Energy-saving transformation system and method for electric furnace for preparing silicon carbide from quartz sand

Abstract

The invention provides an energy-saving transformation system and method for an electric furnace for preparing silicon carbide from quartz sand, and relates to the field of intelligent energy-saving transformation. And performing local time sequence correlation on the time queue of the real-time furnace temperature value by using a data processing and feature extraction mode based on deep learning, so as to automatically recommend the operation power of the negative pressure device according to the full-time-domain context causal coding representation between each real-time furnace temperature local time sequence correlation feature after correlation. And based on the real-time operation power of the negative pressure device, an operation power control instruction is obtained. Through precise temperature control and efficient heat dissipation management, unnecessary energy consumption can be avoided, meanwhile, fluctuation in the production process is reduced, and therefore the energy cost is remarkably reduced.

Description

Technical Field

[0001] The present application relates to the field of intelligent energy-saving transformation, and more particularly, in an embodiment of the present application, to an energy-saving transformation system and method for an electric furnace for preparing silicon carbide from quartz sand. Background Art

[0002] In the modern industrial manufacturing system, silicon carbide, as an important high-performance material, has become an indispensable key material in the fields of electronics, automobiles, new energy, etc., due to its excellent high hardness, high thermal conductivity, high temperature resistance and chemical stability. The electric furnace production process of preparing silicon carbide from quartz sand, as an important preparation method of silicon carbide materials, has a direct impact on the development process and cost-effectiveness of related industries due to its production efficiency, energy consumption level and product quality.

[0003] At present, the traditional electric furnace for preparing silicon carbide with quartz sand has exposed a series of serious problems in actual production. From the process level, the size and process of the traditional electric furnace are relatively simple, which makes it lack sufficient flexibility and precision when dealing with complex production needs. In terms of heat dissipation, due to the limitations of design and structure, the traditional electric furnace dissipates heat slowly, and a large amount of heat accumulates in the furnace, which not only prolongs the production cycle, but also increases the heat loss of the equipment, making energy consumption high. Excessive energy consumption not only increases the production cost of the enterprise, but also poses a serious challenge to the sustainable development of the enterprise.

[0004] Therefore, an energy-saving transformation plan for an electric furnace for preparing silicon carbide from quartz sand is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an energy-saving transformation system and method for a quartz sand-based silicon carbide electric furnace, which first obtains a time queue of real-time furnace temperature values ​​collected by a temperature sensor, and uses deep learning-based data processing and feature extraction to perform local temporal association on the time queue of the real-time furnace temperature values, thereby automatically recommending the operating power of the negative pressure device based on the full-time context causal coding representation between the local temporal association features of each real-time furnace temperature after association, and obtaining the operating power control instruction based on the real-time operating power of the negative pressure device. The present application can avoid unnecessary energy consumption through precise temperature control and efficient heat dissipation management, while reducing fluctuations in the production process, thereby significantly reducing energy costs.

[0006] According to one aspect of the present application, there is provided an energy-saving transformation system for a quartz sand-based electric furnace for preparing silicon carbide, comprising: A real-time furnace temperature value acquisition module is used to obtain a time queue of real-time furnace temperature values ​​acquired by a temperature sensor; A local temporal context encoding module for the real-time furnace temperature value is used to perform local temporal association implicit feature encoding and full-time domain causal context encoding of the furnace temperature on the time queue of the real-time furnace temperature value to obtain the full-time domain temporal semantic features of the furnace temperature, and obtain a mapping result based on the full-time domain temporal semantic features of the furnace temperature, wherein the mapping result is the recommended operating power of the negative pressure device; An operating power control instruction generating module is used to obtain an operating power control instruction based on the mapping result and the real-time operating power of the negative pressure device.

[0007] According to another aspect of the present application, there is provided a method for energy-saving transformation of an electric furnace for preparing silicon carbide from quartz sand, comprising: A time queue for obtaining real-time furnace temperature values ​​collected by a temperature sensor; Performing local temporal association implicit feature encoding and full-time domain causal context encoding of the furnace temperature on the time queue of the real-time furnace temperature value to obtain full-time domain temporal semantic features of the furnace temperature, and obtaining a mapping result based on the full-time domain temporal semantic features of the furnace temperature, wherein the mapping result is a recommended operating power of the negative pressure device; Based on the mapping result and the real-time operating power of the negative pressure device, an operating power control instruction is obtained.

[0008] Compared with the prior art, the energy-saving transformation system and method of a quartz sand electric furnace for preparing silicon carbide provided by the present application first obtains the time queue of the real-time furnace temperature value collected by the temperature sensor, and uses the data processing and feature extraction method based on deep learning to perform local time series association on the time queue of the real-time furnace temperature value, so as to automatically recommend the operating power of the negative pressure device based on the full-time domain context causal coding representation between the local time series association features of each real-time furnace temperature after association, and obtain the operating power control instruction based on it and the real-time operating power of the negative pressure device. The present application can avoid unnecessary energy consumption through precise temperature control and efficient heat dissipation management, while reducing fluctuations in the production process, thereby significantly reducing energy costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0010] Figure 1 This is a system block diagram of an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application.

[0011] Figure 2 This is a data flow diagram of an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application.

[0012] Figure 3 The present invention is a block diagram of a local temporal context encoding module for real-time furnace temperature values ​​in an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application.

[0013] Figure 4 This is a block diagram of a full-time domain causal correlation coding unit for temperature characteristics in a furnace in an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application.

[0014] Figure 5 The present invention is a block diagram of a subunit for constructing a real-time furnace temperature topology characteristic matrix in an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application.

