Energy consumption control method and system applied to annealing furnace

Abstract

The invention discloses an energy consumption control method and system applied to an annealing furnace, and relates to the technical field of annealing furnace control, and the method comprises the steps: building a hot area state matrix through determining the position information of a plurality of hot areas; constructing a coupling strength model of the adjacent hot areas, calculating the coupling strength between the adjacent hot areas, and constructing a coupling strength standard matrix; constructing a training sample and fitting a first compensation factor formula based on a least square method; counting an adjacent hot area set of each target hot area, calculating a thermal disturbance influence weight according to the coupling strength, collecting temperature change values of the adjacent hot areas, and constructing a second compensation factor; and real-time operation parameters of the annealing furnace are obtained and input into the temperature prediction model, a first compensation factor and a second compensation factor are calculated, a temperature control set value executed by the central control system is calculated, and the energy-saving level, the temperature control stability and the process adaptation capacity of the annealing furnace system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of annealing furnace control, and in particular to an energy consumption control method and system applied to an annealing furnace. Background Art

[0002] Annealing furnaces are core equipment in modern industrial heat treatment processes. Current mainstream annealing furnace temperature control systems generally use PID control strategies based on single-hot zone feedback or fixed heating curves set based on experience. This model ignores the thermal coupling between the various hot zones within the furnace and fails to provide coordinated control based on the overall system state, resulting in low thermal energy utilization, slow temperature response, and high energy consumption. In addition, as in the invention patent application with Chinese patent publication number CN104962727A, when the deviation between the actual strip temperature obtained based on the furnace working condition information and the production line status information and the set strip target temperature exceeds the set range, the heating section furnace temperature value is reset and a furnace control message is generated and sent to the basic automation equipment to control the furnace temperature. Although this method takes into account the thermal coupling relationship between the various hot zones within the furnace body, it has hysteresis and does not consider the interference between adjacent hot zones. Therefore, a method for controlling the energy consumption of the annealing furnace is needed that can combine the hot zone status and extract and predict the temperature under complex thermal disturbance conditions, and improve the energy efficiency, adaptability and robustness of the overall control system through systematic data modeling, thereby achieving energy savings. Summary of the Invention

[0003] The object of the present invention is to provide an energy consumption control method and system for an annealing furnace to solve the problems raised in the prior art.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions: an energy consumption control method for an annealing furnace, the method comprising: Step S100: Obtain an internal structural diagram of the annealing furnace, determine the location information of several hot zones and mark them, and arrange multiple sensors inside each hot zone; collect operating parameters of each hot zone according to preset collection time points, and construct a hot zone state matrix; Step S200: Based on the hot zone state matrix and the spatial arrangement of the hot zones, a coupling strength model of adjacent hot zones is constructed, the coupling strength between each adjacent hot zone is calculated, and a coupling strength standard matrix is ​​constructed; Step S300: Compare the predicted temperature of each hot zone with the temperature in the historical operation record, calculate the average value of the temperature control deviation, combine the average difference and variance of the coupling strength of adjacent hot zones, construct a training sample and fit the first compensation factor formula based on the least squares method; Step S400: Counting the adjacent hot zone sets of each target hot zone, calculating the thermal disturbance influence weight according to the coupling strength, collecting the temperature change values ​​of the adjacent hot zones, and constructing a second compensation factor; Step S500: Obtain the real-time operating parameters of the annealing furnace and input them into the temperature prediction model, calculate the real-time predicted temperature, combine the real-time temperature control deviation average value, the real-time average difference of the coupling strength and the real-time variance parameter, input them into the first compensation factor and the second compensation factor formula respectively, calculate the first compensation factor and the second compensation factor, and calculate the temperature control set value executed by the central control system, send the temperature control set value to the annealing furnace controller through the command interface, and adjust the heater power output of the corresponding hot zone.

[0005] Furthermore, step S100 includes: Step S101: Obtain an internal structural diagram of the annealing furnace, determine the location information of each hot zone based on the internal structural diagram and mark it, arrange several sensors inside each hot zone, and monitor the operating parameters of each hot zone according to preset collection time points; Step S102: Obtain the historical operation records of each hot zone, collect the operating parameters corresponding to each collection time point in the historical operation records, obtain the temperature of the heated object, and calculate the temperature difference between the collection time point and the previous collection time point. , calculate the power feedback value at the acquisition time point according to the following formula: ; Where A1 represents the power feedback value, D represents the material mass, E represents the IoT specific heat capacity, and y represents the time difference between the acquisition time point and the previous acquisition time point. The power command value at the acquisition time point is calculated according to the following formula: ; Where A2 represents the power command value, B represents the average current between the acquisition time point and the previous acquisition time point, and C represents the average voltage between the acquisition time point and the previous acquisition time point. The load state at the acquisition time point is calculated according to the following formula: ; Where A represents the load state at the acquisition time point, the temperature monitored at the acquisition time point is F, and the temperature and load state are combined to construct the hot zone state variable S = (A, F); Step S103: aggregating the state variables generated by each hot zone at multiple acquisition time points to form a hot zone state matrix containing a time series, and uploading and storing the state matrix to a central control system; Step S104: Obtain the historical hot zone state matrix in the central control system, use the load state sequence and temperature of the historical hot zone state matrix as training data, and construct a time series model for temperature prediction through a sliding window regression model as F t =f(A t-n , A t-n+1 ,...,A t ), the time series prediction model takes the load status of multiple consecutive sampling time points as input and the temperature at the tth sampling time point as output, where F t It is expressed as the temperature at the tth acquisition time point, A t-n , A t-n+1 ,...,A t They are respectively represented as the load status at the tn, t-n+1, ​​..., t acquisition time points, where n is the window length; The system implements dynamic modeling and structured representation of hot zone status, constructing state variables by combining load status and temperature, enabling the system to accurately perceive the operating status of each hot zone. The introduction of a sliding window prediction model can accurately capture the temporal correlation between load status and temperature changes. Compared with traditional prediction methods based on single points or empirical rules, this improves temperature prediction accuracy and provides data support for subsequent energy consumption optimization control. Constructing a hot zone state matrix and uploading it to the central system facilitates distributed sensing and centralized intelligent control, providing the data foundation for subsequent coupled modeling, compensation control, and other steps. The model is data-driven, avoiding reliance on manual settings or empirical adjustments. It has good scalability and adaptability, and is suitable for deployment and application of annealing furnace equipment of various specifications.

