actual MSWI process. The operating parameter auxiliary decision-making module exhibits differences in laboratory and industrial field applications. Furthermore, based on... Figure 2 The 13 modules in this invention can implement subsystems including, but not limited to, the four subsystems mentioned above. The functions of the 13 modules are described below:

[0034] The multimodal historical data synchronization driver module: This module is designed for the simulated MSWI process with multimodal data built in the laboratory. It enables the synchronous release of historical left grate flame video, right grate flame video and historical process data, that is, to provide a multimodal data source for the MSWI process in the laboratory setting.

[0035] The MSWI process virtual control object module: This module is for MSWI processes with multimodal data that are simulated in the laboratory. It realizes the simulation of MSWI processes that are difficult to build in the laboratory by constructing models in the virtual actuator computer, virtual object computer and virtual instrumentation computer.

[0036] The MSWI process loop control module: This module realizes loop control of the virtual MSWI process;

[0037] The MSWI process monitoring module: This module monitors the virtual MSWI process;

[0038] The operating parameter auxiliary decision-making module: After obtaining relevant operating parameters through the operating parameter reverse receiving server, this module performs comparative analysis and decision-making on the operating parameters, and then directly transmits these operating parameters to the M9 module or the on-site process monitoring system, or transmits the operating parameters to the on-site process monitoring system through the operating parameter OCR recognition method.

[0039] The data acquisition forward isolation module: This module enables the acquisition of all process variables in the MSWI process monitoring module through physical isolation, thereby avoiding any impact on the original control system of the MSWI process;

[0040] The reverse transmission module for operating parameters: This module enables the reverse transmission of optimized operating parameter values ​​obtained from the MSWI process single-objective / multi-objective operation optimization module, as well as operating parameters from the flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and operating parameter detection values ​​from the M6-difficult-to-measure process parameter soft measurement module in a physically isolated manner, so as to avoid affecting the original control system of the MSWI process.

[0041] The MSWI process single-objective / multi-objective operation optimization module: This module optimizes the MSWI process operation parameters based on multimodal data and soft measurement models of difficult-to-measure parameters. It mainly obtains the optimized operation parameter values ​​required by the M10 module and the on-site process monitoring system by optimizing the quality indicators and environmental protection indicators of the MSWI process.

[0042] The soft measurement module for difficult-to-detect process parameters: This module realizes soft measurement modeling of difficult-to-detect parameters such as slag loss on ignition (a product quality parameter) and dioxin (an environmental indicator parameter) based on multimodal data and production reports, providing support for single-objective / multi-objective operation optimization of MSWI process;

[0043] The multimodal data-driven process parameter prediction module: This module realizes single-step / multi-step prediction of process parameters such as furnace temperature, flue gas oxygen content, and steam flow rate based on multimodal data, providing support for loop control of the MSWI process;

[0044] The vision-driven combustion state recognition module: This module realizes the recognition of the combustion state in the furnace through a domain expert-like recognition mechanism for the MSWI process, providing support for the loop control of the MSWI process;

[0045] The flame combustion line quantization module: This module realizes the quantization of the flame combustion line for the MSWI process using a domain expert identification mechanism, providing support for the loop control of the MSWI process;

[0046] The multimodal data acquisition module: This module realizes the acquisition of left grate flame video, right grate flame video and historical process data of the simulated and actual MSWI process with multimodal data, as well as the processing and input of various production reports related to product quality, environmental protection indicators and economic indicators generated in the actual process, providing data support for the M3-M7 modules.

[0047] This invention is based on 13 modules and has 4 subsystems, which will be described in detail below;

[0048] The multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter auxiliary decision-making module, MSWI process monitoring module, MSWI process loop control module, and MSWI process virtual control object module together constitute the laboratory operation optimization algorithm verification subsystem. The multimodal data acquisition module acquires synchronized left and right grate flame image data and historical process data from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, and difficult-to-measure process parameter virtual control object module together form the laboratory operation optimization algorithm verification subsystem. The soft measurement module processes and obtains various types of operating parameters. These parameters are then transmitted to the MSWI process monitoring module via OPC and downloaded to the MSWI process loop control module, which is based on a real PLC / DCS device. The control input is then transmitted as an analog output to the virtual actuator of the MSWI process virtual control object module. The output of the virtual actuator acts on the virtual object to generate the controlled variable output, which is transmitted as an analog input to the MSWI process loop control module via a virtual instrument. This input is then transmitted to the MSWI process monitoring module, and finally to the multimodal data acquisition module via OPC. Finally, the data is fed back to the soft measurement module for difficult-to-measure process parameters, completing the laboratory-oriented operation optimization algorithm verification. The functions of each module in the laboratory operation optimization algorithm verification subsystem are described below:

[0049] Multimodal historical data synchronization driver module: To reflect the dynamic characteristics of the MSWI process and the actual industrial conditions, the collected historical process data and flame video data are synchronously and real-time published. It consists of four components: a network time server, historical right grate flame video publication, historical left grate flame video publication, and historical process data publication. The specific implementation steps are as follows:

[0050] (1) Store the historical process data of the actual MSWI plant in the form of files in the historical process data OPC server, and realize the data publication within the local area network through the OPC protocol;

[0051] (2) Based on the actual MSWI plant, the incinerator flame monitoring is divided into left and right sides, which are played in real time on two image simulation computers respectively.

[0052] (3) Connect the historical data OPC server, the left grate combustion image simulator, the right grate combustion image simulator, and the network time server to the local area network through a switch;

[0053] (4) By using a network time server, the time of process data and image / video is precisely controlled to be at the same moment, thereby enabling the simultaneous display of multimodal information.