[0015] Figure 6 The present invention is a flow chart of a method for energy-saving transformation of an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0017] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0018] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0019] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0020] The traditional electric furnace size process for preparing silicon carbide from quartz sand is relatively simple, but there are problems of slow heat dissipation and high energy consumption. The present application installs a gas collecting hood on the outside of the original electric furnace to effectively collect the gas generated during the production process; at the same time, the dust removal method is improved to bag dust removal, which improves the smoke and dust removal effect. More importantly, by making full use of the structural characteristics of the electric furnace, a pipe for rapid heat removal is installed on the electric furnace. The pipeline forms a mechanism of internal and external pressure difference (i.e., low pressure outside and high pressure inside) through a negative pressure device, which can quickly extract the heat inside the electric furnace, thereby accelerating the heat dissipation process of the furnace. This not only significantly shortens the reaction time of silicon carbide, but also improves the quality of the product, while reducing energy consumption in the entire preparation process.

[0021] In the actual production process of silicon carbide electric furnaces, precise temperature control is crucial to energy efficiency and product quality. The above-mentioned hardware modifications to the electric furnace, including the installation of a gas hood, improved dust removal methods, and the installation of heat removal pipes, have improved the production environment and heat dissipation efficiency to a certain extent, but energy management still needs to be further optimized to achieve more efficient energy-saving goals. The real-time temperature change trend in the furnace is closely related to the operating status of the electric furnace and energy consumption.

[0022] Based on this, the technical concept of the present application is to obtain the time queue of the real-time furnace temperature value collected by the temperature sensor, and use the data processing and feature extraction based on deep learning to perform local time series association on the time queue of the real-time furnace temperature value, so as to automatically recommend the operating power of the negative pressure device according to the full-time domain context causal coding representation between the local time series association features of each real-time furnace temperature after association, and obtain the operating power control instruction based on it and the real-time operating power of the negative pressure device. The present application can avoid unnecessary energy consumption through precise temperature control and efficient heat dissipation management, while reducing fluctuations in the production process, thereby significantly reducing energy costs.

[0023] Figure 1 This is a system block diagram of an energy-saving transformation system for an electric furnace for preparing silicon carbide from quartz sand according to an embodiment of the present application. Figure 2 The data flow diagram of the energy-saving transformation system of the electric furnace for preparing silicon carbide from quartz sand according to the embodiment of the present application is shown in FIG. Figure 1 and Figure 2As shown, according to the energy-saving transformation system 100 of the quartz sand electric furnace for preparing silicon carbide according to the embodiment of the present application, it includes: a real-time furnace temperature value acquisition module 110, which is used to obtain a time queue of real-time furnace temperature values ​​collected by a temperature sensor; a real-time furnace temperature value local temporal context encoding module 120, which is used to perform local temporal association implicit feature encoding and full-time domain causal context encoding of the time queue of the real-time furnace temperature value to obtain the full-time domain temporal semantic features of the furnace temperature, and obtain a mapping result based on the full-time domain temporal semantic features of the furnace temperature, and the mapping result is the recommended operating power of the negative pressure device; an operating power control instruction generation module 130, which is used to obtain an operating power control instruction based on the mapping result and the real-time operating power of the negative pressure device.

[0024] In the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand, the real-time furnace temperature value acquisition module 110 is used to obtain the time queue of the real-time furnace temperature value collected by the temperature sensor. It should be understood that the time queue of the real-time furnace temperature value collected by the temperature sensor can extract the key information reflecting the temperature change inside the electric furnace and convert it into digital time series data for subsequent algorithm processing. Specifically, at the hardware level, the system needs to reasonably arrange the temperature sensor network inside the electric furnace to achieve comprehensive monitoring of the temperature inside the furnace. These temperature sensors usually use professional equipment such as high-precision thermocouples or infrared thermometers, which can accurately sense and record the temperature values ​​of different areas inside the electric furnace. In order to ensure the reliability and coverage of data acquisition, the location of the sensor points needs to comprehensively consider factors such as the electric furnace structure, heat source distribution, and possible temperature gradients. For example, more sensors will be arranged in the core heating area and key heat dissipation parts of the electric furnace, while the density of points can be appropriately reduced in relatively stable areas. In addition, the sensor must also have good anti-interference ability to cope with electromagnetic interference or other environmental noise generated during the operation of the electric furnace. When the temperature sensor is installed, the data acquisition stage is entered next. The main task of this stage is to read the measured values ​​of each sensor regularly through the data acquisition module, and arrange these discrete data points in chronological order to form a time queue. Among them, the acquisition of real-time furnace temperature values ​​is usually built based on embedded systems or industrial controllers, with a timing trigger function, which can send query instructions to sensors within a preset time interval. For example, if the sampling frequency is set to once per second, the data acquisition module will initiate a synchronous reading operation to all sensors every second to obtain a set of values ​​reflecting the temperature distribution in the furnace at the current moment. These values ​​will then be stored in a buffer for further processing. It is worth noting that due to the complex internal environment of the electric furnace and the possible adverse conditions such as high temperature and vibration, a series of safeguards need to be taken in the data acquisition process to ensure data quality. For example, the sensor output can be compensated and corrected through a calibration algorithm to eliminate errors caused by equipment aging or external interference; at the same time, redundant design can be introduced to deploy multiple sensors at key locations to achieve data mutual verification, thereby improving the robustness of the system. In addition, in order to meet the needs of large-scale data transmission, the data acquisition module is usually equipped with a high-speed communication interface, such as Ethernet or wireless communication module, which is used to upload the collected temperature data to the central processing unit or cloud server. Next, the time queue needs to be constructed. A time queue is essentially an ordered data structure in which each element corresponds to a temperature measurement in the furnace at a specific moment. To facilitate subsequent analysis and modeling, the construction of the time queue needs to follow strict timestamp rules to ensure that the timing relationship between each data point is clear and unambiguous. For example, each temperature measurement can be bound to its acquisition time to form a "time-temperature" key-value pair.On this basis, the time queue can also be preprocessed according to actual application requirements, such as removing outliers, filling missing data, or performing smoothing filtering to improve data integrity and availability.