[0006] Furthermore, step S200 includes: Step S201: Based on the historical hot zone state matrix of step S104 and in combination with the spatial arrangement relationship of each hot zone in the annealing furnace, a coupling relationship set of adjacent hot zones is determined, and the temperature and load state of each hot zone in the adjacent hot zones are extracted as the input feature set of the coupling strength model; Step S202: For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of the adjacent hot zones. The calculation formula of the coupling strength model is: ; Among them, G 12 It represents the coupling strength of hot zone 2 to hot zone 1, F1 and F2 represent the temperatures of adjacent hot zones 1 and 2 respectively, A1 and A2 represent the load states of adjacent hot zones 1 and 2 respectively, e1 and e2 represent positive constants to avoid division by zero error, and a1 and a2 represent the weights of temperature and load state respectively; Step S203: Summarize the coupling strength of each adjacent hot zone and calculate the average coupling strength to construct a coupling strength standard matrix of adjacent hot zones. The coupling strength standard matrix is ​​expressed as ; Where H represents the coupling strength standard matrix, G ij It is expressed as the average coupling strength of the coupling strength of hot zone j to hot zone i, N is the total number of hot zones, and the diagonal elements in the coupling strength standard matrix are set to 0, and the elements of non-adjacent hot zones are 0; The coupling strength standard matrix is ​​an N×N sparse matrix, in which the diagonal elements are always set to 0, indicating that the hot zone has no coupling effect on itself. The coupling strength between non-adjacent hot zones is recorded as 0, and only the non-zero coupling values ​​of hot zones with spatial adjacency are retained. This serves as the basic parameter table for subsequent compensation control and disturbance assessment, and is used to support refined power regulation and temperature control setting optimization of the hot zones. The system achieves quantitative modeling of the coupling relationship between thermal zones. Instead of relying on manual experience, it calculates the coupling strength based on temperature and load differences, improving the objectivity and adaptability of the model. The introduction of a coupling strength standard matrix effectively supports the evaluation and coordinated control of thermal zone disturbances, providing high-reliability basic data support for subsequent dual compensation factor modeling; The coupling modeling fully considers the actual spatial layout characteristics of the annealing furnace and only establishes dynamic coupling relationships for adjacent hot zones, avoiding redundant calculations and misleading interference, and improving control accuracy and response efficiency. The coupling formula structure is clear, the physical meaning of the parameters is clear, and it is easy to calibrate and deploy according to actual equipment parameters, with good industrial implementation capabilities and system portability; By characterizing the thermal disturbance path through sparse coupling matrix, the modeling transition from "local thermal control" to "global coordination" is achieved, providing a theoretical basis for efficient coordinated temperature control of multiple thermal zones.

[0007] Furthermore, step S300 includes: Step S301: Preset a number of consecutive days as a training cycle, obtain the operation records of the annealing furnace, input the operation parameters collected at each collection time point in a certain operation record into the time series model for temperature prediction, and obtain the predicted temperature corresponding to each collection time point; Step S302: Obtain the temperature and load state corresponding to each hot zone in the annealing furnace at different acquisition time points, obtain a hot zone state matrix for each hot zone, calculate the coupling strength of each adjacent hot zone based on the coupling strength model, generate a coupling strength matrix, subtract the coupling strength matrix from the coupling strength standard matrix, obtain the coupling strength difference at each acquisition time point, summarize it, and calculate the average difference and variance of the coupling strength; Step S303: Summarize the predicted temperature and temperature corresponding to a certain hot zone at each acquisition time point, and calculate the average temperature control deviation according to the following formula: Among them, Fp represents the average value of temperature control deviation, F' d and F d They are the predicted temperature and temperature corresponding to the dth acquisition time point, respectively, and b is the total number of acquisition time points; Step S304: obtaining the temperature correction value applied by the central control system to each hot zone at different sampling time points, summing up the temperature correction values ​​at all sampling time points, calculating the average temperature correction value and setting it as the first compensation factor; Step S305: Combine the average difference and variance of the coupling strength of a certain adjacent hot zone with the average temperature control deviation of the target hot zone of the adjacent hot zone to construct a training sample of the hot zone compensation factor. Summarize the training samples of the first compensation factor corresponding to all operation records of the hot zone. Take the training samples as input and the first compensation factor as output to establish the formula of the first compensation factor as follows: ; Wherein, Q1 represents the first compensation factor, L1 and L2 represent the average difference and variance of the coupling strength, respectively, k1, k2, and k3 represent the weights of the average difference, variance, and average value of the coupling strength deviation, respectively, and k4 represents the deviation value of the first compensation factor formula. The least squares fitting is performed based on the training samples to obtain the values ​​of k1, k2, k3, and k4, and input them into the first compensation factor formula to establish the final first compensation factor formula; A quantitative mapping relationship between temperature control deviation and coupling disturbance was established. By collaboratively modeling the coupling strength deviation and the actual temperature control deviation, the impact of the coupling disturbance on the control effect of the target hot zone was accurately evaluated. Implemented an adaptive compensation adjustment modeling mechanism, no longer relying on fixed empirical coefficients for parameter adjustment, and improved the adaptability and universality of the compensation factor; The least squares fitting technology is used to ensure strong model convergence and interpretability. The weight parameters have physical meanings, which facilitates actual industrial deployment and subsequent updates and maintenance. Training is based on historical operating data of thermal zones, and it has personalized modeling capabilities. It can establish differentiated compensation strategies for different thermal zones to improve the accuracy of energy efficiency control. The introduction of the first compensation factor effectively reduces the deviation between temperature prediction and actual control, and further enhances the dynamic response capability and stability of the system in strong coupling and multi-disturbance scenarios.

