Search

CN-121832313-B - Carbon optimization control method and system for dedusting system of steel plant based on three-in-one

CN121832313BCN 121832313 BCN121832313 BCN 121832313BCN-121832313-B

Abstract

The invention discloses a carbon optimal control method and system for a dust removal system of a steel plant based on three-stream integration, which comprises the steps of constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence, carrying out time sequence slicing on the sequences to obtain K carbon source flow subsequences, energy consumption flow subsequences and pollution flow subsequences, respectively carrying out characteristic aggregation on the subsequences until a carbon source strength coefficient, a carbon efficiency coefficient and a carbon risk coefficient are obtained, anchoring the carbon source flow subsequences, the energy consumption flow subsequences and the pollution flow subsequences corresponding to the same time sequence number, carrying out sequence coding on the subsequences until K three-stream coupling vectors are generated, constructing a three-stream coupling model for predicting a carbon state evolution track of the dust removal system in a future rated time length according to the K three-stream coupling vectors, and carrying out carbon optimal control on the steel plant based on the three-stream coupling model.

Inventors

  • XU QIAN
  • ZHAO GUANGYUAN
  • WANG FUJIANG

Assignees

  • 柏美智慧科技(上海)股份有限公司

Dates

Publication Date
20260512
Application Date
20260313

Claims (10)