[0025] In the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand, the local temporal context encoding module 120 of the real-time furnace temperature value is used to perform local temporal association implicit feature encoding and full-time domain causal context encoding of the furnace temperature on the time queue of the real-time furnace temperature value to obtain the full-time domain temporal semantic features of the furnace temperature, and based on the full-time domain temporal semantic features of the furnace temperature, obtain a mapping result, which is the recommended operating power of the negative pressure device. In this way, starting from the original data, after multiple steps such as multi-level feature extraction, causal modeling, and parameter mapping, by performing multi-level and multi-dimensional analysis and modeling on the real-time collected furnace temperature data, accurate control instructions are finally generated, which can sense the temperature changes in the furnace in real time and dynamically adjust the operation strategy of the negative pressure device, thereby significantly reducing energy consumption and improving production efficiency.

[0026] Figure 3 The block diagram is a local temporal context encoding module of the real-time furnace temperature value in the energy-saving transformation system of the electric furnace for preparing silicon carbide from quartz sand according to the embodiment of the present application. Figure 3 As shown, in an embodiment of the present application, the real-time furnace temperature value local temporal context encoding module 120 includes: a real-time furnace temperature value local temporal association unit 121, which is used to perform local temporal feature extraction based on a convolutional neural network on the time queue of the real-time furnace temperature value to obtain the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector; a furnace temperature feature full time domain causal association encoding unit 122, which is used to perform furnace temperature feature full time domain context causal association semantic encoding on the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector to obtain the furnace temperature full time domain temporal feature semantic coding vector as the furnace temperature full time domain temporal semantic feature; a mapping result generation unit 123, which is used to obtain the mapping result based on the furnace temperature full time domain temporal feature semantic coding vector, and the mapping result is the recommended operating power of the negative pressure device.

[0027] In the technical solution of the present application, the local temporal sequence association unit 121 of the real-time furnace temperature value is used to extract the local temporal sequence features of the time queue of the real-time furnace temperature value based on the convolutional neural network to obtain the temporal distribution of the local temporal sequence association feature implicit coding vector of the real-time furnace temperature. It is worth mentioning that in the embodiment of the present application, the convolutional neural network is a one-dimensional convolutional neural network. It should be understood that considering that the time queue of the real-time furnace temperature value is one-dimensional time series data, the temperature values ​​at each moment are arranged in chronological order. During the operation of the electric furnace, the change of the furnace temperature often has certain correlation and regularity between adjacent time points. Based on this, in the technical solution of the present application, the local temporal sequence feature extraction based on the one-dimensional convolutional neural network is performed on the time queue of the real-time furnace temperature value to obtain the temporal distribution of the local temporal sequence association feature implicit coding vector of the real-time furnace temperature. It can be understood that the one-dimensional convolutional neural network is the application of the convolutional neural network on one-dimensional data, which is particularly suitable for processing such one-dimensional sequence data. It can effectively capture the local features and patterns in the data without changing the data dimension. Specifically, one-dimensional convolution defines a convolution kernel of a fixed size, which slides along the time axis on the input data. During each slide, the convolution kernel performs a weighted sum operation on the data points in the current window to generate a new output value. In this way, the system can gradually scan the entire time series and extract a set of numerical representations that reflect local features. By sliding the one-dimensional convolution kernel on the one-dimensional data for convolution operations, it is possible to automatically learn the local correlation of temperature data in the time dimension, focus on these local time segments, and capture the short-term trend and fluctuation of temperature. For example, the rate at which the temperature may rise or fall in certain time periods, the regularity of the peak and valley values ​​of the temperature, etc. These local features are very important for understanding the operating status of the electric furnace.

[0028] In the technical solution of the present application, the furnace temperature feature full-time domain causal association coding unit 122 is used to perform furnace temperature feature full-time domain context causal association semantic coding on the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector to obtain the furnace temperature full-time domain temporal feature semantic coding vector as the furnace temperature full-time domain temporal semantic feature. It should be understood that, considering that each real-time furnace temperature local temporal association feature implicit coding vector represents the temporal changes of the furnace temperature within a local time period, in the process of preparing silicon carbide from quartz sand, the temperature changes at different stages affect each other, and there is a complex causal relationship. For example, the steady rise in temperature in the early stage may create conditions for subsequent chemical reactions, and the temperature fluctuations in the subsequent reaction process may feedback and affect the thermal equilibrium established in the early stage. Based on this, in order to place each local temperature feature in the entire production time span for comprehensive consideration, and to dig out the global patterns and causal relationships hidden in the long-term accumulated temperature data, this application performs full-time context causal association semantic encoding of the furnace temperature feature through the time distribution of the implicit coding vector of the real-time furnace temperature local time series correlation feature to obtain the full-time domain time series feature semantic coding vector of the furnace temperature. In this way, the temperature features of each local area are integrated and sublimated to obtain the full-time domain time series feature semantic coding vector of the furnace temperature. This vector integrates the temperature information from local to global, as well as the causal relationship between them, and is a comprehensive, in-depth and semantically understandable representation of the temperature change in the furnace. It not only includes the numerical changes in temperature, but also contains the meaning and role of these changes in the entire production process. It can more accurately reflect the actual demand and change trend of the furnace temperature, thereby providing a reliable basis for automatically recommending the operating power of the negative pressure device.

