SCADA-based thermal power plant auxiliary control system and method

Abstract

The invention relates to the technical field of thermal power plants, and particularly discloses a thermal power plant auxiliary control system and method based on SCADA (supervisory control and data acquisition). Performing time sequence modeling on the hearth temperature, the oxygen content and the pulverized coal concentration data by using a time sequence analysis technology based on deep learning, and then, taking the time sequence joint characteristics of the oxygen content and the pulverized coal concentration as combustion-supporting variable characteristics, and taking the time sequence change characteristics of the hearth temperature as combustion variable characteristics; the influence mechanism of the oxygen content and the pulverized coal concentration on the hearth temperature is excavated by performing fine-grained interactive response analysis on the oxygen content and the pulverized coal concentration, and therefore intelligent recommendation control over the air supply amount and the pulverized coal supply proportion is achieved on the basis. In this way, optimal control over the boiler combustion process of the thermal power plant can be achieved, the combustion efficiency and the heat stability are improved, and then the economic benefits and the environmental protection performance of the power plant are improved.

Description

Technical Field

[0001] This application relates to the technical field of thermal power plants, and more specifically, to a SCADA-based auxiliary control system and method for thermal power plants. Background Art

[0002] As an important source of power supply, the operating efficiency and safety of thermal power plants are directly related to energy utilization efficiency and environmental protection. During the operation of a power plant, the combustion efficiency and stability of the boiler have a crucial impact on the overall performance of the power plant. Among them, the temperature inside the boiler combustion chamber is one of the key factors affecting combustion efficiency and thermal stability. Too high or too low temperature will lead to a decrease in combustion efficiency and affect the economic benefits of the power plant. For example, too high a furnace temperature may cause excessive thermal stress on the boiler heating surface, resulting in equipment damage; while too low a furnace temperature will cause incomplete fuel combustion, increasing energy consumption and pollutant emissions.

[0003] Regarding the control of furnace temperature, traditional auxiliary control methods for thermal power plants usually rely on the experience of operators to manually adjust the furnace temperature. This method often has problems such as slow response speed and low adjustment accuracy, and cannot meet the requirements of modern thermal power plants for efficient, stable, and environmentally friendly operation.

[0004] In recent years, with the development of the Industrial Internet of Things (IIoT) and the progress of sensor technology, the Supervisory Control and Data Acquisition (SCADA) system has been widely used in the power production process. The SCADA system can collect and monitor various operating parameters in a thermal power plant in real time, including key information such as furnace temperature, air supply volume, and pulverized coal supply, providing strong data support for the control of furnace temperature. However, some existing automated auxiliary control systems are mostly based on fixed rules or simple threshold feedback control, and can often only achieve basic automatic adjustment functions, making it difficult to adapt to complex combustion condition changes, and having problems such as unsatisfactory adjustment effects and poor adaptability.

[0005] Therefore, an optimized SCADA-based auxiliary control system and method for thermal power plants are expected. Summary of the Invention

[0006] This application provides a SCADA-based auxiliary control system and method for thermal power plants, which can achieve optimized control of the boiler combustion process in a thermal power plant, improve combustion efficiency and thermal stability, and thus enhance the economic benefits and environmental performance of the power plant.

[0007] In a first aspect, a SCADA-based auxiliary control method for a thermal power plant is provided, including:

[0008] Collect boiler combustion parameters using the SCADA system to obtain a time-series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration;

[0009] Perform time-series encoding based on the parameter sample dimension on the time-series data set of the boiler combustion parameters to obtain a furnace temperature time-series feature implicit encoding vector, an oxygen content time-series feature implicit encoding vector, and a pulverized coal concentration time-series feature implicit encoding vector;

[0010] Fuse the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector to obtain a combustion-aiding variable time-series feature implicit encoding vector;

[0011] Perform fine-grained semantic alignment interaction response analysis on the furnace temperature time-series feature implicit encoding vector and the combustion-aiding variable time-series feature implicit encoding vector to obtain a combustion variable-combustion-aiding variable time-series fine-grained alignment response encoding vector;

[0012] Based on the combustion variable-combustion-aiding variable time-series fine-grained alignment response encoding vector, determine the auxiliary control results of the air supply volume and pulverized coal supply.

[0013] In a second aspect, a secondary control system for a thermal power plant based on SCADA is provided, including:

[0014] A boiler combustion parameter acquisition module for collecting boiler combustion parameters using the SCADA system to obtain a time-series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration;

[0015] A boiler combustion parameter encoding module for performing time-series encoding based on the parameter sample dimension on the time-series data set of the boiler combustion parameters to obtain a furnace temperature time-series feature implicit encoding vector, an oxygen content time-series feature implicit encoding vector, and a pulverized coal concentration time-series feature implicit encoding vector;

[0016] A combustion-aiding variable fusion module for fusing the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector to obtain a combustion-aiding variable time-series feature implicit encoding vector;

[0017] A combustion variable-combustion-aiding variable interaction module for performing fine-grained semantic alignment interaction response analysis on the furnace temperature time-series feature implicit encoding vector and the combustion-aiding variable time-series feature implicit encoding vector to obtain a combustion variable-combustion-aiding variable time-series fine-grained alignment response encoding vector;

[0018] A secondary control result determination module for determining the secondary control results of the air supply volume and pulverized coal supply based on the combustion variable-combustion-aiding variable time-series fine-grained alignment response encoding vector.