[0054] Multimodal data acquisition module: Incinerator flame and process data are important bases for domain experts to judge the stability of MSWI processes. Real-time acquisition and preprocessing of the above multimodal data are particularly important. The specific steps of multimodal data acquisition include:

[0055] (1) Use two identical sets of camera equipment to collect real-time online video of the flames on the left and right grates respectively;

[0056] (2) Transmit the information captured by the camera to the video capture card via a coaxial cable;

[0057] (3) Install the video capture card into the multimodal data acquisition computer and use video decoding to achieve a preliminary display of flame information;

[0058] (4) The acquired flame video is preprocessed by image processing algorithms, including denoising and enhancement.

[0059] (5) The process data on the OPC server is acquired and transferred to the multimodal data acquisition computer via industrial Ethernet using an OPC client.

[0060] (6) Production reports are collected into the multimodal data acquisition computer by means of manual input by domain experts, and the sampling period is set reasonably to achieve synchronous storage after multimodal data acquisition.

[0061] Flame combustion line quantization module: The flame combustion line is the controlled variable required to achieve stable control of the MSWI process. The specific steps for quantization include:

[0062] (1) Using image processing technology, conditional generative adversarial network and cycle consistency generative adversarial network, a complete image library is constructed, which includes a normal sub-library of real combustion lines, a sub-library of real / generated abnormal combustion lines, and a sub-library of generated extreme abnormal flame images of combustion lines.

[0063] (2) Select typical images from a complete image library to form a typical template library and train a twin convolutional neural network;

[0064] (3) Extract combustion line features from new flame images and use twin convolutional neural network similarity metric to achieve adaptation with flame images in the "typical template library";

[0065] (4) For new flame images that are not adapted, the nearest neighbor criterion is used to quantize the combustion line;

[0066] (5) Based on the redundancy discrimination mechanism and combined with the experience of domain experts, the “typical template library” is updated adaptively using non-adaptive images.

[0067] Vision-driven combustion state recognition module: Flame combustion state is the controlled variable required for stable MSWI process control. The specific steps for vision-driven combustion state recognition include:

[0068] (1) The flame image is dehazed and denoised by preprocessing methods such as dehazing algorithm based on artificial multi-exposure image fusion, feature normalization, notch filtering, and median filtering, so as to obtain a clear image.

[0069] (2) Extract multiple physically meaningful features such as brightness, texture, and color from the flame image to represent the image from multiple viewpoints, and reduce these features based on mutual information;

[0070] (3) The above-mentioned reduced features are used as inputs to image classifiers such as support vector machines, deep forests, and convolutional neural networks to establish visual-driven combustion state recognition models for the left and right grates respectively.

[0071] (4) For each new flame image, the combustion state it represents is identified.

[0072] Multimodal data-driven process parameter prediction module: The ability to predict process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration in one or more steps is crucial for domain experts. The specific steps for achieving multimodal data-driven process parameter prediction include:

[0073] (1) Using the multimodal data stored in the M2 module, extract flame image features based on the M4 module;

[0074] (2) Combine flame features with process data in series to form new features for training a prediction model of key process parameters;

[0075] (3) Input the newly acquired multimodal process data from the M2 module into the above-mentioned key process parameter prediction model to obtain one-step or even multi-step prediction outputs of process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration.

[0076] The soft measurement module for difficult-to-detect process parameters enables soft measurement modeling of difficult-to-detect parameters such as slag loss on ignition (LOI) for product quality parameters and dioxins for environmental indicators. This is crucial for single-objective / multi-objective operation optimization of MSWI processes. The specific steps for predicting difficult-to-detect process parameters include:

[0077] (1) Organize the difficult-to-measure process parameter data recorded in various production reports as the output true value of the soft measurement model;

[0078] (2) Select the multimodal data stored in the M2 module for the true value of soft measurement output for difficult-to-measure process parameters, that is, obtain the multimodal data time period corresponding to the true value of soft measurement output;

[0079] (3) For the flame data within the corresponding soft measurement output true value time period, extract the flame image features according to the M4 module, and combine them with the process data in series to form new features for the soft measurement model input.

[0080] (4) Use the above-mentioned reduced features as inputs to regression models such as support vector machines, deep forests, and deep neural networks to establish soft measurement models for difficult-to-detect process parameters;

[0081] (5) Input the newly acquired multimodal process data from the M2 module into the soft measurement model of the above difficult-to-measure process parameters to obtain the soft measurement values ​​of difficult-to-measure process parameters such as slag heat loss on ignition and dioxin concentration.

[0082] MSWI process single-objective / multi-objective operation optimization module: Under different scenarios, considering single or multiple objectives such as flue gas emission indicators, economic benefits, and slag thermal reduction rate, it enables timely adjustment of the setpoints of key controlled variables such as furnace temperature, flue gas oxygen content, and steam flow rate according to the dynamic changes of the MSWI process. This is crucial for achieving optimized control of the MSWI process. The specific steps for achieving single-objective / multi-objective operation optimization of the MSWI process include:

[0083] (1) Based on MSWI process analysis for single / multi-objective optimization, a single / multi-objective optimization model was established with key controlled variables such as furnace temperature, boiler steam flow rate and flue gas oxygen content as decision variables, minimizing pollutant emission indicators and heat reduction rate, and maximizing combustion efficiency and economic indicators.

[0084] (2) Based on the above single / multi-objective optimization model, the multimodal data collected in real time by the M2 module and the soft measurement model of difficult process parameters by the M6 ​​module are used to obtain the optimized setpoints of key controlled variables by intelligent optimization algorithms such as genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm.