[0029] Figure 4 The block diagram of the full time domain causal correlation coding unit of the furnace temperature characteristic in the energy-saving transformation system of the electric furnace for preparing silicon carbide from quartz sand according to the embodiment of the present application. Figure 4As shown, in an embodiment of the present application, the furnace temperature feature full-time domain causal association coding unit 122 includes: a real-time furnace temperature local temporal series implicit feature mining subunit 1221, which is used to perform implicit feature mining on each real-time furnace temperature local temporal series associated feature implicit coding vector in the temporal distribution of the real-time furnace temperature local temporal series associated feature implicit coding vector to obtain the temporal distribution of the real-time furnace temperature local temporal series depth implicit feature coding vector; a real-time furnace temperature topological feature matrix construction subunit 1222, which is used to construct a real-time furnace temperature semantic causal association topological feature matrix of the temporal distribution of the real-time furnace temperature local temporal series depth implicit feature coding vector; a real-time furnace temperature surface context dynamic walking subunit 1223, which is used to combine the real-time furnace temperature semantic causal association topological feature matrix and the real-time The temporal distribution of the implicit coding vector of the local temporal association feature of the furnace temperature is dynamically walked in the real-time furnace temperature graph convolution sequence to obtain the real-time furnace temperature surface context dynamic walking semantic coding vector; the real-time furnace temperature hidden layer context dynamic walking subunit 1224 is used to dynamically walk the real-time furnace temperature semantic causal association topological feature matrix and the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector in the real-time furnace temperature graph convolution sequence to obtain the real-time furnace temperature hidden layer context dynamic walking semantic coding vector; the furnace temperature full time domain time series feature fusion subunit 1225 is used to fuse the real-time furnace temperature surface context dynamic walking semantic coding vector and the real-time furnace temperature hidden layer context dynamic walking semantic coding vector to obtain the furnace temperature full time domain time series feature semantic coding vector.

[0030] In the technical solution of the present application, the real-time furnace temperature local time series implicit feature mining subunit 1221 is used to perform implicit feature mining on each real-time furnace temperature local time series associated feature implicit coding vector in the time series distribution of the real-time furnace temperature local time series associated feature implicit coding vector to obtain the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector, which is expressed as the implicit feature mining formula: ;in, is the time series distribution of the implicit encoding vector of the local time series correlation feature of the real-time furnace temperature, and are the first, second, and third in the time series distribution of the implicit encoding vector of the local time series correlation feature of the real-time furnace temperature. and The real-time furnace temperature local time series correlation feature implicit encoding vector, is the point convolutional coding, is the convolutional coding activation function, and are respectively the first, second, and third in the time series distribution of the local time series depth implicit feature encoding vector of the real-time furnace temperature. , and A real-time furnace temperature local time series deep implicit feature encoding vector, It is the time series distribution of the real-time furnace temperature local time series deep implicit feature encoding vector. It should be understood that the essence of implicit modeling is to alleviate the noise problem and obvious limitations of explicit features, and to generate a broader semantic representation by mapping each real-time furnace temperature local time series associated feature implicit encoding vector in the time series distribution of the original real-time furnace temperature local time series associated feature implicit encoding vector to the latent space, in which important information is efficiently compressed and redundant information is eliminated. In this process, important information is efficiently compressed, and redundant information is eliminated. For example, in the time series of furnace temperature changes, there may be some high-frequency fluctuations or random disturbances. Although these phenomena seem important on the surface, they do not actually reflect the essential laws of the operation state of the electric furnace. Through implicit feature mining, the system can automatically identify and filter out these irrelevant information while retaining those truly representative features.

[0031] In the technical solution of the present application, the real-time furnace temperature topological feature matrix construction subunit 1222 is used to construct the real-time furnace temperature semantic causal association topological feature matrix of the time series distribution of the local time series depth implicit feature encoding vector of the real-time furnace temperature. It should be understood that in the actual operation process of the quartz sand preparation silicon carbide electric furnace, the change in temperature is not only affected by its own state at the current moment, but may also have deep interactions with multiple time points before or after. This complex causal relationship cannot be fully described by simple local feature extraction, so it is necessary to construct a higher-level structured representation, that is, the real-time furnace temperature semantic causal association topological feature matrix.

[0032] Figure 5 The block diagram of the subunit of the real-time furnace temperature topology characteristic matrix construction in the energy-saving transformation system of the electric furnace for preparing silicon carbide from quartz sand according to the embodiment of the present application. Figure 5As shown, in an embodiment of the present application, the real-time furnace temperature topological feature matrix construction subunit 1222 includes: a real-time furnace temperature correlation factor calculation secondary subunit 1222-1, which is used to calculate the semantic causal correlation factor between any two real-time furnace temperature local temporal depth implicit feature coding vectors in the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector to obtain a real-time furnace temperature semantic causal correlation topological matrix composed of multiple real-time furnace temperature semantic causal correlation factors; a gated activation causal triggering secondary subunit 1222-2, which is used to perform gated activation causal triggering on the real-time furnace temperature semantic causal correlation topological matrix to obtain the real-time furnace temperature semantic causal correlation topological feature matrix.

[0033] In an embodiment of the present application, the real-time furnace temperature association factor calculation secondary subunit 1222-1 is used to: calculate the association matrix between any two real-time furnace temperature local time series depth implicit feature coding vectors in the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector to obtain the time series distribution of the real-time furnace temperature time series association matrix; calculate the semantic causal association factors of each real-time furnace temperature time series association matrix in the time series distribution of the real-time furnace temperature time series association matrix to obtain the real-time furnace temperature semantic causal association topological matrix composed of multiple real-time furnace temperature semantic causal association factors, the real-time furnace temperature semantic causal association factor It is related to the mean, variance, maximum value and real-time furnace temperature causal association bias value of the corresponding real-time furnace temperature time series association matrix, wherein, in response to the variance of the real-time furnace temperature time series association matrix being greater than or equal to a predetermined threshold, the weighted average of the distances between any two real-time furnace temperature local time series depth implicit feature coding vectors in the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector is used as the real-time furnace temperature causal association bias value; in response to the variance of the real-time furnace temperature time series association matrix being less than the predetermined threshold, the weighted mean of the real-time furnace temperature time series association matrix is ​​used as the real-time furnace temperature causal association bias value.