[0008] Furthermore, step S400 includes: Step S401: Count all adjacent hot zone sets of a target hot zone, collect the coupling strength of each adjacent hot zone to the target hot zone in the coupling strength standard matrix, and calculate the thermal disturbance influence weight of the adjacent hot zone to the target hot zone according to the following formula: ; Among them, q r It is expressed as the thermal disturbance influence weight of the adjacent hot zone r on the target hot zone g, G rg It is expressed as the coupling strength of the adjacent hot zone r to the target hot zone g, G ng It is expressed as the coupling strength of the adjacent hot zone n to the target hot zone g, and m is the total number of adjacent hot zones; Step S402: Obtain the temperature change value of each adjacent hot zone between two consecutive acquisition time points, and calculate the second compensation factor of the target hot zone according to the following second compensation factor formula: ; Among them, Q2 represents the second compensation factor, q ng It is expressed as the weight of the thermal disturbance influence of the adjacent hot zone n on the target hot zone g, Expressed as the temperature change value of adjacent hot zone n; The second compensation factor dynamically reflects the temperature fluctuation trend of adjacent hot zones, featuring strong real-time performance and fast response, improving the system's adaptability to sudden coupling disturbances. The thermal disturbance weight is derived based on a standard coupling strength matrix, with a reasonable weight distribution that matches the actual coupling degree between hot zones, ensuring the calculation results have engineering significance. The first compensation factor is based on historical statistical modeling, while the second compensation factor is based on real-time disturbance trend modeling. The two complement each other to build a complete disturbance perception and compensation framework. The thermal disturbance weight is extracted based on the coupling matrix and is applicable to any hot zone arrangement structure and scale, with good portability and scalability. In an industrial environment with multiple hot zones, strong coupling, and frequent disturbances, the second compensation factor significantly reduces the temperature control error fluctuation, effectively improving the stability and energy efficiency of the annealing furnace operation.

[0009] Furthermore, step S500 includes: Step S501: Acquire the real-time operating parameters of the annealing furnace, input them into the time series model of temperature prediction, calculate the real-time predicted temperature corresponding to each acquisition time point, collect the real-time temperature at the acquisition time point, calculate the real-time temperature control deviation average value, calculate the real-time average difference and real-time variance of the coupling strength, and input them into the first compensation factor formula to calculate the first compensation factor; Step S502: obtaining the real-time temperature change value of each adjacent hot zone between two consecutive acquisition time points, and inputting it into the second compensation factor formula to calculate the second compensation factor; Step S503: Obtain the target temperature setting value preset by the central control system as F s , and calculate the temperature setpoint executed by the central control system according to the following formula: ; Among them, F' s The temperature setting value is represented as a temperature setting value executed by the central control system, and the temperature setting value executed by the central control system is sent to the annealing furnace controller through the temperature control command interface, and the heater power output of the corresponding hot zone is adjusted according to the temperature setting value; A dynamic adaptive adjustment mechanism for control instructions is implemented, and the set temperature is calculated based on the three-dimensional fusion of real-time prediction, historical deviation, and disturbance trend, breaking through the traditional constant set value control method and improving the control sensitivity. It integrates two types of compensation factors to cover multi-scale disturbance responses. The first compensation factor models statistical deviations, while the second compensation factor models disturbance trends. The two factors work together to dynamically adapt the setpoint to the system's operating status. Reduce the steady-state deviation and dynamic overshoot of the temperature control system caused by coupling disturbance, improve temperature control stability and response speed, and significantly optimize the temperature control curve; Enhance the system's anti-interference ability in highly complex thermal fields, making it suitable for industrial annealing scenarios with multiple hot zones and high-frequency disturbances, and improving product consistency and process stability; It realizes the transformation of temperature control from "fixed value driven" to "model driven", and has good industrial promotion and engineering practical value.