  1. 1. The carbon optimization control method for the dust removal system of the steel plant based on three-in-one is characterized by comprising the following steps: s1, constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence of the dust removal system in a historical observation period; S2, carrying out time sequence slicing on the carbon source flow sequence, the energy consumption flow sequence and the pollution flow sequence to respectively obtain K carbon source flow subsequences, energy consumption flow subsequences and pollution flow subsequences; The carbon source flow subsequence, the energy consumption flow subsequence and the pollution flow subsequence of the same slice share the same time sequence number; S3, respectively performing characteristic aggregation on the carbon source subsequence, the energy consumption subsequence and the polluted stream subsequence with the same time sequence number until K carbon source intensity coefficients, carbon efficiency coefficients and carbon risk coefficients corresponding to the same time sequence number are obtained; Wherein the subsequence comprises an instantaneous carbon source vector, an instantaneous energy consumption vector and an instantaneous pollution vector in continuous J sampling windows; S4, anchoring the carbon source subsequence, the energy consumption flow subsequence and the pollution flow subsequence corresponding to the same time sequence number in the K carbon source subsequences, the energy consumption flow subsequence and the pollution flow subsequence with the same time sequence number; S5, carrying out sequence coding on the carbon source sub-sequence, the energy consumption sub-sequence and the pollution stream sub-sequence corresponding to the same time sequence number until K three-stream coupling vectors are generated, wherein each three-stream coupling vector inherits the corresponding time sequence number; S6, constructing a three-stream coupling model for predicting a carbon state evolution track of the dust removal system in a future rated time length according to the K three-stream coupling vectors; And S7, performing carbon optimization control on the steel plant based on the three-stream coupling model.
  2. 2. The three-in-one-based carbon optimization control method for a dust removal system of a steel plant according to claim 1, wherein constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence of the dust removal system in a historical observation period comprises: S1-1, acquiring instantaneous state parameters in M sampling windows, wherein the instantaneous state parameters comprise instantaneous carbon source parameters, instantaneous energy consumption parameters and instantaneous pollution parameters; S1-2, characterizing instantaneous state parameters to generate M instantaneous state features corresponding to each sampling window, wherein the instantaneous state features comprise M instantaneous carbon source features, M instantaneous energy consumption features and M instantaneous pollution features; S1-3, performing feature stitching on the instantaneous state features of each sampling window until instantaneous carbon source vectors, instantaneous energy consumption vectors and instantaneous pollution vectors corresponding to M sampling windows are obtained; S1-4, arranging the instantaneous carbon source vector, the instantaneous energy consumption vector and the instantaneous pollution vector in sequence according to the time sequence of the sampling window of the instantaneous carbon source vector, the instantaneous energy consumption vector and the instantaneous pollution vector, and constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence of the dust removal system in a historical observation period.
  3. 3. The three-in-one-based carbon optimization control method for the dust removal system of the steel plant according to claim 2, wherein the step of obtaining the instantaneous state parameters in the M sampling windows comprises the following steps: s1-1-1, establishing a sliding sampling window on an operation time axis of a dust removal system of a steel plant; s1-1-2, acquiring instantaneous state parameters in a sampling window by category based on a preset parameter label, wherein the instantaneous state parameters comprise N instantaneous carbon source parameters, N instantaneous energy consumption parameters and N instantaneous pollution parameters; S1-1-3, sliding a sampling window with a standard sampling step length until the instantaneous state parameters in M sampling windows are continuously acquired.
  4. 4. The three-in-one-based carbon optimization control method for the dust removal system of the steel plant according to claim 2, wherein the method is characterized by carrying out characterization on instantaneous state parameters to generate M instantaneous state features corresponding to each sampling window, and comprises the following steps: s1-2-1, determining a target category parameter in instantaneous carbon source parameters, instantaneous energy consumption parameters and instantaneous pollution parameters; s1-2-2, anchoring a maximum state parameter and a minimum state parameter in M instantaneous state parameters of target class parameters; S1-2-3, carrying out difference between the maximum state parameter and the minimum state parameter to obtain total state deviation; s1-2-4, selecting a current state parameter in M instantaneous state parameters; s1-2-5, carrying out difference between the current state parameter and the minimum state parameter to obtain current state deviation; S1-2-6, carrying out ratio operation on the total state deviation and the current state deviation to generate an instantaneous state characteristic corresponding to the current state parameter; s1-2-7, traversing M instantaneous state parameters until the instantaneous state characteristics corresponding to M sampling windows are obtained.
  5. 5. The three-in-one-based carbon optimization control method for a dust removal system of a steel plant according to claim 4, wherein the characteristic polymerization step of the carbon source intensity coefficient comprises: a1, selecting a target carbon source subsequence corresponding to the current time sequence number from the K carbon source subsequences with the same time sequence number; A2, calculating vector modular lengths of all components component by component in continuous J instantaneous carbon source vectors contained in the target carbon source subsequence; A3, carrying out accumulation operation on vector modular length of all instantaneous carbon source vectors in the target carbon source subsequence to obtain carbon source accumulation total corresponding to the subsequence; A4, performing product operation on the accumulated total amount of the carbon sources and a preset carbon emission factor of the unit material to generate a carbon source intensity coefficient corresponding to the current time sequence number; a5, traversing the K carbon source subsequences with the same time sequence number, and repeatedly generating the carbon source intensity coefficients until the carbon source intensity coefficients corresponding to the K time sequence numbers are obtained.