[0085] Operating parameter-assisted decision-making module: The single-step / multi-step predicted values ​​and soft measurement values ​​of key operating parameter models are crucial for domain experts to implement control strategies and determine whether to apply optimized operating parameters to the control of MSWI processes. The steps to implement operating parameter-assisted decision-making are as follows:

[0086] (1) Based on the M3-M6 module, the single-step / multi-step predicted values ​​and soft measurement values ​​of the key operating parameter model are collected to the operating parameter reverse receiving server;

[0087] (2) Based on the single-step / multi-step predicted values ​​and soft measurement values ​​of the key operating parameter models in different time periods, a statistical comparative analysis is conducted with the actual values ​​of these key operating parameters. The reliability of the single-step / multi-step predicted values ​​and soft measurement values ​​of these key operating parameter models is assisted by methods such as active learning by domain experts and automatic judgment by setting thresholds.

[0088] (3) Combining the reliability of the single-step / multi-step predicted values ​​and soft measurement values ​​of the above key operating parameter models with the optimized operating parameters derived from M7, the reliable operating parameters of domain expert decision-making or automatic decision-making are transmitted to the M10 module.

[0089] MSWI process monitoring module: Includes functions such as decision-making on whether to optimize operating parameters and real-time monitoring of the operation process. The latter includes interfaces for combustion process, grate operating status, boiler status, flue gas treatment, variable trend charts, and parameter settings. The specific steps are as follows:

[0090] (1) Develop a configuration monitoring system that includes interfaces for incineration process, grate operation status, boiler status, flue gas treatment, variable trend diagram and parameter setting, and perform real-time monitoring and data display of MSWI process;

[0091] (2) The process variable values ​​sent in real time from the loop control module to the OPC Server are received by the OPC Client and displayed in a graphical way on the interface of the incineration process, grate operation status, etc., so as to realize the monitoring function of the whole process and transmit them to the M2 module.

[0092] (3) Make decisions on whether to use optimized operating parameters based on production needs, production indicators and expert experience, and then set and modify control loop parameters.

[0093] (4) Download the determined control loop parameters to the M11-MSWI process loop control module.

[0094] MSWI process loop control module: Based on actual manufacturer's PLC / DCS equipment, this module constructs a logic loop control system for the MSWI process. The specific steps for implementing MSWI process loop control include:

[0095] (1) Power-on startup of hardware devices is achieved based on CPU, input, output and communication modules from PLC / DCS manufacturers;

[0096] (2) Control network communication is achieved by connecting to the M10 module via industrial Ethernet;

[0097] (3) Write the start-stop, PID, alarm and interlock control programs of the MSWI process in ladder diagram language to realize the logic loop control function;

[0098] (4) Connect to the virtual actuator computer of the M12 module through the AO / DO module, and connect to the virtual instrument device computer of the M12 module through the AI / DI module to realize data interaction with the M12 module.

[0099] MSWI Process Virtual Control Object Module: The virtual actuators and measuring instruments in this module need to exchange data effectively with the M11 module to support the latter's operation. The implementation steps of the MSWI Process Virtual Control Object are as follows:

[0100] (1) Based on the actual operating data of actuators such as dampers, fans and hydraulic drives, a data-driven virtual actuator model is constructed; based on the actual operating data of sensor devices such as temperature, flow and pressure, a data-driven dashed sensor model is constructed; similarly, a virtual object model is constructed using a data-driven approach.

[0101] (2) Based on data acquisition boards and terminal boards, the I / O modules in the PLC / DCS control system are connected to the O / I terminals on the terminal boards via twisted-pair cables, and connected to the MSWI process virtual incineration object computer via Ethernet.

[0102] (3) The DI / DO (or AI / AO) modules in the actuator model computer and the DO / DI (or AO / AI) modules in the PLC / DCS control system respectively transmit data in real time.

[0103] (4) The AI / AO in the detection instrument model computer and the AO / AI module in the PLC / DCS control system respectively transmit data in real time.

[0104] The multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together constitute the laboratory process parameter modeling algorithm simulated real-time verification subsystem. The multimodal data acquisition module acquires synchronized flame image data from the left and right grates, as well as historical process data, from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module achieve process parameter prediction, combustion state recognition, combustion line quantization, and process parameter soft measurement results, simulating real-time multimodal data synchronization and release in an industrial setting. The various modules of the laboratory process parameter modeling algorithm simulated real-time verification subsystem are described below:

[0105] Multimodal historical data synchronization driver module: Incinerator flame and process data are crucial for domain experts to assess MSWI process stability, making real-time acquisition and preprocessing of this multimodal data essential. Specific steps for multimodal data acquisition include:

[0106] (1) Use two identical sets of camera equipment to collect real-time online video of the flames on the left and right grates respectively;

[0107] (2) Transmit the information captured by the camera to the video capture card via a coaxial cable;

[0108] (3) Install the video capture card into the multimodal data acquisition computer and use video decoding to achieve a preliminary display of flame information;

[0109] (4) The acquired flame video is preprocessed by image processing algorithms, including denoising and enhancement.

[0110] (5) The process data originating from the OPC server is acquired and transmitted to the multimodal data acquisition computer via industrial Ethernet using an OPC client.

[0111] (6) Production reports are collected into the multimodal data acquisition computer by means of manual input by domain experts, and the sampling period is set reasonably to achieve synchronous storage after multimodal data acquisition.

[0112] Multimodal data acquisition module: Incinerator flame and process data are crucial for domain experts to assess MSWI process stability. Real-time acquisition and preprocessing of this multimodal data are therefore essential. Specific steps for multimodal data acquisition include:

[0113] (1) Use two identical sets of camera equipment to collect real-time online video of the flames on the left and right grates respectively;

[0114] (2) Transmit the information captured by the camera to the video capture card via a coaxial cable;

[0115] (3) Install the video capture card into the multimodal data acquisition computer and use video decoding to achieve a preliminary display of flame information;

[0116] (4) The acquired flame video is preprocessed by using image recognition software in the computer, such as noise reduction and enhancement.

[0117] (5) The process data on the OPC server is acquired and transferred to the multimodal data acquisition computer via industrial Ethernet using an OPC client.