[0010] In order to better implement the above method, an energy consumption control system for an annealing furnace is also proposed. The system includes a hot zone state module, a coupling strength module, a first compensation factor module, a second compensation factor module and a real-time control module; Hot zone status module: Obtain the internal structure diagram of the annealing furnace, determine the location information of several hot zones and mark them, and arrange multiple sensors inside each hot zone; collect the operating parameters of each hot zone according to the preset collection time point, and construct the hot zone status matrix; Coupling strength module: Based on the hot zone state matrix and the spatial arrangement of the hot zones, a coupling strength model of adjacent hot zones is constructed, the coupling strength between each adjacent hot zone is calculated, and a coupling strength standard matrix is ​​constructed; First compensation factor module: compares the predicted temperature of each hot zone with the temperature in the historical operation record, calculates the average temperature control deviation, combines the average difference and variance of the coupling strength of adjacent hot zones, constructs a training sample, and fits the first compensation factor formula based on the least squares method; Second compensation factor module: Counts the adjacent hot zone sets of each target hot zone, calculates the thermal disturbance influence weight according to the coupling strength, collects the temperature change values ​​of the adjacent hot zones, and constructs the second compensation factor; Real-time control module: obtains the real-time operating parameters of the annealing furnace and inputs them into the temperature prediction model, calculates the real-time predicted temperature, combines the real-time temperature control deviation average value, the real-time average difference of the coupling strength and the real-time variance parameter, inputs them into the first compensation factor and the second compensation factor formula respectively, calculates the first compensation factor and the second compensation factor, and calculates the temperature control set value executed by the central control system, sends the temperature control set value to the annealing furnace controller through the command interface, and adjusts the heater power output of the corresponding hot zone.

[0011] Furthermore, the hot zone state module includes a hot zone state variable unit and a temperature prediction model unit: Hot zone state variable unit: obtains an internal structural diagram of the annealing furnace, determines the location information of each hot zone based on the internal structural diagram and marks it, arranges several sensors inside each hot zone, and monitors the operating parameters of each hot zone according to preset collection time points; obtains the historical operation records of each hot zone, collects the operating parameters corresponding to each collection time point in the historical operation records, calculates the load state at the collection time point, and combines the temperature and load state to construct the hot zone state variable; Temperature prediction model unit: The state variables generated by each hot zone at multiple acquisition time points are aggregated to form a hot zone state matrix containing a time series, and the state matrix is ​​uploaded and stored in the central control system; the historical hot zone state matrix in the central control system is obtained, and the load state sequence and temperature of the historical hot zone state matrix are used as training data. A time series model for temperature prediction is constructed through a sliding window regression model. The time series prediction model takes the load state at multiple consecutive sampling time points as input and the temperature at the acquisition time point as output.

[0012] Furthermore, the coupling strength module includes a coupling strength calculation unit and a coupling strength standard matrix establishment unit: Coupling strength calculation unit: Based on the historical hot zone state matrix and the spatial arrangement relationship of each hot zone in the annealing furnace, the coupling relationship set of adjacent hot zones is determined. The temperature and load status of each hot zone in the adjacent hot zones are extracted as the input feature set of the coupling strength model. For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of the adjacent hot zones. Establish a coupling strength standard matrix unit: summarize the coupling strength of each adjacent hot zone and calculate the average coupling strength to construct a coupling strength standard matrix of adjacent hot zones, in which the diagonal elements are set to 0 and the elements of non-adjacent hot zones are 0.

[0013] Furthermore, the first compensation factor module includes a training sample construction unit and a first compensation factor formula establishment unit: Construct a training sample unit: preset several consecutive days as the training cycle, obtain the operation record of the annealing furnace, input the operation parameters collected at each collection time point in a certain operation record into the time series model of temperature prediction, and obtain the predicted temperature corresponding to each collection time point; obtain the temperature and load state corresponding to each hot zone in the annealing furnace at different collection time points, obtain the hot zone state matrix of each hot zone, calculate the coupling strength of each adjacent hot zone based on the coupling strength model, generate a coupling strength matrix, subtract the coupling strength matrix from the coupling strength standard matrix, and obtain the temperature of each collection time point. The coupling strength differences of the points are summarized to calculate the average coupling strength difference and variance; the predicted temperature and temperature corresponding to a certain hot zone at each acquisition time point are summarized to calculate the average temperature control deviation; the temperature correction value applied by the central control system during the control process of each hot zone at different acquisition time points is obtained, the temperature correction values ​​of all sampling time points are summarized, the temperature correction average is calculated and set as the first compensation factor; the average coupling strength difference and variance of a certain adjacent hot zone are combined with the average temperature control deviation of the target hot zone of the adjacent hot zone to construct the training sample of the hot zone compensation factor; Establish a first compensation factor formula unit: summarize the training samples of the first compensation factor corresponding to all the operation records of the hot zone, use the training samples as input and the first compensation factor as output, establish the first compensation factor formula, perform least squares fitting based on the training samples, obtain the values ​​of k1, k2, k3, and k4, and input them into the first compensation factor formula to establish the final first compensation factor formula.