[0019] A SCADA-based auxiliary control system and method for thermal power plants provided by this application utilize the SCADA system to collect parameters such as furnace temperature, oxygen content, and pulverized coal concentration during the boiler combustion process, and use time series analysis technology based on deep learning to perform time series modeling on the furnace temperature, oxygen content, and pulverized coal concentration data to capture the time series change trends of the furnace temperature, oxygen content, and pulverized coal concentration. Then, taking the time series joint features of oxygen content and pulverized coal concentration as the combustion-aiding variable features, and taking the time series change features of furnace temperature as the combustion variable features, through fine-grained interaction response analysis of the two, to explore the influence mechanism of oxygen content and pulverized coal concentration on furnace temperature, and thus realize intelligent recommended control of air supply volume and pulverized coal supply ratio on this basis. In this way, the optimized control of the boiler combustion process in thermal power plants can be achieved, improving combustion efficiency and thermal stability, and further enhancing the economic benefits and environmental protection performance of the power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below only relate to some embodiments of this application and do not limit this application.

[0021] Figure 1 It is a schematic flow chart of the SCADA-based auxiliary control method for thermal power plants in the embodiments of this application.

[0022] Figure 2 It is a schematic diagram of data flow of the SCADA-based auxiliary control method for thermal power plants in the embodiments of this application.

[0023] Figure 3 It is a schematic flow chart of step S2 in the SCADA-based auxiliary control method for thermal power plants in the embodiments of this application.

[0024] Figure 4 It is a schematic flow chart of step S4 in the SCADA-based auxiliary control method for thermal power plants in the embodiments of this application.

[0025] Figure 5 It is a schematic block diagram of the SCADA-based auxiliary control system for thermal power plants in the embodiments of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the drawings. Obviously, the described embodiments are only some embodiments of this application, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of this application without creative efforts also belong to the scope of protection of this application.

[0027] In view of the above technical problems, the technical concept of this application is as follows: Use the SCADA system to collect the furnace temperature, oxygen content, and pulverized coal concentration parameters during the boiler combustion process, and use the time series analysis technology based on deep learning to perform time series modeling on the furnace temperature, oxygen content, and pulverized coal concentration data to capture the time series change trends of the furnace temperature, oxygen content, and pulverized coal concentration. Then, use the time series joint features of the oxygen content and pulverized coal concentration as the combustion-aiding variable features, and use the time series change features of the furnace temperature as the combustion variable features. Through fine-grained interaction response analysis of the two, the influence mechanism of the oxygen content and pulverized coal concentration on the furnace temperature is explored, and on this basis, intelligent recommended control of the air supply volume and pulverized coal supply ratio is realized. In this way, the optimal control of the boiler combustion process in a thermal power plant can be achieved, the combustion efficiency and thermal stability can be improved, and further the economic benefits and environmental protection performance of the power plant can be enhanced.

[0028] Based on this, in the technical solution of this application, Figure 1 is a schematic flowchart of the SCADA-based auxiliary control method for a thermal power plant according to an embodiment of this application. Figure 2 is a schematic diagram of data flow of the SCADA-based auxiliary control method for a thermal power plant according to an embodiment of this application. As Figure 1 and Figure 2 shown, the SCADA-based auxiliary control method for a thermal power plant includes: S1, using the SCADA system to collect boiler combustion parameters to obtain a time series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration; S2, performing time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain a furnace temperature time series feature implicit encoding vector, an oxygen content time series feature implicit encoding vector, and a pulverized coal concentration time series feature implicit encoding vector; S3, fusing the oxygen content time series feature implicit encoding vector and the pulverized coal concentration time series feature implicit encoding vector to obtain a combustion-aiding variable time series feature implicit encoding vector; S4, performing fine-grained semantic alignment interaction response analysis on the furnace temperature time series feature implicit encoding vector and the combustion-aiding variable time series feature implicit encoding vector to obtain a combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector; S5, based on the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector, determining the auxiliary control results of the air supply volume and pulverized coal supply.

[0029] Exemplarily, in step S1, a SCADA system is used to collect boiler combustion parameters to obtain a time-series data set of boiler combustion parameters. Among them, the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration. It should be understood that the furnace temperature is a key indicator for measuring the boiler combustion state, which is directly related to the thermal efficiency and operation stability of the boiler. The oxygen content refers to the oxygen concentration entering the boiler combustion chamber, which plays a crucial role in the combustion process. An appropriate oxygen content can ensure that the fuel burns fully in the boiler, thereby improving the combustion efficiency. When the oxygen content is too low, the fuel burns incompletely, and the unburned pulverized coal will be discharged with the flue gas, not only causing energy waste but also increasing the maintenance cost of the boiler. On the contrary, when the oxygen content is too high, excessive air entering the furnace will carry away a large amount of heat, reducing the thermal efficiency of the boiler and increasing the emissions of pollutants such as nitrogen oxides (NOx). Therefore, precisely controlling the oxygen content is the key to maintaining a high combustion efficiency. At the same time, the pulverized coal concentration also has a direct impact on the combustion efficiency and stability of the boiler. Too high a pulverized coal concentration may lead to incomplete combustion and increase the carbon content in fly ash; while too low a pulverized coal concentration may cause insufficient combustion and reduce the furnace temperature. That is to say, the oxygen content and pulverized coal concentration, as the necessary conditions and main fuel for the combustion reaction, jointly affect the change of the furnace temperature, and thus have an important impact on the combustion efficiency and stability of the boiler. Based on this, in the technical solution of this application, in order to achieve optimal control of the boiler combustion process, it is expected to further determine the optimal ratio of air supply volume and pulverized coal supply by analyzing the dynamic influence of oxygen content and pulverized coal concentration on the furnace temperature.