[0118] (6) Production reports are collected into the multimodal data acquisition computer by means of manual input by domain experts, and the sampling period is set reasonably to achieve synchronous storage after multimodal data acquisition.

[0119] Flame combustion line quantization module: The flame combustion line is the controlled variable required to achieve stable control of the MSWI process. The specific steps for quantization include:

[0120] (1) Using image processing technology, conditional generative adversarial network and cycle consistency generative adversarial network, a complete image library is constructed, which includes a normal sub-library of real combustion lines, a sub-library of real / generated abnormal combustion lines, and a sub-library of generated extreme abnormal flame images of combustion lines.

[0121] (2) Select typical images from a complete image library to form a typical template library and train a twin convolutional neural network;

[0122] (3) Extract combustion line features from new flame images and use twin convolutional neural network similarity metric to achieve adaptation with flame images in the "typical template library";

[0123] (4) For new flame images that are not adapted, the nearest neighbor criterion is used to quantize the combustion line;

[0124] (5) Based on the redundancy discrimination mechanism and combined with the experience of domain experts, the “typical template library” is updated adaptively using non-adaptive images.

[0125] Vision-driven combustion state recognition module: Flame combustion state is the controlled variable required for stable MSWI process control. The specific steps for vision-driven combustion state recognition include:

[0126] (1) The flame image is dehazed and denoised by preprocessing methods such as dehazing algorithm based on artificial multi-exposure image fusion, feature normalization, notch filtering, and median filtering, so as to obtain a clear image.

[0127] (2) Extract multiple physically meaningful features such as brightness, texture, and color from the flame image to represent the image from multiple viewpoints, and reduce these features based on mutual information;

[0128] (3) The above-mentioned reduced features are used as inputs to image classifiers such as support vector machines, deep forests, and convolutional neural networks to establish visual-driven combustion state recognition models for the left and right grates respectively.

[0129] (4) For each new flame image, the combustion state it represents is identified.

[0130] Multimodal data-driven process parameter prediction module: The ability to predict process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration in one or more steps is crucial for domain experts. The specific steps for achieving multimodal data-driven process parameter prediction include:

[0131] (1) Using the multimodal data stored in the M2 module, extract flame image features based on the M4 module;

[0132] (2) Combine flame features with process data in series to form new features for training a prediction model of key process parameters;

[0133] (3) Input the newly acquired multimodal process data from the M2 module into the above-mentioned key process parameter prediction model to obtain one-step or even multi-step prediction outputs of process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration.

[0134] The soft measurement module for difficult-to-detect process parameters enables soft measurement modeling of difficult-to-detect parameters such as slag loss on ignition (LOI) for product quality parameters and dioxins for environmental indicators. This is crucial for single-objective / multi-objective operation optimization of MSWI processes. The specific steps for predicting difficult-to-detect process parameters include:

[0135] (1) Organize the difficult-to-measure process parameter data recorded in various production reports as the output true value of the soft measurement model;

[0136] (2) Select the multimodal data stored in the M2 module for the true value of soft measurement output for difficult-to-measure process parameters, that is, obtain the multimodal data time period corresponding to the true value of soft measurement output;

[0137] (3) For the flame data within the corresponding soft measurement output true value time period, extract the flame image features according to the M4 module, and combine them with the process data in series to form new features for the soft measurement model input.

[0138] (4) Use the above-mentioned reduced features as inputs to regression models such as support vector machines, deep forests, and deep neural networks to establish soft measurement models for difficult-to-detect process parameters;

[0139] (5) Input the newly acquired multimodal process data from the M2 module into the soft measurement model of the above difficult-to-measure process parameters to obtain the soft measurement values ​​of difficult-to-measure process parameters such as slag heat loss on ignition and dioxin concentration.

[0140] The industrial field data acquisition and process parameter modeling subsystem comprises the data forward acquisition isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module. It transmits industrial field data to the forward server via OPC, then through physical isolation and forward acquisition before being transmitted to the multimodal data acquisition module. The flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module enable the synchronous release of process parameter prediction, combustion state recognition, combustion line quantification, and process parameter soft measurement results based on real-time multimodal data from the industrial field. Each module of the industrial field data acquisition and process parameter modeling subsystem is described below:

[0141] Forward Data Acquisition and Isolation Module: This module acquires process data from the PLC / DCS control system network during MSWI using the OPC client provided by the industrial DCS system manufacturer. It then publishes the acquired real-time data via the OPC communication protocol through physical isolation. The implementation steps for forward data acquisition and isolation are as follows:

[0142] (1) Using the PLC / DCS manufacturer's OPC server protocol, process data is collected from the field process monitoring system to the forward server and published to the outside world in the form of OPC server;

[0143] (2) Connect the forward server and the physically isolated forward acquisition machine to the same local area network through a switch;

[0144] (3) The collected process data is transmitted to the forward data analysis server in a physically isolated forward transmission manner using a unidirectional optical fiber transmission method, including process variable grouping, naming and sampling time settings;

[0145] (4) The physical isolation forward transmission and forward data analysis server are connected to the same local area network through a switch. The forward data analysis server provides data services to the M2 module in the OPC service mode.

[0146] Multimodal data acquisition module: Incinerator flame and process data are crucial for domain experts to assess MSWI process stability. Real-time acquisition and preprocessing of this multimodal data are therefore essential. Specific steps for multimodal data acquisition include:

[0147] (1) Use two identical sets of camera equipment to collect real-time online video of the flames on the left and right grates respectively;

[0148] (2) Transmit the information captured by the camera to the video capture card via a coaxial cable;

[0149] (3) Install the video capture card into the multimodal data acquisition computer and use video decoding to achieve a preliminary display of flame information;

[0150] (4) The acquired flame video is preprocessed by using image recognition software in the computer, such as noise reduction and enhancement.

[0151] (5) The process data on the OPC server is acquired and transferred to the multimodal data acquisition computer via industrial Ethernet using an OPC client.