[0014] Compared with the existing technology, the present invention has the following advantages: by constructing a thermal zone coupling strength model and a standard coupling matrix, the spatial thermal disturbance relationship between each thermal zone is clearly characterized, avoiding the neglect or empirical estimation of thermal coupling in traditional temperature control methods, and enhancing the integrity and synergy of multi-thermal zone temperature control; The first compensation factor is based on historical temperature control deviations and coupled disturbance modeling to effectively correct model prediction errors. The second compensation factor responds to temperature changes in adjacent hot zones and quickly compensates for sudden disturbances. The combination of the two takes into account both "long-term trends" and "short-term disturbances," improving control accuracy and steady-state capabilities. A load-state-driven sliding window prediction model is used to predict temperature trends before control. This overcomes the time lag caused by traditional temperature control methods that rely on static settings or feedback adjustments, and allows for early response to thermal load changes. Real-time calculation of temperature control deviations, coupled disturbances, and compensation values, and their integration into the control set values, enables intelligent decision-making capabilities of the central control system, reduces the burden of manual intervention and parameter debugging, and precisely controls temperature regulation and energy consumption, reducing overheating, underheating, and repeated heating. This improves the consistency and pass rate of product heat treatment, indirectly leading to reduced production costs and energy savings. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of an energy consumption control method applied to an annealing furnace according to the present invention; Figure 2 The diagram is a structural diagram of an energy consumption control system applied to an annealing furnace according to the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 The present invention provides a technical solution: an energy consumption control method for an annealing furnace, the method comprising: Step S100: Obtain an internal structural diagram of the annealing furnace, determine the location information of several hot zones and mark them, and arrange multiple sensors inside each hot zone; collect operating parameters of each hot zone according to preset collection time points, and construct a hot zone state matrix; Wherein, step S100 includes: Step S101: Obtain an internal structural diagram of the annealing furnace, determine the location information of each hot zone based on the internal structural diagram and mark it, arrange several sensors inside each hot zone, and monitor the operating parameters of each hot zone according to preset collection time points; Step S102: Obtain the historical operation records of each hot zone, collect the operating parameters corresponding to each collection time point in the historical operation records, obtain the temperature of the heated object, and calculate the temperature difference between the collection time point and the previous collection time point. , calculate the power feedback value at the acquisition time point according to the following formula: ; Where A1 represents the power feedback value, D represents the material mass, E represents the IoT specific heat capacity, and y represents the time difference between the acquisition time point and the previous acquisition time point. The power command value at the acquisition time point is calculated according to the following formula: ; Where A2 represents the power command value, B represents the average current between the acquisition time point and the previous acquisition time point, and C represents the average voltage between the acquisition time point and the previous acquisition time point. The load state at the acquisition time point is calculated according to the following formula: ; Where A represents the load state at the acquisition time point, the temperature monitored at the acquisition time point is F, and the temperature and load state are combined to construct the hot zone state variable S = (A, F); Step S103: aggregating the state variables generated by each hot zone at multiple acquisition time points to form a hot zone state matrix containing a time series, and uploading and storing the state matrix to a central control system; Step S104: Obtain the historical hot zone state matrix in the central control system, use the load state sequence and temperature of the historical hot zone state matrix as training data, and construct a time series model for temperature prediction through a sliding window regression model as F t =f(A t-n , A t-n+1 ,...,A t ), the time series prediction model takes the load status of multiple consecutive sampling time points as input and the temperature at the tth sampling time point as output, where F t It is expressed as the temperature at the tth acquisition time point, A t-n , A t-n+1 ,...,A t They are respectively represented as the load status at the tn, t-n+1, ​​..., t acquisition time points, where n is the window length; For example, consider an industrial annealing furnace divided into three hot zones (Z1, Z2, and Z3). Sensors are deployed in each hot zone to collect the following operating parameters: current (A), voltage (V), material mass (kg), specific heat capacity (kJ / kg·°C), and real-time temperature (°C). The data collection cycle is once every minute. This example displays data records from five consecutive time points, t1-t5. The data records of five consecutive acquisition time points in hot zone Z2 are shown in Table 1: Table 1

[0018] The load status of the Z2 hot zone at each sampling point is calculated as: t2=87.76%, t3=94.79%, t4=98.38%, t5=99.18%.

[0019] Step S200: Based on the hot zone state matrix and the spatial arrangement of the hot zones, a coupling strength model of adjacent hot zones is constructed, the coupling strength between each adjacent hot zone is calculated, and a coupling strength standard matrix is ​​constructed; Wherein, step S200 includes: Step S201: Based on the historical hot zone state matrix of step S104 and in combination with the spatial arrangement relationship of each hot zone in the annealing furnace, a coupling relationship set of adjacent hot zones is determined, and the temperature and load state of each hot zone in the adjacent hot zones are extracted as the input feature set of the coupling strength model; Step S202: For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of the adjacent hot zones. The calculation formula of the coupling strength model is: ; Among them, G 12 It represents the coupling strength of hot zone 2 to hot zone 1, F1 and F2 represent the temperatures of adjacent hot zones 1 and 2 respectively, A1 and A2 represent the load states of adjacent hot zones 1 and 2 respectively, e1 and e2 represent positive constants to avoid division by zero error, and a1 and a2 represent the weights of temperature and load state respectively; Step S203: Summarize the coupling strength of each adjacent hot zone and calculate the average coupling strength to construct a coupling strength standard matrix of adjacent hot zones. The coupling strength standard matrix is ​​expressed as ; Where H represents the coupling strength standard matrix, G ij It is expressed as the average coupling strength of the coupling strength of hot zone j to hot zone i, N is the total number of hot zones, and the diagonal elements in the coupling strength standard matrix are set to 0, and the elements of non-adjacent hot zones are 0; For example, according to the hot zone state matrix, at time t3, the data of the three hot zones (Z1, Z2, and Z3) are: Z1 temperature is 715°C, load state is 77.76%; Z2 temperature is 730°C, load state is 94.79%; Z3 temperature is 735°C, load state is 100.00%; According to the furnace arrangement, the adjacent pairs are (Z1, Z2), (Z2, Z1), (Z2, Z3), (Z3, Z2); Calculate G according to the coupling strength model 12 (Z1←Z2)=0.5243, G 21 (Z2←Z1)=0.3660, G 23 (Z2←Z3)=0.4799, G 32 (Z3←Z2)=0.5213; The average value of the coupling strength of each adjacent thermal zone is calculated, and a 3x3 standard coupling matrix is constructed according to the order of the thermal zone numbers Z1→Z2→Z3: .