[0030] In one embodiment, the SCADA system will establish a communication connection with sensors installed inside and around the boiler. These sensors are responsible for measuring key combustion parameters such as furnace temperature, oxygen content, and pulverized coal concentration. Whenever the sensors detect changes in these parameters, they will send the data to the central computer or server of the SCADA system. This process is usually periodic, that is, the latest readings are collected at set time intervals (such as every second, every minute), thereby forming a series of data points arranged in chronological order, which is the time-series data set of the boiler combustion parameters.

[0031] Exemplarily, in step S2, a time series encoding based on the parameter sample dimension is performed on the time series data set of the boiler combustion parameters to obtain a hidden encoding vector of the furnace temperature time series feature, a hidden encoding vector of the oxygen content time series feature, and a hidden encoding vector of the pulverized coal concentration time series feature. It should be understood that by performing time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters, the dynamic change characteristics of the furnace temperature, oxygen content, and pulverized coal concentration in the time series can be captured. Specifically, by encoding the time series data set, the patterns and rules hidden behind a large number of observations can be extracted. For example, for the furnace temperature, the encoded hidden vector can reflect information such as the trend, periodicity, and abnormal conditions of the temperature fluctuation over time; for the oxygen content and pulverized coal concentration, the key factors and change patterns affecting the combustion efficiency can be revealed through their respective hidden encoding vectors. This fine-grained feature representation helps to uncover the complex interaction effects between different parameters, especially those non-linear or with lag effects. In this application, adopting the time series encoding method based on the parameter sample dimension can also enhance the model's ability to predict future states. This is because the hidden vector obtained after the encoding process not only contains the information at the current moment but also integrates the historical information over a period of time in the past, enabling the model to better understand and simulate the dynamic behavior in the actual physical process during the learning process. This is crucial for achieving intelligent recommendation control because it allows the system to adjust the air supply volume and pulverized coal supply ratio according to the current and past operating conditions, thereby achieving the purpose of optimizing the combustion process.

[0032] In one embodiment, as Figure 3 shown, performing a time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain a hidden encoding vector of the furnace temperature time series feature, a hidden encoding vector of the oxygen content time series feature, and a hidden encoding vector of the pulverized coal concentration time series feature includes: S21, performing data segmentation on the time series data set of the boiler combustion parameters to obtain a time series data set of the furnace temperature, a time series data set of the oxygen content, and a time series data set of the pulverized coal concentration; S22, respectively inputting the time series data set of the furnace temperature, the time series data set of the oxygen content, and the time series data set of the pulverized coal concentration into a sequence encoder based on a bidirectional gated recurrent unit to obtain the hidden encoding vector of the furnace temperature time series feature, the hidden encoding vector of the oxygen content time series feature, and the hidden encoding vector of the pulverized coal concentration time series feature.

[0033] Exemplarily, in step S21, the time-series dataset of the boiler combustion parameters is sliced to obtain the time-series dataset of the furnace temperature, the time-series dataset of the oxygen content, and the time-series dataset of the pulverized coal concentration. It should be understood that considering that although the three parameters of furnace temperature, oxygen content, and pulverized coal concentration are interrelated during the boiler combustion process, each has its own unique variation law and physical meaning. Therefore, in order to more specifically analyze the time-series variation characteristics of each boiler combustion parameter, the present application further slices the time-series dataset of the boiler combustion parameters to form the time-series dataset of the furnace temperature, the time-series dataset of the oxygen content, and the time-series dataset of the pulverized coal concentration, so as to realize the construction of independent datasets, in order to avoid the mutual interference between different parameter data during the subsequent time-series data analysis process, thereby improving the accuracy of data analysis.

[0034] Exemplarily, in step S22, the time-series dataset of the furnace temperature, the time-series dataset of the oxygen content, and the time-series dataset of the pulverized coal concentration are respectively input into a sequence encoder based on a bidirectional gated recurrent unit to obtain the time-series feature hidden coding vector of the furnace temperature, the time-series feature hidden coding vector of the oxygen content, and the time-series feature hidden coding vector of the pulverized coal concentration. It should be understood that in order to fully capture the time-series variation characteristics of the furnace temperature, oxygen content, and pulverized coal concentration during the boiler combustion process, the present application uses a bidirectional gated recurrent unit (Bi-directional Gated Recurrent Unit, abbreviated as Bi-GRU) network, which has excellent performance in the field of time-series analysis, to perform time-series modeling on the time-series dataset of the furnace temperature, the time-series dataset of the oxygen content, and the time-series dataset of the pulverized coal concentration respectively. It should be understood that the bidirectional gated recurrent unit (Bi-GRU) is an improved recurrent neural network (RNN) model. By introducing a gating mechanism and a bidirectional propagation path, it can effectively capture the long-term dependence relationships in the time-series dataset and reduce the problems of gradient vanishing or gradient explosion. In the technical solution of the present application, using the bidirectional gated recurrent unit (Bi-GRU) network to scan and learn the time-series datasets of the furnace temperature, oxygen content, and pulverized coal concentration from both the forward and backward directions can comprehensively extract the variation trends, periodicity, and the correlations with the data at adjacent time moments of each parameter in the time series, and compress and encode them into a fixed-length hidden vector representation to obtain the corresponding time-series feature hidden coding vector of the furnace temperature, the time-series feature hidden coding vector of the oxygen content, and the time-series feature hidden coding vector of the pulverized coal concentration, providing accurate data support for subsequent intelligent recommendation control.