[0152] (6) Production reports are collected into the multimodal data acquisition computer by means of manual input by domain experts, and the sampling period is set reasonably to achieve synchronous storage after multimodal data acquisition.

[0153] Flame combustion line quantization module: The flame combustion line is the controlled variable required to achieve stable control of the MSWI process. The specific steps for quantization include:

[0154] (1) Using image processing technology, conditional generative adversarial network and cycle consistency generative adversarial network, a complete image library is constructed, which includes a normal sub-library of real combustion lines, a sub-library of real / generated abnormal combustion lines, and a sub-library of generated extreme abnormal flame images of combustion lines.

[0155] (2) Select typical images from a complete image library to form a typical template library and train a twin convolutional neural network;

[0156] (3) Extract combustion line features from new flame images and use twin convolutional neural network similarity metric to achieve adaptation with flame images in the "typical template library";

[0157] (4) For new flame images that are not adapted, the nearest neighbor criterion is used to quantize the combustion line;

[0158] (5) Based on the redundancy discrimination mechanism and combined with the experience of domain experts, the “typical template library” is updated adaptively using non-adaptive images.

[0159] Vision-driven combustion state recognition module: Flame combustion state is the controlled variable required for stable MSWI process control. The specific steps for vision-driven combustion state recognition include:

[0160] (1) The flame image is dehazed and denoised by preprocessing methods such as dehazing algorithm based on artificial multi-exposure image fusion, feature normalization, notch filtering, and median filtering, so as to obtain a clear image.

[0161] (2) Extract multiple physically meaningful features such as brightness, texture, and color from the flame image to represent the image from multiple viewpoints, and reduce these features based on mutual information;

[0162] (3) The above-mentioned reduced features are used as inputs to image classifiers such as support vector machines, deep forests, and convolutional neural networks to establish visual-driven combustion state recognition models for the left and right grates respectively.

[0163] (4) For each new flame image, the combustion state it represents is identified.

[0164] Multimodal data-driven process parameter prediction module: The ability to predict process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration in one or more steps is crucial for domain experts. The specific steps for achieving multimodal data-driven process parameter prediction include:

[0165] (1) Using the multimodal data stored in the M2 module, extract flame image features based on the M4 module;

[0166] (2) Combine flame features with process data in series to form new features for training a prediction model of key process parameters;

[0167] (3) Input the newly acquired multimodal process data from the M2 module into the above-mentioned key process parameter prediction model to obtain one-step or even multi-step prediction outputs of process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration.

[0168] The soft measurement module for difficult-to-detect process parameters enables soft measurement modeling of difficult-to-detect parameters such as slag loss on ignition (LOI) for product quality parameters and dioxins for environmental indicators. This is crucial for single-objective / multi-objective operation optimization of MSWI processes. The specific steps for predicting difficult-to-detect process parameters include:

[0169] (1) Organize the difficult-to-measure process parameter data recorded in various production reports as the output true value of the soft measurement model;

[0170] (2) Select the multimodal data stored in the M2 module for the true value of soft measurement output for difficult-to-measure process parameters, that is, obtain the multimodal data time period corresponding to the true value of soft measurement output;

[0171] (3) For the flame data within the corresponding soft measurement output true value time period, extract the flame image features according to the M4 module, and combine them with the process data in series to form new features for the soft measurement model input.

[0172] (4) Use the above-mentioned reduced features as inputs to regression models such as support vector machines, deep forests, and deep neural networks to establish soft measurement models for difficult-to-detect process parameters;

[0173] (5) Input the newly acquired multimodal process data from the M2 module into the soft measurement model of the above difficult-to-measure process parameters to obtain the soft measurement values ​​of difficult-to-measure process parameters such as slag heat loss on ignition and dioxin concentration.

[0174] The data forward acquisition isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter reverse transmission module, and operation parameter auxiliary decision-making module together constitute the industrial field auxiliary decision-making and operation optimization subsystem. Process data is transmitted to the forward server of the data acquisition forward isolation module via OPC. After physical isolation and forward acquisition, it is transmitted to the multimodal data acquisition module. After processing by the flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, and MSWI process single-objective / multi-objective operation optimization module, various types of key operation parameter prediction values ​​and soft measurement values, as well as optimized operation parameter values, are obtained. These are then transmitted to the operation parameter reverse transmission module. After passing through the server, the data is physically isolated and transmitted back to the operating parameter receiving server. Following auxiliary decision analysis in the operating parameter auxiliary decision module, and based on the MSWI plant's safety requirements, optimized operating parameters are transmitted to the field monitoring system and PLC / DCS system using OPC protocol or OCR recognition. The control quantities are then transmitted as analog outputs to the actual MSWI process actuators, acting on the actual objects comprising solid waste storage and transportation, solid waste combustion, waste heat exchange, flue gas purification, and flue gas emission stages. The data is then collected via instrumentation through analog inputs to the PLC / DCS system and field monitoring system, and then transmitted via OPC to the data acquisition forward isolation module and multimodal data acquisition module. Finally, it is fed back to the MSWI process single-objective / multi-objective operation optimization module, completing the implementation of the operation optimization algorithm for the actual industrial site. The functions of each module in the industrial site auxiliary decision and operation optimization subsystem are described below:

[0175] Forward Data Acquisition and Isolation Module: This module acquires process data from the PLC / DCS control system network during MSWI using the OPC client provided by the industrial DCS system manufacturer. It then publishes the acquired real-time data via the OPC communication protocol through physical isolation. The implementation steps for forward data acquisition and isolation are as follows:

[0176] (1) Using the PLC / DCS manufacturer's OPC server protocol, process data is collected from the field process monitoring system to the forward server and published to the outside world in the form of OPC server;

[0177] (2) Connect the forward server and the physically isolated forward acquisition machine to the same local area network through a switch;

[0178] (3) The collected process data is transmitted to the forward data analysis server in a physically isolated forward transmission manner using a unidirectional optical fiber transmission method, including process variable grouping, naming and sampling time settings;

[0179] (4) The physical isolation forward transmission and forward data analysis server are connected to the same local area network through a switch. The forward data analysis server provides data services to the M2 module in the OPC service mode.