[0020] Step S300: comparing the predicted temperature and the temperature of each thermal zone in the historical operation record, calculating the average value of the temperature control deviation, combining the average difference and variance of the coupling strength of the adjacent thermal zone, constructing a training sample, and fitting a first compensation factor formula based on the least square method; Wherein, step S300 comprises: Step S301: presetting a plurality of consecutive days as a training period, obtaining the operation record of the annealing furnace, in a certain operation record, inputting the operation parameters collected at each collection time point into the time series model of temperature prediction, and obtaining the predicted temperature corresponding to each collection time point; Step S302: obtaining the temperature and load state corresponding to each collection time point of each thermal zone in the annealing furnace, obtaining the thermal zone state matrix of each thermal zone, calculating the coupling strength of each adjacent thermal zone based on the coupling strength model, generating the coupling strength matrix, subtracting the coupling strength matrix from the standard coupling strength matrix, obtaining the coupling strength difference of each collection time point, and collecting the coupling strength difference to calculate the average difference and variance of the coupling strength; Step S303: collecting the predicted temperature and the temperature corresponding to each collection time point of a certain thermal zone, and calculating the average value of the temperature control deviation according to the following formula: Wherein, Fp represents the average value of the temperature control deviation, F' d And F d respectively represent the predicted temperature and the temperature corresponding to the dth collection time point, and b represents the total number of collection time points; Step S304: obtaining the temperature correction value applied in the control process of the central control system corresponding to each collection time point of each thermal zone, collecting the temperature correction values of all sampling time points, calculating the average value of the temperature correction, and setting the average value as the first compensation factor; Step S305: combining the average difference and variance of the coupling strength of a certain adjacent thermal zone with the average value of the temperature control deviation of the target thermal zone of the adjacent thermal zone, constructing the training sample of the compensation factor of the thermal zone, collecting the training sample of the first compensation factor corresponding to all operation records of the thermal zone, taking the training sample as the input and the first compensation factor as the output, and establishing the first compensation factor formula as: ; Wherein, Q1 represents the first compensation factor, L1 and L2 represent the average difference and variance of coupling strength respectively, k1, k2, k3 represent the weight values of the average difference, variance and temperature control deviation average value of coupling strength respectively, k4 represents the deviation value of the first compensation factor formula, the least square fitting is performed according to the training sample, the numerical values of k1, k2, k3 and k4 are obtained, and are input into the first compensation factor formula, and the final first compensation factor formula is established. For example, the data of three consecutive days is set as a training period, 5 time points are sampled each day, a total of 15 sampling data, and the sliding window length is 3. Taking the 5th time point of the 3rd day of the Z2 hot area as an example, the load states in the sliding window are 92.06%, 94.03% and 94.67% respectively, the predicted temperature is 739.5, and the measured temperature is 738. The states of Z1, Z2 and Z3 at the 5th time point of the 3rd day are shown in Table 2. Table 2

[0021] The calculated G 21 (Z2←Z1)=0.4991, G 23 (Z2←Z3)=0.6839, the average value is 0.5915, the corresponding standard coupling matrix value is 0.4087, and the coupling strength difference is 0.1828. The following data is taken as a group of training samples as shown in Table 3: Table 3

[0022] A plurality of groups of running data are fitted by using the least square method, and finally the fitting coefficients: the numerical values of k1, k2, k3 and k4 are 12, 20, 0.6 and 0.05 respectively.

[0023] Step S400: count the set of adjacent hot areas of each target hot area, calculate the thermal disturbance influence weight according to the coupling strength, collect the temperature change value of the adjacent hot area, and construct the second compensation factor; Wherein, step S400 includes: Step S401: count all adjacent hot area sets of a target hot area, collect the coupling strength of each adjacent hot area to the target hot area in the coupling strength standard matrix, and calculate the thermal disturbance influence weight of the adjacent hot area to the target hot area according to the following formula: ; Wherein, q r represents the thermal disturbance influence weight of adjacent hot area r to target hot area g, G rg represents the coupling strength of adjacent hot area r to target hot area g, G ngIt is expressed as the coupling strength of the adjacent hot zone n to the target hot zone g, and m is the total number of adjacent hot zones; Step S402: Obtain the temperature change value of each adjacent hot zone between two consecutive acquisition time points, and calculate the second compensation factor of the target hot zone according to the following second compensation factor formula: ; Among them, Q2 represents the second compensation factor, q ng It is expressed as the weight of the thermal disturbance influence of the adjacent hot zone n on the target hot zone g, Expressed as the temperature change value of adjacent hot zone n; For example, G 21 The standard coupling value of (Z2←Z1) is 0.3353, G 23 The standard coupling value of (Z2←Z3) is 0.4821, and the calculated thermal perturbation influence weights are 0.4103 and 0.5897, respectively.