[0035] Exemplarily, in step S3, the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector are fused to obtain the combustion-aiding variable time-series feature implicit encoding vector. It should be understood that considering that during the boiler combustion process, the oxygen content and the pulverized coal concentration jointly act on the combustion reaction, and both have an important synergistic effect on the change of the furnace temperature. Therefore, in this application, the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector are further concatenated to jointly represent the time-series features of the oxygen content and the pulverized coal concentration, forming the combustion-aiding variable time-series feature implicit encoding vector, so as to comprehensively consider the comprehensive effect of the oxygen content and the pulverized coal concentration on the combustion process during the subsequent boiler combustion auxiliary control decision-making process.

[0036] In one embodiment, fusing the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector to obtain the combustion-aiding variable time-series feature implicit encoding vector includes: concatenating the oxygen content time-series feature implicit encoding vector and the pulverized coal concentration time-series feature implicit encoding vector to obtain the combustion-aiding variable time-series feature implicit encoding vector.

[0037] Exemplarily, in step S4, a fine-grained semantic alignment interaction response analysis is performed on the furnace temperature time-series feature implicit encoding vector and the combustion-aiding variable time-series feature implicit encoding vector to obtain the combustion variable-combustion-aiding variable time-series fine-grained alignment response encoding vector. That is, a time-series interaction response analysis is further performed on the furnace temperature time-series feature implicit encoding vector and the combustion-aiding variable time-series feature implicit encoding vector to reveal the dynamic influence mechanism of the oxygen content and the pulverized coal concentration on the furnace temperature. In particular, considering that the influence of the oxygen content and the pulverized coal concentration on the furnace temperature may be non-linear and time-delayed, a simple linear model cannot accurately capture this complex dynamic relationship. In response to this, this application proposes a fine-grained semantic alignment interaction response analysis method, by constructing a semantic flow field between the furnace temperature time-series feature implicit encoding vector and the combustion-aiding variable time-series feature implicit encoding vector, to capture the time-delay effect and interaction between different variables, so as to more accurately reveal the influence of the oxygen content and the pulverized coal concentration on the furnace temperature.

[0038] In one embodiment, as Figure 4As shown, perform fine-grained semantic alignment interaction response analysis on the implicit encoding vector of the furnace temperature time series feature and the implicit encoding vector of the combustion-aiding variable time series feature to obtain the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector, including: S41, constructing a semantic flow field between the implicit encoding vector of the furnace temperature time series feature and the implicit encoding vector of the combustion-aiding variable time series feature to obtain the combustion variable-combustion-aiding variable time series feature semantic flow field; S42, performing feature alignment processing on the implicit encoding vector of the furnace temperature time series feature and the implicit encoding vector of the combustion-aiding variable time series feature based on the combustion variable-combustion-aiding variable time series feature semantic flow field to obtain the aligned implicit encoding vector of the furnace temperature time series feature and the aligned implicit encoding vector of the combustion-aiding variable time series feature; S43, performing time series interaction response encoding on the aligned implicit encoding vector of the furnace temperature time series feature and the aligned implicit encoding vector of the combustion-aiding variable time series feature to obtain the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector.

[0039] In one embodiment, in step S41, constructing a semantic flow field between the implicit encoding vector of the furnace temperature time series feature and the implicit encoding vector of the combustion-aiding variable time series feature to obtain the combustion variable-combustion-aiding variable time series feature semantic flow field includes: performing feature dimension modulation on the implicit encoding vector of the furnace temperature time series feature and the implicit encoding vector of the combustion-aiding variable time series feature to obtain a dimension-modulated implicit encoding vector of the furnace temperature time series feature and a dimension-modulated implicit encoding vector of the combustion-aiding variable time series feature with the same feature dimension. Specifically, this process can be expressed by the formula:

[0040] v′1 = Sigmoid[Conv 1×1 (v1)],

[0041] v′2 = Sigmoid[Conv 1×1 (v2)],

[0042] where, v1 represents the implicit encoding vector of the furnace temperature time series feature, v2 represents the implicit encoding vector of the combustion-aiding variable time series feature, Conv 1×1 (·) represents the point convolution operation, Sigmoid represents the Sigmoid activation function, v′1 represents the dimension-modulated implicit encoding vector of the furnace temperature time series feature, and v′2 represents the dimension-modulated implicit encoding vector of the combustion-aiding variable time series feature.

[0043] Perform temporal correlation encoding on the latent encoding vector of the temporal features of the dimension-modulated furnace temperature and the latent encoding vector of the temporal features of the dimension-modulated combustion-aiding variables to obtain a combustion variable-combustion-aiding variable temporal feature correlation encoding matrix; perform multi-scale convolution and upsampling processing on the combustion variable-combustion-aiding variable temporal feature correlation encoding matrix to construct the combustion variable-combustion-aiding variable temporal feature semantic flow field. Specifically, this process can be expressed by the formula:

[0044]

[0045] where, (·) T represents the transpose of a vector, represents matrix multiplication operation, L is the feature scale value of the latent encoding vector of the temporal features of the dimension-modulated furnace temperature and the latent encoding vector of the temporal features of the dimension-modulated combustion-aiding variables, Conv 5×5 (·) represents a 5×5 convolution operation, Conv 3×3 (·) represents a 3×3 convolution operation, Upsampling represents upsampling operation, and Ω represents the combustion variable-combustion-aiding variable temporal feature semantic flow field.