  6. 6. The three-in-one based carbon optimization control method for a dust removal system of a steel plant according to claim 4, wherein the characteristic polymerization step of the carbon efficiency coefficient comprises: B1, anchoring a target energy consumption flow sub-sequence and a target pollution flow sub-sequence corresponding to the same time sequence number in the K energy consumption flow sub-sequences and the pollution flow sub-sequences with the same time sequence number; b2, extracting fan power components from continuous J instantaneous energy consumption vectors contained in the target energy consumption flow sub-sequence, and performing time integral operation to obtain an energy consumption total value corresponding to the sub-sequence; B3, extracting a discharge concentration component and a pipe network air quantity component from continuous J instantaneous pollution vectors contained in the target pollution flow subsequence, calculating the product of the discharge concentration component and the pipe network air quantity component, and performing time integral operation to obtain the dust removal total amount corresponding to the subsequence; B4, taking the total energy consumption value as a dividend, taking the total dust removal amount as a divisor, and carrying out ratio operation to generate a carbon efficiency coefficient corresponding to the current time sequence number; And B5, traversing the K subsequence pairs with the same time sequence number, and repeatedly generating the carbon efficiency coefficient until the carbon efficiency coefficients corresponding to the K time sequence numbers are obtained.
  7. 7. The three-in-one-based carbon optimization control method for a dust removal system of a steel plant according to claim 4, wherein the characteristic aggregation step of the carbon risk coefficient comprises the following steps: C1, selecting a target pollution stream subsequence corresponding to the current time sequence number from the K pollution stream subsequences with the same time sequence number; c2, presetting an environment-friendly emission threshold, and anchoring emission concentration components component by component in continuous J instantaneous pollution vectors contained in the target pollution flow subsequence; C3, carrying out difference calculation on the emission concentration component and the environment-friendly emission threshold value, screening out a difference value larger than zero as an instantaneous exceeding amplitude, and setting the difference value smaller than or equal to zero as zero; C4, performing accumulation operation on all instantaneous exceeding amplitudes in the target pollution flow subsequence, and multiplying the accumulated instantaneous exceeding amplitudes by the time length of a sampling window to obtain a exceeding total area corresponding to the subsequence; c5, mapping the out-of-standard total area into a carbon risk coefficient corresponding to the current time sequence number; And C6, traversing the K polluted stream subsequences with the same time sequence number, and repeatedly generating carbon risk coefficients until the carbon risk coefficients corresponding to the K time sequence numbers are obtained.
  8. 8. The three-in-one-based carbon optimization control method for the dust removal system of the steel plant according to claim 7, wherein the carbon source subsequence, the energy consumption subsequence and the contaminated stream subsequence corresponding to the same sequence number are subjected to sequence coding until K three-stream coupling vectors are generated, and the method comprises the following steps: s5-1, anchoring a target carbon source subsequence, a target energy consumption stream subsequence and a target pollution stream subsequence corresponding to the same time sequence number K in a K subsequence set with the same time sequence number; S5-2, aligning sampling windows of the target carbon source flow subsequence, the target energy consumption flow subsequence and the target pollution flow subsequence respectively; S5-3, performing channel splicing on the three subsequences aligned with the sampling window, and constructing a three-stream fusion sequence under the time sequence number; s5-4, carrying out average pooling on each channel of the three-stream fusion sequence, extracting a fixed dimension vector representing the three-stream comprehensive state under the time sequence number, and defining the fixed dimension vector as a three-stream coupling vector of the time sequence number k; s5-5, marking the time sequence number k on the generated three-stream coupling vector so as to inherit the time sequence representation of the three-stream coupling vector; S5-6, traversing K identical time sequence numbers, and repeatedly generating three-stream coupling vectors until K three-stream coupling vectors with independent time sequence numbers are generated.
  9. 9. The three-in-one-based carbon optimization control method for the dust removal system of the steel plant, which is disclosed in claim 8, is characterized in that a three-stream coupling model for predicting the carbon state evolution track of the dust removal system in the future rated time length is constructed according to K three-stream coupling vectors, and comprises the following steps: S6-1, anchoring target input vectors and supervision labels pair by pair along the time sequence numbering direction of the three-stream coupling vectors; The supervision tag is characterized in that a carbon source intensity coefficient, a carbon efficiency coefficient and a carbon risk coefficient are associated with a three-stream coupling vector of a target input vector distance rated prediction duration P along a time sequence numbering direction; s6-2, pairing the target input vector and the supervision label, and constructing a carbon track sample for carbon state evolution track prediction until G carbon track samples are obtained, wherein G=K-P, and G is smaller than K; S6-3, inputting G monitoring samples into a time sequence prediction model for iterative monitoring training, and generating a three-stream coupling model for predicting the carbon state evolution track of the dust removal system in the future rated time.
  10. 10. Carbon optimization control system of dust pelletizing system of iron and steel plant based on three-stream unification, its characterized in that includes: A flow sequence construction unit for constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence of the dust removal system in a historical observation period; The subsequence slicing unit is used for performing time sequence slicing on the carbon source flow sequence, the energy consumption flow sequence and the pollution flow sequence to respectively obtain K carbon source flow subsequences, energy consumption flow subsequences and pollution flow subsequences; The carbon source flow subsequence, the energy consumption flow subsequence and the pollution flow subsequence of the same slice share the same time sequence number; the characteristic aggregation unit is used for respectively carrying out characteristic aggregation on the carbon source subsequence, the energy consumption subsequence and the pollution flow subsequence with the same time sequence number until K carbon source intensity coefficients, carbon efficiency coefficients and carbon risk coefficients corresponding to the same time sequence number are obtained; Wherein the subsequence comprises an instantaneous carbon source vector, an instantaneous energy consumption vector and an instantaneous pollution vector in continuous J sampling windows; A subsequence anchoring unit, configured to anchor, in K carbon source subsequences, energy consumption flow subsequences, and pollution flow subsequences with the same time sequence number, a carbon source subsequence, an energy consumption flow subsequence, and a pollution flow subsequence corresponding to the same time sequence number; The vector coding unit is used for carrying out sequence coding on the carbon source subsequence, the energy consumption subsequence and the pollution stream subsequence corresponding to the same time sequence number until K three-stream coupling vectors are generated, wherein each three-stream coupling vector inherits the corresponding time sequence number; the coupling model construction unit is used for constructing a three-stream coupling model for predicting the carbon state evolution track of the dust removal system in the future rated time length according to the K three-stream coupling vectors; and the carbon optimization control unit is used for carrying out carbon optimization control on the steel plant based on the three-stream coupling model.