[0024] Step S500: obtaining the real-time operating parameters of the annealing furnace and inputting them into the temperature prediction model, calculating the real-time predicted temperature, combining the real-time temperature control deviation average value, the real-time average difference of the coupling strength, and the real-time variance parameter, respectively inputting them into the first compensation factor and the second compensation factor formula, calculating the first compensation factor and the second compensation factor, and calculating the temperature control set value executed by the central control system, sending the temperature control set value to the annealing furnace controller through the command interface, and adjusting the heater power output of the corresponding hot zone; Wherein, step S500 includes: Step S501: Acquire the real-time operating parameters of the annealing furnace, input them into the time series model of temperature prediction, calculate the real-time predicted temperature corresponding to each acquisition time point, collect the real-time temperature at the acquisition time point, calculate the real-time temperature control deviation average value, calculate the real-time average difference and real-time variance of the coupling strength, and input them into the first compensation factor formula to calculate the first compensation factor; Step S502: obtaining the real-time temperature change value of each adjacent hot zone between two consecutive acquisition time points, and inputting it into the second compensation factor formula to calculate the second compensation factor; Step S503: Obtain the target temperature setting value preset by the central control system as F s , and calculate the temperature setpoint executed by the central control system according to the following formula: ; Among them, F' s The temperature setting value is represented as a temperature setting value executed by the central control system, and the temperature setting value executed by the central control system is sent to the annealing furnace controller through the temperature control command interface, and the heater power output of the corresponding hot zone is adjusted according to the temperature setting value; For example, the real-time data of Z2 is shown in Table 4: Table 4

[0025] The calculated real-time temperature control deviation is 2.5; Assume that at time t, the temperature of the adjacent hot zone Z1 is 716 and the load state is 92.93%; the temperature of the adjacent hot zone Z3 is 745 and the load state is 45.85%; The calculated coupling strengths are 0.5072 and 0.6761 respectively, the average real-time coupling strength is 0.5917, the standard coupling matrix value is 0.4087, the calculated average difference is 0.183, the variance is 0.0406, and the first compensation factor calculated according to the first compensation factor formula is 3.833; The temperature change of the adjacent hot zone Z1 is +2; the temperature change of the adjacent hot zone Z3 is +4. The calculated second compensation factor is 3.1794. The preset target temperature setting value is 740, and the execution temperature setting value is 747.01. The central control system sends 747.01 to the hot zone Z2 controller through the temperature control command interface. The controller dynamically adjusts the heater output power according to this temperature setting value.

[0026] In order to better implement the above method, an energy consumption control system for an annealing furnace is also proposed. The system includes a hot zone state module, a coupling strength module, a first compensation factor module, a second compensation factor module and a real-time control module; Hot zone status module: Obtain the internal structure diagram of the annealing furnace, determine the location information of several hot zones and mark them, and arrange multiple sensors inside each hot zone; collect the operating parameters of each hot zone according to the preset collection time point, and construct the hot zone status matrix; Among them, the hot zone state module includes the hot zone state variable unit and the temperature prediction model unit: Hot zone state variable unit: obtains an internal structural diagram of the annealing furnace, determines the location information of each hot zone based on the internal structural diagram and marks it, arranges several sensors inside each hot zone, and monitors the operating parameters of each hot zone according to preset collection time points; obtains the historical operation records of each hot zone, collects the operating parameters corresponding to each collection time point in the historical operation records, calculates the load state at the collection time point, and combines the temperature and load state to construct the hot zone state variable; Temperature prediction model unit: The state variables generated by each hot zone at multiple acquisition time points are aggregated to form a hot zone state matrix containing a time series, and the state matrix is ​​uploaded and stored in the central control system; the historical hot zone state matrix in the central control system is obtained, and the load state sequence and temperature of the historical hot zone state matrix are used as training data. A time series model for temperature prediction is constructed through a sliding window regression model. The time series prediction model takes the load state at multiple consecutive sampling time points as input and the temperature at the acquisition time point as output.

[0027] Coupling strength module: Based on the hot zone state matrix and the spatial arrangement of the hot zones, a coupling strength model of adjacent hot zones is constructed, the coupling strength between each adjacent hot zone is calculated, and a coupling strength standard matrix is ​​constructed; The coupling strength module includes a coupling strength calculation unit and a coupling strength standard matrix establishment unit: Coupling strength calculation unit: Based on the historical hot zone state matrix and the spatial arrangement relationship of each hot zone in the annealing furnace, the coupling relationship set of adjacent hot zones is determined. The temperature and load status of each hot zone in the adjacent hot zones are extracted as the input feature set of the coupling strength model. For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of the adjacent hot zones. Establish a coupling strength standard matrix unit: summarize the coupling strength of each adjacent hot zone and calculate the average coupling strength to construct a coupling strength standard matrix of adjacent hot zones, in which the diagonal elements are set to 0 and the elements of non-adjacent hot zones are 0.