[0046] That is, perform feature dimension modulation on the latent encoding vector of the temporal features of the furnace temperature and the latent encoding vector of the temporal features of the combustion-aiding variables to obtain a dimension-modulated latent encoding vector of the temporal features of the furnace temperature and a dimension-modulated latent encoding vector of the temporal features of the combustion-aiding variables with the same feature dimension, including: perform point convolution processing based on the Sigmoid function on the latent encoding vector of the temporal features of the furnace temperature and the latent encoding vector of the temporal features of the combustion-aiding variables respectively to obtain the dimension-modulated latent encoding vector of the temporal features of the furnace temperature and the dimension-modulated latent encoding vector of the temporal features of the combustion-aiding variables.

[0047] It should be understood that, first, feature dimension modulation is performed on the latent encoded vector of the furnace temperature time series characteristics and the latent encoded vector of the combustion-aiding variable time series characteristics. This process ensures that data from two different sources can be compared and analyzed within the same feature space. Through feature dimension modulation, the system can adjust the original encoded vectors, which may have different dimensions or scales, to a consistent state, thereby generating a dimension-modulated latent encoded vector of the furnace temperature time series characteristics and a dimension-modulated latent encoded vector of the combustion-aiding variable time series characteristics with the same feature dimension. This step provides a basis for subsequent deeper interaction analysis, ensuring that data collected at different time points or under different conditions can be effectively processed within a unified framework. Next, temporal correlation encoding is performed on these two dimension-modulated latent encoded vectors to construct a combustion variable-combustion-aiding variable time series feature correlation encoding matrix. This matrix not only records the feature values at each moment, but more importantly, it captures the correlation between the furnace temperature and the combustion-aiding variables that changes over time. This temporal correlation encoding can reveal the interaction patterns between the two at different moments, and how they jointly affect the combustion efficiency and stability of the boiler. For example, it can show the change trend of the furnace temperature when the oxygen content increases; or the specific impact on the combustion effect when the pulverized coal concentration fluctuates. By deeply exploring these correlations, valuable information can be provided for optimizing the combustion process. Finally, to further enhance the understanding of the complex relationship between the combustion variables and the combustion-aiding variables, the system performs multi-scale convolution and upsampling processing on the above-mentioned correlation encoding matrix to construct a combustion variable-combustion-aiding variable time series feature semantic flow field. The multi-scale convolution operation allows the model to capture local and global feature information at different levels, ensuring that even subtle changes are not overlooked. The upsampling process helps to restore the spatial details that may be lost during the convolution process, ensuring that the finally generated semantic flow field contains both macroscopic trends and microscopic accuracy. The obtained semantic flow field comprehensively reflects the dynamic relationship between the furnace temperature and the combustion-aiding variables.

[0048] Exemplarily, in step S42, feature alignment processing is performed on the latent encoded vector of the furnace temperature time series characteristics and the latent encoded vector of the combustion-aiding variable time series characteristics based on the combustion variable-combustion-aiding variable time series feature semantic flow field to obtain an aligned latent encoded vector of the furnace temperature time series characteristics and an aligned latent encoded vector of the combustion-aiding variable time series characteristics. Specifically, this process can be expressed by the formula:

[0049]

[0050] where, v 1t and v 2t respectively represent the aligned latent encoded vector of the furnace temperature time series characteristics and the aligned latent encoded vector of the combustion-aiding variable time series characteristics.

[0051] That is, feature alignment processing is performed on the constructed semantic flow field. During this process, the system dynamically adjusts the feature intensities and directions at each position in the implicit encoding vector of the furnace temperature time-series features and the implicit encoding vector of the combustion-aiding variable time-series features, so that information from different data sources can achieve fine-grained matching and interaction within the same semantic space. This alignment processing ensures that even data collected at different time points or under different conditions can be effectively compared and combined, enhancing the model's ability to understand subtle changes during the combustion process. After alignment, the implicit encoding vectors of the furnace temperature time-series features and the implicit encoding vectors of the combustion-aiding variable time-series features already have higher similarity and consistency, laying a solid foundation for further interaction response analysis.

[0052] In one embodiment, in step S43, performing temporal interaction response encoding on the aligned implicit encoding vector of the furnace temperature time-series features and the aligned implicit encoding vector of the combustion-aiding variable time-series features to obtain the combustion variable - combustion-aiding variable temporal fine-grained alignment response encoding vector includes: performing a linear transformation on the aligned implicit encoding vector of the furnace temperature time-series features to obtain a query vector and a value vector; performing a linear transformation on the aligned implicit encoding vector of the combustion-aiding variable time-series features to obtain a key vector. Specifically, this process can be represented by the formula:

[0053]

[0054] where, W 1q 、W 1v and W 2k respectively represent the query embedding matrix, the value embedding matrix, and the key embedding matrix, b 1q 、b 1v and b 2k respectively represent different bias terms, V 1q 、V 1v and V 2k respectively represent the query vector, the value vector, and the key vector.

[0055] Inputting the query vector, the value vector, and the key vector into a fine-grained response encoding module based on the Transformer structure to obtain the combustion variable - combustion-aiding variable temporal fine-grained alignment response encoding vector. Specifically, this process can be represented by the formula:

[0056]

[0057] where, d is the feature scale value of the key vector, softmax(·) represents the normalized exponential function, and V i represents the combustion variable - combustion-aiding variable temporal fine-grained alignment response encoding vector.