[0034] The real-time furnace temperature correlation factor calculation secondary subunit 1222-1 is expressed as a semantic causal correlation calculation formula: ;in, is matrix multiplication, yes The transposed vector of yes and The real-time furnace temperature timing correlation matrix between yes The variance of To obtain The maximum value in yes The mean of is the real-time furnace temperature causal correlation offset value, yes The corresponding real-time furnace temperature semantic causal association factor, for and The distance between yes The number of vectors in is a predetermined threshold, and is a weighted hyperparameter, and are the real-time furnace temperature semantic causal association factors at each position in the real-time furnace temperature semantic causal association topological matrix, It is the semantic causal association topological matrix of the real-time furnace temperature. It should be understood that after the implicit feature mining is completed, the model enters the semantic causal association calculation stage. The goal of this step is to derive the causal association between the two features from the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector. In the specific implementation, this process can be implemented in combination with causal modeling technology. In particular, in the technical solution of the present application, it constructs a real-time furnace temperature temporal association matrix between any two real-time furnace temperature local temporal depth implicit feature coding vectors in the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector, and uses the causal association energy measurement function to express the causal association between any two real-time furnace temperature local temporal depth implicit feature coding vectors in the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector to obtain a semantic causal association factor.

[0035] In particular, by considering low-level causal associations in complex systems as molecular-level relationships inferred based on statistical correlations, intervention predictions of causal energy can be further performed on the basis of global fine-grained statistical association representations, so as to study the fine-grained structure of causal associations and their dynamic regulation based on the high-dimensional and heterogeneous representation of causal omics. In particular, when the aggregate distribution representation of the causal graph is greater than a predetermined threshold, source data integration is biased based on the matrix graph node effect representation of the semantic causal association factor, while when the aggregate distribution representation of the causal graph is less than a predetermined threshold, condensed structural modeling can be performed directly through feature pattern integration compression. In this way, not only can the causal energy in the description system be encoded, but also the implicit causal intervention prediction results can be condensed, thereby obtaining a more efficient revelation of key causal associations.

[0036] In an embodiment of the present application, the gated activation causal triggering secondary subunit 1222-2 is used to perform gated activation causal triggering on the real-time furnace temperature semantic causal association topology matrix to obtain the real-time furnace temperature semantic causal association topology feature matrix, which is expressed as a gated activation causal triggering formula: ;in, is the real-time furnace temperature semantic causal association topological feature matrix, is a nonlinear activation function, is the normalized threshold, For Perform gated activation processing. It should be understood that after the semantic causal association factors are arranged as the real-time furnace temperature semantic causal association topological moments, the real-time furnace temperature semantic causal association topological matrix is ​​dynamically modeled using the gated activation causal trigger network. The role of the gated activation causal trigger network is to perform more detailed feature extraction and modeling of the causal topological relationship. Its core is the dynamic gating mechanism and nonlinear activation function to dynamically trigger the topological matrix representation. The gating mechanism allows the model to distinguish key causal paths in a dynamic context, thereby strengthening important associations and reducing noise interference. At the same time, the activation function in the network enhances the model's expressiveness by introducing nonlinearity, capturing high-order regularities hidden in complex causal structures. Specifically, considering that traditional linear models are often difficult to capture high-order regularities hidden in complex causal structures, the existence of activation functions enables the model to better adapt to diverse data distributions. For example, commonly used activation functions such as ReLU or Leaky ReLU can effectively avoid the gradient vanishing problem by truncating negative values ​​or introducing small slopes, thereby improving training efficiency and model performance. In the technical solution of the present application, the activation function helps the model reveal those potential, deep causal relationships through nonlinear mapping.

[0037] In the technical solution of the present application, the real-time furnace temperature surface context dynamic walking subunit 1223 is used to perform a real-time furnace temperature graph convolution sequence dynamic walking on the real-time furnace temperature semantic causal association topological feature matrix and the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector to obtain the real-time furnace temperature surface context dynamic walking semantic coding vector, which is expressed as the real-time furnace temperature surface dynamic walking formula: ;in, It is the graph convolution processing, It is the real-time furnace temperature surface context dynamic walk semantic encoding vector. It should be understood that after the gated activation causal triggering deeply optimizes the real-time furnace temperature semantic causal association topological structure, the real-time furnace temperature graph convolution sequence dynamic walk mechanism generates the real-time furnace temperature surface context dynamic walk semantic encoding vector and the real-time furnace temperature hidden context dynamic walk semantic encoding vector in a hierarchical manner through the dynamic graph representation learning based on the graph convolutional neural network (GCN). Specifically, in order to generate the real-time furnace temperature surface context dynamic walk semantic encoding vector, the real-time furnace temperature graph convolution sequence dynamic walk mechanism simulates the propagation of features in the topological structure, from local to global, and recursively aggregates the explicit semantics between nodes. Among them, the core idea of ​​the graph convolutional neural network is to capture the relationship between nodes and their impact on the overall system behavior by performing convolution operations on graph structure data. In this technical solution, the real-time furnace temperature semantic causal association topological feature matrix defines the connection relationship between nodes, and the temporal distribution of the real-time furnace temperature local temporal association feature implicit encoding vector provides the initial feature representation of each node. Through graph convolution operations, the system can propagate these initial features along the topological structure and continuously aggregate the information of adjacent nodes during the propagation process to generate new feature representations.