[0180] Multimodal data acquisition module: Incinerator flame and process data are crucial for domain experts to assess MSWI process stability. Real-time acquisition and preprocessing of this multimodal data are therefore essential. Specific steps for multimodal data acquisition include:

[0181] (1) Use two identical sets of camera equipment to collect real-time video of the left / right flames at the industrial site;

[0182] (2) Transmit the information captured by the camera to the video capture card via a coaxial cable;

[0183] (3) Install the video capture card into the multimodal data acquisition computer and use video decoding to achieve a preliminary display of flame information;

[0184] (4) The acquired flame video is preprocessed by using image recognition software in the computer, such as noise reduction and enhancement.

[0185] (5) The process data on the OPC server is acquired and transferred to the acquisition computer of the M2 module via industrial Ethernet using the OPC client method;

[0186] (6) Production reports are collected into the multimodal data acquisition computer by means of manual input by domain experts, and the sampling period is set reasonably to achieve synchronous storage after multimodal data acquisition.

[0187] Flame combustion line quantization module: The flame combustion line is the controlled variable required to achieve stable control of the MSWI process. The specific steps for quantization include:

[0188] (1) Using image processing technology, conditional generative adversarial network and cycle consistency generative adversarial network, a complete image library is constructed, which includes a normal sub-library of real combustion lines, a sub-library of real / generated abnormal combustion lines, and a sub-library of generated extreme abnormal flame images of combustion lines.

[0189] (2) Select typical images from a complete image library to form a typical template library and train a twin convolutional neural network;

[0190] (3) Extract combustion line features from new flame images and use twin convolutional neural network similarity metric to achieve adaptation with flame images in the "typical template library";

[0191] (4) For new flame images that are not adapted, the nearest neighbor criterion is used to quantize the combustion line;

[0192] (5) Based on the redundancy discrimination mechanism and combined with the experience of domain experts, the “typical template library” is updated adaptively using non-adaptive images.

[0193] Vision-driven combustion state recognition module: Flame combustion state is the controlled variable required for stable MSWI process control. The specific steps for vision-driven combustion state recognition include:

[0194] (1) The flame image is dehazed and denoised by preprocessing methods such as dehazing algorithm based on artificial multi-exposure image fusion, feature normalization, notch filtering, and median filtering, so as to obtain a clear image.

[0195] (2) Extract multiple physically meaningful features such as brightness, texture, and color from the flame image to represent the image from multiple viewpoints, and reduce these features based on mutual information;

[0196] (3) The above-mentioned reduced features are used as inputs to image classifiers such as support vector machines, deep forests, and convolutional neural networks to establish visual-driven combustion state recognition models for the left and right grates respectively.

[0197] (4) For each new flame image, the combustion state it represents is identified.

[0198] Multimodal data-driven process parameter prediction module: The ability to predict process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration in one or more steps is crucial for domain experts. The specific steps for achieving multimodal data-driven process parameter prediction include:

[0199] (1) Using the multimodal data stored in the M2 module, extract flame image features based on the M4 module;

[0200] (2) Combine flame features with process data in series to form new features for training a prediction model of key process parameters;

[0201] (3) Input the newly acquired multimodal process data from the M2 module into the above-mentioned key process parameter prediction model to obtain one-step or even multi-step prediction outputs of process parameters such as furnace temperature, flue gas oxygen content, steam flow rate, CO emission concentration, and NOx emission concentration.

[0202] The soft measurement module for difficult-to-detect process parameters enables soft measurement modeling of difficult-to-detect parameters such as slag loss on ignition (LOI) for product quality parameters and dioxins for environmental indicators. This is crucial for single-objective / multi-objective operation optimization of MSWI processes. The specific steps for predicting difficult-to-detect process parameters include:

[0203] (1) Organize the difficult-to-measure process parameter data recorded in various production reports as the output true value of the soft measurement model;

[0204] (2) Select the multimodal data stored in the M2 module for the true value of soft measurement output for difficult-to-measure process parameters, that is, obtain the multimodal data time period corresponding to the true value of soft measurement output;

[0205] (3) For the flame data within the corresponding soft measurement output true value time period, extract the flame image features according to the M4 module, and combine them with the process data in series to form new features for the soft measurement model input.

[0206] (4) Use the above-mentioned reduced features as inputs to regression models such as support vector machines, deep forests, and deep neural networks to establish soft measurement models for difficult-to-detect process parameters;

[0207] (5) Input the newly acquired multimodal process data from the M2 module into the soft measurement model of the above difficult-to-measure process parameters to obtain the soft measurement values ​​of difficult-to-measure process parameters such as slag heat loss on ignition and dioxin concentration.

[0208] MSWI process single-objective / multi-objective operation optimization module: Under different scenarios, considering single or multiple objectives such as flue gas emission indicators, economic benefits, and slag thermal reduction rate, it enables timely adjustment of the setpoints of key controlled variables such as furnace temperature, flue gas oxygen content, and steam flow rate according to the dynamic changes of the MSWI process. This is crucial for achieving optimized control of the MSWI process. The specific steps for achieving single-objective / multi-objective operation optimization of the MSWI process include:

[0209] (1) Based on MSWI process analysis for single / multi-objective optimization, a single / multi-objective optimization model was established with key controlled variables such as furnace temperature, boiler steam flow rate and flue gas oxygen content as decision variables, minimizing pollutant emission indicators and heat reduction rate, and maximizing combustion efficiency and economic indicators.