[0058] That is, after the feature alignment is completed, a transformer structure is further used to perform cross-domain temporal interaction response encoding between the implicit encoding vector of the furnace temperature temporal features and the implicit encoding vector of the combustion-aiding variable temporal features. Specifically, first, through a linear transformation operation, the system converts the aligned implicit encoding vector of the furnace temperature temporal features into a query vector and a value vector. The query vector is used to represent the focus of the furnace temperature features at the current moment or time period, while the value vector contains the specific information at that moment. This transformation enables the system to more accurately locate the feature parts related to specific combustion states, thereby better understanding the change patterns of the furnace temperature and the underlying physical mechanisms. For example, in a certain combustion stage, if the furnace temperature suddenly rises, the query vector can help identify the key driving factors of this change, while the value vector provides the specific numerical information about this temperature increase. At the same time, the system also performs a linear transformation on the aligned implicit encoding vector of the combustion-aiding variable temporal features to generate a key vector. The role of the key vector is similar to an "index", which can indicate which combustion-aiding variable features are most relevant to the content focused on by the current query vector. This step is crucial for revealing how combustion-aiding variables affect the furnace temperature, as different combustion-aiding variables may play important roles at different time points. For example, when the oxygen content increases, it may immediately cause the furnace temperature to rise; while the change in the pulverized coal concentration may show a certain lag effect. The key vector helps the system identify and quantify these complex causal relationships, enabling it to accurately capture the key influencing factors even in non-linear interaction scenarios. Next, the query vector, value vector, and key vector are input into a fine-grained response encoding module based on the transformer structure. The transformer structure is based on the multi-head self-attention mechanism and the feed-forward neural network. By calculating the similarity between the query vector and the key vector to guide the allocation of attention weights, it dynamically retrieves information related to the furnace temperature features from the combustion-aiding variable temporal features, and finally generates a combustion variable-combustion-aiding variable temporal fine-grained alignment response encoding vector. In this way, the ability to understand and model the complex temporal relationship between the furnace temperature and the combustion-aiding variables can be effectively enhanced, providing more accurate decision-making support for the control of the air supply volume and the pulverized coal supply ratio.

[0059] Exemplarily, in step S5, based on the combustion variable - combustion - assisting variable time - series fine - grained alignment response encoding vector, the auxiliary control results of the air supply volume and pulverized coal supply are determined. In one embodiment, determining the auxiliary control results of the air supply volume and pulverized coal supply based on the combustion variable - combustion - assisting variable time - series fine - grained alignment response encoding vector includes: inputting the combustion variable - combustion - assisting variable time - series fine - grained alignment response encoding vector into an auxiliary control module based on a decoder to obtain the auxiliary control results, where the auxiliary control results include the recommended ratios of the air supply volume and pulverized coal supply. It should be understood that through the fine - grained semantic alignment interaction response analysis performed in the previous steps, the complex dynamic relationships between the furnace temperature, oxygen content, and pulverized coal concentration have been captured. These relationships are often non - linear and may have hysteresis effects. Therefore, in order to extract specific control strategies from such a complex representation, a specially designed decoder is needed to process. The role of the decoder is to decode the actual operable information from the implicit encoding vector, that is, the recommended ratios of the air supply volume and pulverized coal supply, in order to achieve the purpose of optimizing combustion efficiency. Since the detailed interaction information between the combustion variables and combustion - assisting variables has been obtained in the previous steps, the decoder can use this information to make more accurate predictions and suggestions.

[0060] In one embodiment of the present application, inputting the combustion variable - combustion - assisting variable time - series fine - grained alignment response encoding vector into an auxiliary control module based on a decoder to obtain the auxiliary control results includes: using the decoder to perform decoding regression on the combustion variable - combustion - assisting variable time - series fine - grained alignment response encoding vector according to the following formula to obtain the auxiliary control results for representing the recommended ratios of the air supply volume and pulverized coal supply; where the formula is: where Y is the auxiliary control result for representing the recommended ratios of the air supply volume and pulverized coal supply, and W is the weight matrix of the auxiliary control module based on the decoder.

[0061] In summary, the auxiliary control method for a thermal power plant based on SCADA according to the embodiments of the present application is elucidated. It uses the SCADA system to collect the parameters of the furnace temperature, oxygen content, and pulverized coal concentration during the boiler combustion process, and uses time - series analysis technology based on deep learning to perform time - series modeling on the furnace temperature, oxygen content, and pulverized coal concentration data to capture the time - series change trends of the furnace temperature, oxygen content, and pulverized coal concentration. Then, taking the time - series joint features of the oxygen content and pulverized coal concentration as the combustion - assisting variable features, and taking the time - series change features of the furnace temperature as the combustion variable features, through fine - grained interaction response analysis of the two, the influence mechanism of the oxygen content and pulverized coal concentration on the furnace temperature is explored, and on this basis, intelligent recommended control of the air supply volume and pulverized coal supply ratio is realized. In this way, the optimization of the boiler combustion process in a thermal power plant can be achieved.

[0062] Figure 5Schematic block diagram of the auxiliary control system of a thermal power plant based on SCADA according to an embodiment of the present application. As Figure 5 shown, the auxiliary control system 100 of the thermal power plant based on SCADA includes: a boiler combustion parameter acquisition module 110, configured to use the SCADA system to acquire boiler combustion parameters to obtain a time series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration; a boiler combustion parameter encoding module 120, configured to perform time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain a furnace temperature time series feature implicit encoding vector, an oxygen content time series feature implicit encoding vector, and a pulverized coal concentration time series feature implicit encoding vector; a combustion aid variable fusion module 130, configured to fuse the oxygen content time series feature implicit encoding vector and the pulverized coal concentration time series feature implicit encoding vector to obtain a combustion aid variable time series feature implicit encoding vector; a combustion variable-combustion aid variable interaction module 140, configured to perform a fine-grained semantic alignment interaction response analysis on the furnace temperature time series feature implicit encoding vector and the combustion aid variable time series feature implicit encoding vector to obtain a combustion variable-combustion aid variable time series fine-grained alignment response encoding vector; and an auxiliary control result determination module 150, configured to determine the auxiliary control results of the air supply volume and the pulverized coal supply based on the combustion variable-combustion aid variable time series fine-grained alignment response encoding vector.