[0038] In the technical solution of the present application, the real-time furnace temperature hidden layer context dynamic walking subunit 1224 is used to dynamically walk the real-time furnace temperature graph convolution sequence of the real-time furnace temperature semantic causal association topological feature matrix and the real-time furnace temperature local time series depth implicit feature coding vector to obtain the real-time furnace temperature hidden layer context dynamic walking semantic coding vector, which is expressed as the real-time furnace temperature hidden layer dynamic walking formula: ;in, is the real-time furnace temperature hidden context dynamic walk semantic encoding vector. It should be understood that for the real-time furnace temperature hidden context dynamic walk semantic encoding vector, the real-time furnace temperature graph convolution sequence dynamic walk mechanism has carried out a deeper semantic exploration for implicit features. Unlike the surface context dynamic walk semantic encoding vector, the hidden context dynamic walk semantic encoding vector pays more attention to capturing deep-level feature representations, especially those high-order information and deep patterns hidden in complex systems. Hidden semantics enables the potential embedding of features, so it is necessary to pay attention to the multi-hop propagation of high-order information and the distributed decoupling of deep features during the walk process to avoid the over-smoothing phenomenon caused by the propagation of deep topological features. This hidden layer representation provides greater generalization ability for feature expression by capturing profound temporal dependencies and complex semantic patterns. Specifically, multi-hop propagation refers to gradually expanding the receptive field of nodes through multiple iterations of graph convolution operations, so that each node can perceive a wider range of contextual information. For example, in the first layer of convolution operations, each node only exchanges information with its direct neighbors; while in subsequent layers, the receptive field of nodes gradually expands to capture the influence of more indirect neighbors. In this way, the system can not only retain the details of local features, but also reveal the deep-level laws hidden in the global scope. At the same time, in order to avoid the over-smoothing phenomenon caused by the propagation of deep topological features, the dynamic walk mechanism of the real-time furnace temperature graph convolution sequence also introduces a distributed decoupling strategy. The so-called over-smoothing phenomenon means that as the number of graph convolution layers increases, the node features may gradually tend to be consistent, thereby losing their uniqueness. In this way, the distributed decoupling strategy can ensure that the features of each layer can maintain sufficient discrimination.

[0039] In the technical solution of the present application, the furnace temperature full time domain time series feature fusion subunit 1225 is used to fuse the real-time furnace temperature surface context dynamic walk semantic coding vector and the real-time furnace temperature hidden layer context dynamic walk semantic coding vector to obtain the furnace temperature full time domain time series feature semantic coding vector, which is expressed as: ;in, is the fusion weight parameter, is the semantic encoding vector of the full-time domain time series feature of the furnace temperature. It should be understood that the semantic representation of the surface layer and the hidden layer captures explicit and implicit semantic information respectively. Specifically, the surface context dynamic walk semantic encoding vector mainly focuses on the propagation of local features and the explicit relationship modeling, which can effectively reflect the temperature change law and its direct causal relationship in a short time; while the hidden context dynamic walk semantic encoding vector pays more attention to the deep feature expression, and is good at capturing long-term dependencies and high-order patterns hidden in complex systems. Considering that these two feature dimensions are complementary, feature fusion is required in the final stage. In the specific implementation, the importance distribution of surface and hidden features in the fusion process can be dynamically adjusted through strategies such as weighted summation, attention mechanism or multi-head fusion to ensure that the comprehensive features have stronger distinguishing ability and representation integrity. Among them, the weighted summation dynamically adjusts the importance distribution of the two features by introducing a learnable fusion weighting parameter. Specifically, when the surface feature is more representative at a certain moment, the model will give it a higher weight; otherwise, it will pay more attention to the hidden feature. This method can not only make full use of the advantages of the two features, but also avoid the limitations that may be brought by a single feature. The core idea of ​​the attention mechanism is to automatically calculate the importance score of each feature based on the current context information and assign weights accordingly. Through the attention mechanism, the system can adaptively adjust the degree of attention to the two features, thereby improving the robustness and generalization ability of the model. The multi-head fusion strategy is to input the surface and hidden layer features into multiple independent sub-networks for processing, and then merge the output results of each sub-network. The advantage of this method is that it can model features from multiple angles, thereby obtaining richer and more diverse representations. In this way, the system can not only better adapt to complex data distribution, but also significantly enhance the ability to express features.

[0040] In an embodiment of the present application, the mapping result generating unit 123 is used to obtain the mapping result based on the full-time domain temporal feature semantic coding vector of the furnace temperature, and the mapping result is the recommended operating power of the negative pressure device. Specifically, in an embodiment of the present application, the mapping result generating unit is used to: input the full-time domain temporal feature semantic coding vector of the furnace temperature into a decoder-based heat dissipation parameter mapper to obtain the mapping result, and the mapping result is the recommended operating power of the negative pressure device. That is, the full-time domain temporal feature semantic coding vector of the furnace temperature obtained by performing full-time domain context causal association using the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector is classified and processed to automatically recommend the operating power of the negative pressure device. It should be understood that in the production of electric furnaces for preparing silicon carbide from quartz sand, there is an intrinsic connection between the furnace temperature and the operating power of the negative pressure device. The change in the furnace temperature needs to be controlled and regulated by adjusting the operating power of the negative pressure device to maintain appropriate heat dissipation conditions. The semantic coding vector of the full-time-domain temporal series feature of the furnace temperature contains comprehensive and in-depth information about the temperature changes in the furnace. Through the decoder-based heat dissipation parameter mapper, these temperature feature information can be converted into parameters related to the operating power of the negative pressure device, thereby establishing a mapping relationship between temperature features and power. In other words, the decoder is a model component that can convert the encoded feature vector into a specific output value. The decoder-based heat dissipation parameter mapper has the ability to decode and convert the semantic information contained in the full-time-domain temporal series feature semantic coding vector of the furnace temperature. It can understand the temperature changes and requirements represented by these temperature features based on pre-trained model parameters and algorithms, and map them to appropriate negative pressure device operating power values, thereby realizing the conversion from temperature features to power parameters.