[0210] (2) Based on the above single / multi-objective optimization model, the multimodal data collected in real time by the M2 module and the soft measurement model of difficult process parameters by the M6 ​​module are used to obtain the optimized setpoints of key controlled variables by intelligent optimization algorithms such as genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm.

[0211] The reverse transmission module for operating parameters obtains single-step / multi-step predicted values ​​and soft measurement values ​​of key operating parameter models, as well as optimized operating parameter values, from a reverse server. This data is then transmitted via physical isolation and transmitted in OPC format through a reverse data analysis server. The implementation steps of the M7-operating parameter reverse transmission module are as follows:

[0212] (1) Collect and store the single-step / multi-step predicted values ​​and soft measurement values ​​of the key operating parameter model and the optimized operating parameter values ​​through the reverse server computer;

[0213] (2) Connect the reverse server and the physically isolated reverse acquisition machine to the same local area network through a switch, and then transmit the data to the physically isolated reverse acquisition machine;

[0214] (3) The collected process data is transmitted to the reverse data analysis server in a physically isolated reverse transmission manner using a unidirectional optical fiber transmission method, including process variable grouping, naming and sampling time settings;

[0215] (4) The physically isolated reverse transmission and reverse data analysis server are connected to the same local area network through a switch. The reverse data analysis server provides data services to the M9 module using OPC service.

[0216] Operating parameter-assisted decision-making module: The single-step / multi-step predicted values ​​and soft measurement values ​​of key operating parameter models are crucial for domain experts to implement control strategies and determine whether to apply optimized operating parameters to the control of MSWI processes. The steps to implement operating parameter-assisted decision-making are as follows:

[0217] (1) Collect the single-step / multi-step predicted values ​​and soft measurement values ​​of the key operating parameter model from the M8 module to the operating parameter reverse receiving server;

[0218] (2) Based on the single-step / multi-step predicted values ​​and soft measurement values ​​of the key operating parameter models in different time periods, a statistical comparative analysis is conducted with the actual values ​​of these key operating parameters. The reliability of the single-step / multi-step predicted values ​​and soft measurement values ​​of these key operating parameter models is assisted by methods such as active learning by domain experts and automatic judgment by setting thresholds.

[0219] (3) Based on the reliability of the single-step / multi-step predicted values ​​and soft measurement values ​​of the above key operating parameter models and the optimized operating parameters, two methods are used for data transmission: when the isolation device of the system is fully trusted in the industrial field, the optimized operating parameters of domain expert decision-making or automatic decision-making are transmitted to the field process control system; when higher-level safety issues are required in the industrial field, the optimized operating parameters are transmitted to the field process control system based on the operating parameter OCR recognition technology.

[0220] This invention provides a multi-purpose modular operation optimization system for urban solid waste incineration processes. The system includes a multimodal historical data synchronization and driving module, a MSWI process virtual control object module, an MSWI process loop control module, an MSWI process monitoring module, an operation parameter auxiliary decision-making module, a data acquisition forward isolation module, an operation parameter reverse transmission module, an MSWI process single-objective / multi-objective operation optimization module, a soft measurement module for difficult-to-measure process parameters, a multimodal data-driven process parameter prediction module, a vision-driven combustion state recognition module, a flame combustion line quantification module, and a multimodal data acquisition module. This system is applicable to both laboratory and industrial sites, avoiding the limitations of traditional simulation methods. Overcoming the drawbacks of overly complex or overly simple experimental systems, this system can be modularly built according to specific needs. It can also effectively meet the requirements of isolated data acquisition and software porting in industrial settings. This system can effectively integrate multimodal data synchronous prediction and operation optimization control. It takes into account the security requirements of different MSWI enterprises and can choose OPC server or OCR recognition method for transmission of operating parameters. It has strong practicality and adaptability. This system can realize functions such as laboratory operation optimization algorithm verification, laboratory process parameter modeling algorithm pseudo-real-time verification, industrial site data acquisition and process parameter modeling, and industrial site auxiliary decision-making and operation optimization. It is easy to use.