[0063] In one embodiment, the boiler combustion parameter encoding module is configured to: perform data segmentation on the time series data set of the boiler combustion parameters to obtain a time series data set of furnace temperature, a time series data set of oxygen content, and a time series data set of pulverized coal concentration; and input the time series data set of furnace temperature, the time series data set of oxygen content, and the time series data set of pulverized coal concentration into a sequence encoder based on a bidirectional gated recurrent unit to obtain the furnace temperature time series feature implicit encoding vector, the oxygen content time series feature implicit encoding vector, and the pulverized coal concentration time series feature implicit encoding vector.

[0064] Here, those skilled in the art can understand that the specific operations of the various modules and units in the above-mentioned auxiliary control system of the thermal power plant based on SCADA have been described in detail in the description of the Figures 1 to 4 auxiliary control method of the thermal power plant based on SCADA, and therefore, the repeated description thereof will be omitted.

[0065] An embodiment of the present application further provides a computer program product, which includes computer program code. When the computer program code runs on a computer, the computer is enabled to implement the methods in the above embodiments of the present application.

[0066] The embodiments of the present application also provide a computer-readable storage medium storing computer instructions, which, when running on a computer, enable the computer to implement the methods in the above embodiments of the present application.

[0067] The embodiments of the present application also provide a chip including a circuit for executing the methods in the above embodiments of the present application.

[0068] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0069] In the description of the embodiments of the present application, unless otherwise specified, " / " means "or". For example, A / B may represent A or B; herein, "and / or" is an association relationship describing associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In the present application, "at least one" means one or more, and "a plurality" means two or more. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of single (item) or plural items. For example, at least one (item) of a, b, or c may represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c may be single or multiple.

[0070] In the embodiments of the present application, prefix words such as "first" and "second" are only used to distinguish different described objects and have no limiting effect on the position, order, priority, quantity, content, etc. of the described objects. The use of ordinal words and other prefix words for distinguishing described objects in the embodiments of the present application does not constitute a limitation on the described objects. The statement of the described objects refers to the description in the context of the claims or embodiments, and should not constitute an unnecessary limitation due to the use of such prefix words.

[0071] In the several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0072] In various embodiments of the present application, if there is no special description and logical conflict, the terms and / or descriptions among the various embodiments are consistent and can be cited mutually. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0073] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0074] In addition, in each embodiment of the present application, the various functional units may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0075] As described above, the above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should be covered by the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A SCADA-based auxiliary control method for thermal power plants, characterized in that, Including: Collecting boiler combustion parameters by using a SCADA system to obtain a time series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration; Performing time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain a furnace temperature time series feature implicit encoding vector, an oxygen content time series feature implicit encoding vector, and a pulverized coal concentration time series feature implicit encoding vector; Fusing the oxygen content time series feature implicit encoding vector and the pulverized coal concentration time series feature implicit encoding vector to obtain a combustion aid variable time series feature implicit encoding vector; Performing fine-grained semantic alignment interaction response analysis on the furnace temperature time series feature implicit encoding vector and the combustion aid variable time series feature implicit encoding vector to obtain a combustion variable - combustion aid variable time series fine-grained alignment response encoding vector; Based on the combustion variable - combustion aid variable time series fine-grained alignment response encoding vector, determining the auxiliary control results of the air supply volume and pulverized coal supply.

2. The SCADA-based auxiliary control method for a thermal power plant according to claim 1, wherein, Performing time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain a furnace temperature time series feature implicit encoding vector, an oxygen content time series feature implicit encoding vector, and a pulverized coal concentration time series feature implicit encoding vector, including: Performing data segmentation on the time series data set of the boiler combustion parameters to obtain a time series data set of furnace temperature, a time series data set of oxygen content, and a time series data set of pulverized coal concentration; Inputting the time series data set of furnace temperature, the time series data set of oxygen content, and the time series data set of pulverized coal concentration into a sequence encoder based on a bidirectional gated recurrent unit respectively to obtain the furnace temperature time series feature implicit encoding vector, the oxygen content time series feature implicit encoding vector, and the pulverized coal concentration time series feature implicit encoding vector.

3. The SCADA-based auxiliary control method for thermal power plants according to claim 2, wherein Fusing the oxygen content time series feature implicit encoding vector and the pulverized coal concentration time series feature implicit encoding vector to obtain a combustion aid variable time series feature implicit encoding vector, including: Cascading and splicing the oxygen content time series feature implicit encoding vector and the pulverized coal concentration time series feature implicit encoding vector to obtain the combustion aid variable time series feature implicit encoding vector.

4. The SCADA-based auxiliary control method for thermal power plants according to claim 3, wherein Performing fine-grained semantic alignment interaction response analysis on the furnace temperature time series feature implicit encoding vector and the combustion aid variable time series feature implicit encoding vector to obtain a combustion variable - combustion aid variable time series fine-grained alignment response encoding vector, including: Constructing a semantic flow field between the furnace temperature time series feature implicit encoding vector and the combustion aid variable time series feature implicit encoding vector to obtain a combustion variable - combustion aid variable time series feature semantic flow field; Performing feature alignment processing on the furnace temperature time series feature implicit encoding vector and the combustion aid variable time series feature implicit encoding vector based on the combustion variable - combustion aid variable time series feature semantic flow field to obtain an aligned furnace temperature time series feature implicit encoding vector and an aligned combustion aid variable time series feature implicit encoding vector; Performing time series interaction response encoding on the aligned furnace temperature time series feature implicit encoding vector and the aligned combustion aid variable time series feature implicit encoding vector to obtain the combustion variable - combustion aid variable time series fine-grained alignment response encoding vector.