[0041] In the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand, the operating power control instruction generation module 130 is used to obtain the operating power control instruction based on the mapping result and the real-time operating power of the negative pressure device, including: inputting the mapping result and the real-time operating power of the negative pressure device into a PID-based power controller to obtain the operating power control instruction. That is, the mapping result gives the recommended operating power of the negative pressure device based on the temperature characteristics in the furnace, and the actual operating power of the negative pressure device may differ from the recommended value due to various factors (such as equipment aging, voltage fluctuations, load changes, etc.). Inputting both into the power controller at the same time allows the controller to fully understand the deviation between the ideal power and the actual power, so as to make more accurate adjustment decisions. The PID (proportional-integral-differential) controller is a classic and widely used control algorithm, which can generate a suitable control signal to adjust the system output according to the system error (i.e., the difference between the recommended power and the actual power) through the calculation of the three links of proportion, integration and differentiation. For the control of the operating power of the negative pressure device, the PID controller can effectively handle complex dynamic changes, respond quickly to power deviations, and has good stability and robustness. In this way, the mapping results and the real-time operating power are comprehensively calculated by the PID-based power controller to generate an accurate operating power control instruction. The instruction takes into account factors such as the size of the power deviation, the change trend, and the historical accumulated deviation, and can accurately guide the negative pressure device to adjust the operating power to achieve accurate control of the temperature in the furnace. It is worth mentioning that the control instruction specifically refers to the operation signal calculated by the operating power control instruction generation module based on the PID controller algorithm, which is used to guide the negative pressure device to adjust its operating power. In a specific embodiment of the present application, the control instruction is used to adjust the wind speed and air volume of the negative pressure device (induced draft fan). Among them, the proportional link in the PID controller directly generates a control signal based on the error; the integral link considers the historical accumulation of the error to eliminate the static deviation; the differential link predicts the change trend of the error, so as to respond in advance. Finally, the PID controller combines the outputs of these three links to generate an accurate control signal, so that the actual operating power of the negative pressure device gradually approaches the ideal value recommended by the mapping result.

[0042] In summary, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand based on the embodiment of the present application is explained, which obtains the time queue of the real-time furnace temperature value collected by the temperature sensor, and uses the data processing and feature extraction based on deep learning to perform local time series association on the time queue of the real-time furnace temperature value, so as to automatically recommend the operating power of the negative pressure device according to the full-time domain context causal coding representation between the local time series association features of each real-time furnace temperature after association, and obtain the operating power control instruction based on it and the real-time operating power of the negative pressure device. The present application can avoid unnecessary energy consumption through precise temperature control and efficient heat dissipation management, while reducing fluctuations in the production process, thereby significantly reducing energy costs.

[0043] As described above, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand according to the embodiment of the present application can be implemented in various terminal devices. In one example, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand can be integrated into the terminal device as a software module and / or a hardware module. For example, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand can also be one of the many hardware modules of the terminal device.

[0044] Alternatively, in another example, the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand and the terminal device may also be separate devices, and the energy-saving transformation system 100 for preparing silicon carbide electric furnace from quartz sand may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0045] Figure 6 The flowchart is a method for energy-saving transformation of a quartz sand furnace for preparing silicon carbide according to an embodiment of the present application. Figure 6 As shown, according to the energy-saving transformation method of the quartz sand electric furnace for preparing silicon carbide according to the embodiment of the present application, it includes: S110, obtaining a time queue of real-time furnace temperature values ​​collected by a temperature sensor; S120, performing local temporal association implicit feature encoding and full-time domain causal context encoding of the time queue of real-time furnace temperature values ​​to obtain the full-time domain temporal semantic features of the furnace temperature, and based on the full-time domain temporal semantic features of the furnace temperature, obtaining a mapping result, and the mapping result is the recommended operating power of the negative pressure device; S130, obtaining an operating power control instruction based on the mapping result and the real-time operating power of the negative pressure device.

[0046] Here, those skilled in the art can understand that the specific operations of each step in the energy-saving transformation method of the quartz sand preparation silicon carbide electric furnace have been referred to above. Figures 1 to 5 The description of the energy-saving transformation system of the quartz sand preparation silicon carbide electric furnace has been introduced in detail, and therefore, its repeated description will be omitted.

[0047] In summary, the energy-saving transformation method of the electric furnace for preparing silicon carbide from quartz sand based on the embodiment of the present application is explained, which obtains the time queue of the real-time furnace temperature value collected by the temperature sensor, and uses the data processing and feature extraction based on deep learning to perform local time series association on the time queue of the real-time furnace temperature value, so as to automatically recommend the operating power of the negative pressure device based on the full-time domain context causal coding representation between the local time series association features of each real-time furnace temperature after association, and obtain the operating power control instruction based on it and the real-time operating power of the negative pressure device. The present application can avoid unnecessary energy consumption through precise temperature control and efficient heat dissipation management, while reducing fluctuations in the production process, thereby significantly reducing energy costs.

[0048] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0049] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0050] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0051] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0052] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0053] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit of the technical solution of the present invention.

Claims

1. An energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand, characterized in that: include: A real-time furnace temperature value acquisition module is used to obtain a time queue of real-time furnace temperature values ​​acquired by a temperature sensor; A local temporal context encoding module for the real-time furnace temperature value is used to perform local temporal association implicit feature encoding and full-time domain causal context encoding of the furnace temperature on the time queue of the real-time furnace temperature value to obtain the full-time domain temporal semantic features of the furnace temperature, and obtain a mapping result based on the full-time domain temporal semantic features of the furnace temperature, wherein the mapping result is the recommended operating power of the negative pressure device; An operating power control instruction generating module is used to obtain an operating power control instruction based on the mapping result and the real-time operating power of the negative pressure device.

2. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 1 is characterized in that: The real-time furnace temperature value local temporal context encoding module includes: A local time series association unit for the real-time furnace temperature value is used to extract local time series features of the time queue of the real-time furnace temperature value based on a convolutional neural network to obtain a time series distribution of an implicit coding vector of a local time series association feature of the real-time furnace temperature; A furnace temperature feature full-time domain causal association coding unit is used to perform furnace temperature feature full-time domain context causal association semantic coding on the temporal distribution of the real-time furnace temperature local temporal association feature implicit coding vector to obtain a furnace temperature full-time domain temporal series feature semantic coding vector as the furnace temperature full-time domain temporal series semantic feature; A mapping result generating unit is used to obtain the mapping result based on the full-time domain temporal feature semantic coding vector of the furnace temperature, and the mapping result is the recommended operating power of the negative pressure device.

3. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 2 is characterized in that: The convolutional neural network is a one-dimensional convolutional neural network.

4. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 3 is characterized in that: The furnace temperature characteristic full time domain causal correlation coding unit includes: A real-time furnace temperature local time series implicit feature mining subunit is used to perform implicit feature mining on each real-time furnace temperature local time series associated feature implicit coding vector in the time series distribution of the real-time furnace temperature local time series associated feature implicit coding vector to obtain the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector; A real-time furnace temperature topological feature matrix construction subunit is used to construct a real-time furnace temperature semantic causal association topological feature matrix of the time series distribution of the real-time furnace temperature local time series depth implicit feature encoding vector; The real-time furnace temperature surface context dynamic walking subunit is used to dynamically walk the real-time furnace temperature graph convolution sequence of the real-time furnace temperature semantic causal association topological feature matrix and the real-time furnace temperature local temporal association feature implicit coding vector to obtain the real-time furnace temperature surface context dynamic walking semantic coding vector; A real-time furnace temperature hidden layer context dynamic walking subunit is used to dynamically walk the real-time furnace temperature graph convolution sequence of the real-time furnace temperature semantic causal association topological feature matrix and the real-time furnace temperature local time series depth implicit feature coding vector to obtain a real-time furnace temperature hidden layer context dynamic walking semantic coding vector; The furnace temperature full time domain time series feature fusion subunit is used to fuse the real-time furnace temperature surface context dynamic wandering semantic coding vector and the real-time furnace temperature hidden layer context dynamic wandering semantic coding vector to obtain the furnace temperature full time domain time series feature semantic coding vector.

5. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 4 is characterized in that: The real-time furnace temperature topology characteristic matrix construction subunit includes: A real-time furnace temperature correlation factor calculation secondary subunit is used to calculate the semantic causal correlation factor between any two real-time furnace temperature local temporal depth implicit feature coding vectors in the temporal distribution of the real-time furnace temperature local temporal depth implicit feature coding vector to obtain a real-time furnace temperature semantic causal correlation topological matrix composed of multiple real-time furnace temperature semantic causal correlation factors; The gated activation causal triggering secondary subunit is used to perform gated activation causal triggering on the real-time furnace temperature semantic causal association topology matrix to obtain the real-time furnace temperature semantic causal association topology feature matrix.

6. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 5 is characterized in that: The real-time furnace temperature correlation factor calculation secondary subunit is used to: Calculate the association matrix between any two real-time furnace temperature local time series depth implicit feature coding vectors in the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector to obtain the time series distribution of the real-time furnace temperature time series association matrix; Calculating the semantic causal association factors of each real-time furnace temperature time series association matrix in the time series distribution of the real-time furnace temperature time series association matrix to obtain the real-time furnace temperature semantic causal association topological matrix composed of multiple real-time furnace temperature semantic causal association factors, wherein the real-time furnace temperature semantic causal association factor is related to the mean, variance, maximum value and real-time furnace temperature causal association bias value of the corresponding real-time furnace temperature time series association matrix; In which, in response to the variance of the real-time furnace temperature time series association matrix being greater than or equal to a predetermined threshold, a weighted average of the distances between any two real-time furnace temperature local time series depth implicit feature coding vectors in the time series distribution of the real-time furnace temperature local time series depth implicit feature coding vector is used as the real-time furnace temperature causal association bias value; In response to the variance of the real-time furnace temperature time series association matrix being less than the predetermined threshold, a weighted mean of the real-time furnace temperature time series association matrix is ​​used as the real-time furnace temperature causal association bias value.

7. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 6 is characterized in that: The mapping result generating unit is used to input the full-time domain temporal feature semantic coding vector of the furnace temperature into a decoder-based heat dissipation parameter mapper to obtain the mapping result, and the mapping result is the recommended operating power of the negative pressure device.

8. The energy-saving transformation system for preparing silicon carbide electric furnace from quartz sand according to claim 7 is characterized in that: The operating power control instruction generating module is used to: input the mapping result and the real-time operating power of the negative pressure device into a PID-based power controller to obtain the operating power control instruction.

9. An energy-saving transformation method for preparing silicon carbide electric furnace from quartz sand, characterized in that: include: A time queue for obtaining real-time furnace temperature values ​​collected by a temperature sensor; Performing local temporal association implicit feature encoding and full-time domain causal context encoding of the furnace temperature on the time queue of the real-time furnace temperature value to obtain full-time domain temporal semantic features of the furnace temperature, and obtaining a mapping result based on the full-time domain temporal semantic features of the furnace temperature, wherein the mapping result is a recommended operating power of the negative pressure device; Based on the mapping result and the real-time operating power of the negative pressure device, an operating power control instruction is obtained.

10. The energy-saving transformation method of the electric furnace for preparing silicon carbide from quartz sand according to claim 9, characterized in that: The time queue of the real-time furnace temperature value is subjected to local time series associated implicit feature encoding and full time domain causal context encoding of the furnace temperature to obtain full time domain time series semantic features of the furnace temperature, and based on the full time domain time series semantic features of the furnace temperature, a mapping result is obtained, and the mapping result is the recommended operating power of the negative pressure device, including: Performing local time series feature extraction based on a convolutional neural network on the time queue of the real-time furnace temperature value to obtain a time series distribution of an implicit coding vector of a local time series correlation feature of the real-time furnace temperature; Performing furnace temperature feature full-time domain context causal association semantic coding on the time series distribution of the real-time furnace temperature local time series association feature implicit coding vector to obtain the furnace temperature full-time domain time series feature semantic coding vector as the furnace temperature full-time domain time series semantic feature; Based on the semantic coding vector of the full-time domain temporal feature of the temperature in the furnace, the mapping result is obtained, and the mapping result is the recommended operating power of the negative pressure device.

Images