[0221] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A multi-purpose modular operation optimization system for urban solid waste incineration processes, characterized in that, include: The system comprises a multimodal historical data synchronization driving module, an MSWI process virtual control object module, an MSWI process loop control module, an MSWI process monitoring module, an operating parameter auxiliary decision-making module, a data acquisition forward isolation module, an operating parameter reverse transmission module, an MSWI process single-objective / multi-objective operation optimization module, a difficult-to-measure process parameter soft measurement module, a multimodal data-driven process parameter prediction module, a vision-driven combustion state recognition module, a flame combustion line quantization module, and a multimodal data acquisition module. The multimodal historical data synchronization driving module is connected to the multimodal data acquisition module. The multimodal data acquisition module is connected to the flame combustion line quantization module, the vision-driven combustion state recognition module, the multimodal data-driven process parameter prediction module, the difficult-to-measure process parameter soft measurement module, and the MSWI process single-objective / multi-objective operation optimization module. The flame combustion line quantization module and the vision-driven... The combustion state identification module, the multimodal data-driven process parameter prediction module, the difficult-to-measure process parameter soft measurement module, and the MSWI process single-objective / multi-objective operation optimization module are connected to the operation parameter auxiliary decision-making module and the operation parameter reverse transmission module. The operation parameter reverse transmission module is connected to the operation parameter auxiliary decision-making module. The operation parameter auxiliary decision-making module is connected to the MSWI process monitoring module and the field process monitoring system. The field process monitoring system is connected to the data acquisition forward isolation module. The data acquisition forward isolation module is connected to the multimodal data acquisition module. The MSWI process monitoring module is connected to the MSWI process loop control module. The MSWI process loop control module is connected to the MSWI process virtual control object module. The field process monitoring system is connected to the field loop control system. The field loop control system is connected to the actuators and instruments. The multimodal historical data synchronization driver module is used to provide multimodal data sources for the MSWI process; The MSWI process virtual control object module is used to simulate the MSWI process built in the laboratory; The MSWI process loop control module is used to realize loop control of the virtual MSWI process; The MSWI process monitoring module is used to monitor the virtual MSWI process; The operating parameter auxiliary decision-making module is used to acquire operating parameters and perform comparative analysis and decision-making. The data acquisition forward isolation module is used to acquire data from all process variables in the monitoring module of the virtual MSWI process through physical isolation. The reverse transmission module for operating parameters is used to transmit the optimized values ​​of operating parameters obtained from the MSWI process single-objective / multi-objective operation optimization module, as well as the operating parameters from the flame combustion line quantification module, the vision-driven combustion state recognition module, the multimodal data-driven process parameter prediction module, and the operating parameter detection values ​​from the difficult-to-measure process parameter soft measurement module in a physically isolated manner. The MSWI process single-objective / multi-objective operation optimization module is used to optimize the MSWI process operation parameters based on multimodal data and a soft measurement model for difficult-to-measure parameters; The soft measurement module for difficult-to-detect process parameters is used to perform soft measurement modeling for parameters that are difficult to detect based on multimodal data and production reports. The multimodal data-driven process parameter prediction module is used to realize single-step / multi-step prediction of process parameters such as furnace temperature, flue gas oxygen content, and steam flow rate based on multimodal data. The vision-driven combustion state recognition module is used to recognize the combustion state in the furnace using a domain expert-like recognition mechanism for the MSWI process. The flame combustion line quantization module is used to quantify the flame combustion line using a domain expert identification mechanism for the MSWI process. The multimodal data acquisition module is used to acquire simulated and actual MSWI processes with multimodal data, including left grate flame video, right grate flame video, and historical process data, as well as process and input various production reports related to product quality, environmental indicators, and economic indicators generated in the actual process. The data acquisition forward isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter reverse transmission module, and operation parameter auxiliary decision-making module together constitute the industrial field auxiliary decision-making and operation optimization subsystem. Process data is transmitted to the forward server of the data acquisition forward isolation module via OPC. After physical isolation and forward acquisition, it is transmitted to the multimodal data acquisition module. After processing by the flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, and MSWI process single-objective / multi-objective operation optimization module, various types of key operation parameter predicted values ​​and soft measurement values, as well as optimized operation parameter values, are obtained. After being transmitted to the reverse server of the operating parameter reverse transmission module, the parameters are physically isolated and then transmitted back to the operating parameter reverse receiving server. After auxiliary decision analysis by the operating parameter auxiliary decision module, the optimized operating parameters are transmitted to the field monitoring system and PLC / DCS system using OPC protocol or OCR recognition method according to the safety requirements of the MSWI plant. The control quantity is then transmitted to the actual MSWI process actuator in the form of analog output, acting on the actual object consisting of solid waste storage and transportation, solid waste combustion, waste heat exchange, flue gas purification and flue gas emission stages. The analog input is then collected by the instrument device and sent to the PLC / DCS system and the field monitoring system. It is then transmitted to the data acquisition forward isolation module and the multimodal data acquisition module via OPC, and then fed back to the MSWI process single-objective / multi-objective operation optimization module to complete the implementation of the operation optimization algorithm for the actual industrial site.

2. The multi-purpose modular operation optimization system for urban solid waste incineration process according to claim 1, characterized in that, The multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, difficult-to-measure process parameter soft measurement module, MSWI process single-objective / multi-objective operation optimization module, operation parameter auxiliary decision-making module, MSWI process monitoring module, MSWI process loop control module, and MSWI process virtual control object module together constitute the laboratory operation optimization algorithm verification subsystem. The multimodal data acquisition module acquires synchronized left and right grate flame image data and historical process data from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together form the laboratory operation optimization algorithm verification subsystem. The process parameter soft measurement module and the difficult-to-measure process parameter soft measurement module process and obtain various types of operating parameters. The operating parameters are transmitted to the MSWI process monitoring module via OPC and downloaded to the MSWI process loop control module based on real PLC / DCS equipment. Then, the control quantity is transmitted to the virtual actuator of the MSWI process virtual control object module as an analog output. The output of the virtual actuator then acts on the virtual object to generate the controlled variable output, which is transmitted to the MSWI process loop control module as an analog input through a virtual instrument device. It is then transmitted to the MSWI process monitoring module, and then transmitted to the multimodal data acquisition module via OPC. Finally, it is fed back to the difficult-to-measure process parameter soft measurement module to complete the verification of the operation optimization algorithm for laboratory use.

3. The multi-purpose modular operation optimization system for urban solid waste incineration process according to claim 1, characterized in that, The multimodal historical data synchronization driving module, multimodal data acquisition module, flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together constitute a laboratory process parameter modeling algorithm simulated real-time verification subsystem. The multimodal data acquisition module acquires synchronized left and right grate flame image data and historical process data from the multimodal historical data synchronization driving module. The flame combustion line quantization module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module realize the process parameter prediction, combustion state recognition, combustion line quantization, and process parameter soft measurement results simulated in real-time multimodal data synchronization and release in an industrial setting.

4. The multi-purpose modular operation optimization system for urban solid waste incineration process according to claim 1, characterized in that, The data acquisition forward isolation module, multimodal data acquisition module, flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module together constitute the industrial field data acquisition and process parameter modeling subsystem. The industrial field data is transmitted to the forward server via OPC, and then transmitted to the multimodal data acquisition module after physical isolation forward acquisition. The flame combustion line quantification module, vision-driven combustion state recognition module, multimodal data-driven process parameter prediction module, and difficult-to-measure process parameter soft measurement module realize the process parameter prediction, combustion state recognition, combustion line quantification, and process parameter soft measurement results based on the real-time multimodal data of the industrial field.

Images