5. The SCADA-based auxiliary control method for thermal power plants according to claim 4, characterized in that Construct a semantic flow field between the implicit encoding vector of the furnace temperature time series features and the implicit encoding vector of the combustion-aiding variable time series features to obtain a combustion variable-combustion-aiding variable time series feature semantic flow field, including: Perform feature dimension modulation on the implicit encoding vector of the furnace temperature time series features and the implicit encoding vector of the combustion-aiding variable time series features to obtain a dimension-modulated implicit encoding vector of the furnace temperature time series features and a dimension-modulated implicit encoding vector of the combustion-aiding variable time series features with the same feature dimension; Perform time series correlation encoding on the dimension-modulated implicit encoding vector of the furnace temperature time series features and the dimension-modulated implicit encoding vector of the combustion-aiding variable time series features to obtain a combustion variable-combustion-aiding variable time series feature correlation encoding matrix; Perform multi-scale convolution and upsampling processing on the combustion variable-combustion-aiding variable time series feature correlation encoding matrix to construct the combustion variable-combustion-aiding variable time series feature semantic flow field.

6. The SCADA-based auxiliary control method for thermal power plants according to claim 5, characterized in that Perform feature dimension modulation on the implicit encoding vector of the furnace temperature time series features and the implicit encoding vector of the combustion-aiding variable time series features to obtain a dimension-modulated implicit encoding vector of the furnace temperature time series features and a dimension-modulated implicit encoding vector of the combustion-aiding variable time series features with the same feature dimension, including: Perform point convolution processing based on the Sigmoid function on the implicit encoding vector of the furnace temperature time series features and the implicit encoding vector of the combustion-aiding variable time series features respectively to obtain the dimension-modulated implicit encoding vector of the furnace temperature time series features and the dimension-modulated implicit encoding vector of the combustion-aiding variable time series features.

7. The SCADA-based auxiliary control method for thermal power plants according to claim 6, wherein Perform time series interaction response encoding on the aligned implicit encoding vector of the furnace temperature time series features and the aligned implicit encoding vector of the combustion-aiding variable time series features to obtain the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector, including: Perform a linear transformation on the aligned implicit encoding vector of the furnace temperature time series features to obtain a query vector and a value vector; Perform a linear transformation on the aligned implicit encoding vector of the combustion-aiding variable time series features to obtain a key vector; Input the query vector, the value vector, and the key vector into a fine-grained response encoding module based on the transformer structure to obtain the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector.

8. The SCADA-based auxiliary control method for thermal power plants according to claim 7, characterized in that, Based on the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector, determine the auxiliary control results of the air supply volume and the pulverized coal supply, including: Input the combustion variable-combustion-aiding variable time series fine-grained alignment response encoding vector into an auxiliary control module based on a decoder to obtain the auxiliary control results, and the auxiliary control results include the recommended ratios of the air supply volume and the pulverized coal supply.

9. A SCADA-based auxiliary control system for a thermal power plant, characterized in that, Including: A boiler combustion parameter acquisition module, configured to use the SCADA system to acquire boiler combustion parameters to obtain a time series data set of boiler combustion parameters, where the boiler combustion parameters include furnace temperature, oxygen content, and pulverized coal concentration; A boiler combustion parameter encoding module, configured to perform time series encoding based on the parameter sample dimension on the time series data set of the boiler combustion parameters to obtain an implicit encoding vector of furnace temperature time series features, an implicit encoding vector of oxygen content time series features, and an implicit encoding vector of pulverized coal concentration time series features; A combustion-supporting variable fusion module, which is used to fuse the implicit encoding vectors of the time-series features of the oxygen content and the implicit encoding vectors of the time-series features of the pulverized coal concentration to obtain the implicit encoding vectors of the time-series features of the combustion-supporting variables; A combustion variable-combustion-supporting variable interaction module, which is used to perform a fine-grained semantic alignment interaction response analysis on the implicit encoding vectors of the time-series features of the furnace temperature and the implicit encoding vectors of the time-series features of the combustion-supporting variables to obtain the fine-grained alignment response encoding vectors of the time-series of the combustion variables and the combustion-supporting variables; An auxiliary control result determination module, which is used to determine the auxiliary control results of the air supply volume and the pulverized coal supply based on the fine-grained alignment response encoding vectors of the time-series of the combustion variables and the combustion-supporting variables.

10. The SCADA-based auxiliary control system for thermal power plants according to claim 9, wherein The boiler combustion parameter encoding module is used for: Performing data segmentation on the time-series data set of the boiler combustion parameters to obtain a time-series data set of the furnace temperature, a time-series data set of the oxygen content, and a time-series data set of the pulverized coal concentration; Respectively inputting the time-series data set of the furnace temperature, the time-series data set of the oxygen content, and the time-series data set of the pulverized coal concentration into a sequence encoder based on a bidirectional gated recurrent unit to obtain the implicit encoding vectors of the time-series features of the furnace temperature, the implicit encoding vectors of the time-series features of the oxygen content, and the implicit encoding vectors of the time-series features of the pulverized coal concentration.

Images