Description

Carbon optimization control method and system for dedusting system of steel plant based on three-in-one Technical Field The invention relates to carbon optimal control of a dust removal system, in particular to a carbon optimal control method and a carbon optimal control system of a dust removal system of a steel plant based on three-in-one. Background In the control logic of the existing dust removing system of the steel plant, the instantaneous carbon source parameters (such as tapping frequency and feeding beat) and the instantaneous energy consumption parameters are difficult to be strictly matched on a time axis, so that the production low-load period is difficult to identify, the phenomenon that a fan still keeps high-power operation and is accompanied by a large amount of carbon sources is frequently caused, meanwhile, the traditional method only pays attention to whether the instantaneous pollution parameters exceed a limit value or not, ignores the accumulation effect of the exceeding amplitude and duration time, and cannot calculate the carbon risk coefficient representing the comprehensive environmental protection risk, and more importantly, the prior art is difficult to deduce future carbon state evolution tracks due to lack of the excavation of the coupling rules among the historical carbon source sequences, the energy consumption flow sequences and the pollution flow sequences, and the factors cause the defect that the dust removing system of the steel plant commonly has three-stream splitting, namely the carbon source (production beat), the energy consumption flow (fan energy consumption) and the pollution flow (emission state) are mutually independent in the operation process, and cannot form a collaborative optimization closed loop. The lag and splitting control mode ensures that the system can not generate a feedforward compensation instruction before the discharge exceeds the standard, and can not dynamically adjust the frequency of the fan when the system is prejudged to be in an inefficient working condition, so that the dust removal system is in a high-carbon consumption and high-risk passive running state for a long time, and the requirement of refined carbon management and control in the steel industry is difficult to meet. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a carbon optimization control method and a carbon optimization control system for a dust removal system of a steel plant based on three-in-one, and the technical problems in the background art are solved by introducing three-stream coupling vectors. In order to achieve the above purpose, the invention is realized by the following technical scheme: In a first aspect, the invention provides a carbon optimization control method for a dust removal system of a steel plant based on three-in-one, which comprises the following steps: s1, constructing a carbon source flow sequence, an energy consumption flow sequence and a pollution flow sequence of the dust removal system in a historical observation period; S2, carrying out time sequence slicing on the carbon source flow sequence, the energy consumption flow sequence and the pollution flow sequence to respectively obtain K carbon source flow subsequences, energy consumption flow subsequences and pollution flow subsequences; The carbon source flow subsequence, the energy consumption flow subsequence and the pollution flow subsequence of the same slice share the same time sequence number; S3, respectively performing characteristic aggregation on the carbon source subsequence, the energy consumption subsequence and the polluted stream subsequence with the same time sequence number until K carbon source intensity coefficients, carbon efficiency coefficients and carbon risk coefficients corresponding to the same time sequence number are obtained; Wherein the subsequence comprises an instantaneous carbon source vector, an instantaneous energy consumption vector and an instantaneous pollution vector in continuous J sampling windows; S4, anchoring the carbon source subsequence, the energy consumption flow subsequence and the pollution flow subsequence corresponding to the same time sequence number in the K carbon source subsequences, the energy consumption flow subsequence and the pollution flow subsequence with the same time sequence number; S5, carrying out sequence coding on the carbon source sub-sequence, the energy consumption sub-sequence and the pollution stream sub-sequence corresponding to the same time sequence number until K three-stream coupling vectors are generated, wherein each three-stream coupling vector inherits the corresponding time sequence number; S6, constructing a three-stream coupling model for predicting a carbon state evolution track of the dust removal system in a future rated time length according to the K three-stream coupling vectors; And S7, performing carbon optimization control on the steel plant based on the t