[0028] First compensation factor module: compares the predicted temperature of each hot zone with the temperature in the historical operation record, calculates the average temperature control deviation, combines the average difference and variance of the coupling strength of adjacent hot zones, constructs a training sample, and fits the first compensation factor formula based on the least squares method; The first compensation factor module includes a training sample construction unit and a first compensation factor formula establishment unit: Construct a training sample unit: preset several consecutive days as the training cycle, obtain the operation record of the annealing furnace, input the operation parameters collected at each collection time point in a certain operation record into the time series model of temperature prediction, and obtain the predicted temperature corresponding to each collection time point; obtain the temperature and load state corresponding to each hot zone in the annealing furnace at different collection time points, obtain the hot zone state matrix of each hot zone, calculate the coupling strength of each adjacent hot zone based on the coupling strength model, generate a coupling strength matrix, subtract the coupling strength matrix from the coupling strength standard matrix, and obtain the temperature of each collection time point. The coupling strength differences of the points are summarized to calculate the average coupling strength difference and variance; the predicted temperature and temperature corresponding to a certain hot zone at each acquisition time point are summarized to calculate the average temperature control deviation; the temperature correction value applied by the central control system during the control process of each hot zone at different acquisition time points is obtained, the temperature correction values ​​of all sampling time points are summarized, the temperature correction average is calculated and set as the first compensation factor; the average coupling strength difference and variance of a certain adjacent hot zone are combined with the average temperature control deviation of the target hot zone of the adjacent hot zone to construct the training sample of the hot zone compensation factor; Establish a first compensation factor formula unit: summarize the training samples of the first compensation factor corresponding to all the operation records of the hot zone, use the training samples as input and the first compensation factor as output, establish the first compensation factor formula, perform least squares fitting based on the training samples, obtain the values ​​of k1, k2, k3, and k4, and input them into the first compensation factor formula to establish the final first compensation factor formula.

[0029] Second compensation factor module: Counts the adjacent hot zone sets of each target hot zone, calculates the thermal disturbance influence weight according to the coupling strength, collects the temperature change values ​​of the adjacent hot zones, and constructs the second compensation factor; Real-time control module: obtains the real-time operating parameters of the annealing furnace and inputs them into the temperature prediction model, calculates the real-time predicted temperature, combines the real-time temperature control deviation average value, the real-time average difference of the coupling strength and the real-time variance parameter, inputs them into the first compensation factor and the second compensation factor formula respectively, calculates the first compensation factor and the second compensation factor, and calculates the temperature control set value executed by the central control system, sends the temperature control set value to the annealing furnace controller through the command interface, and adjusts the heater power output of the corresponding hot zone.

[0030] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An energy consumption control method for an annealing furnace, characterized in that: Methods include: Step S100: Obtain an internal structural diagram of the annealing furnace, determine the location information of several hot zones and mark them, and arrange multiple sensors inside each hot zone; collect operating parameters of each hot zone according to preset collection time points, and construct a hot zone state matrix; Step S200: Based on the hot zone state matrix and the spatial arrangement of the hot zones, a coupling strength model of adjacent hot zones is constructed, the coupling strength between each adjacent hot zone is calculated, and a coupling strength standard matrix is ​​constructed; Step S300: Compare the predicted temperature of each hot zone with the temperature in the historical operation record, calculate the average value of the temperature control deviation, combine the average difference and variance of the coupling strength of adjacent hot zones, construct a training sample and fit the first compensation factor formula based on the least squares method; Step S400: Counting the adjacent hot zone sets of each target hot zone, calculating the thermal disturbance influence weight according to the coupling strength, collecting the temperature change values ​​of the adjacent hot zones, and constructing a second compensation factor; Step S500: Obtain the real-time operating parameters of the annealing furnace and input them into the temperature prediction model, calculate the real-time predicted temperature, combine the real-time temperature control deviation average value, the real-time average difference of the coupling strength and the real-time variance parameter, input them into the first compensation factor and the second compensation factor formula respectively, calculate the first compensation factor and the second compensation factor, and calculate the temperature control set value executed by the central control system, send the temperature control set value to the annealing furnace controller through the command interface, and adjust the heater power output of the corresponding hot zone.

2. The energy consumption control method for an annealing furnace according to claim 1, characterized in that: The step S100 includes the following steps: Step S101: Obtain an internal structural diagram of the annealing furnace, determine the location information of each hot zone based on the internal structural diagram and mark it, arrange several sensors inside each hot zone, and monitor the operating parameters of each hot zone according to preset collection time points; Step S102: Obtain the historical operation records of each hot zone, collect the operating parameters corresponding to each collection time point in the historical operation records, obtain the temperature of the heated object, and calculate the temperature difference between the collection time point and the previous collection time point. , calculate the power feedback value at the acquisition time point according to the following formula: ； Where A1 represents the power feedback value, D represents the material mass, E represents the IoT specific heat capacity, and y represents the time difference between the acquisition time point and the previous acquisition time point. The power command value at the acquisition time point is calculated according to the following formula: ； Where A2 represents the power command value, B represents the average current between the acquisition time point and the previous acquisition time point, and C represents the average voltage between the acquisition time point and the previous acquisition time point. The load state at the acquisition time point is calculated according to the following formula: ； Where A represents the load state at the acquisition time point, the temperature monitored at the acquisition time point is F, and the temperature and load state are combined to construct the hot zone state variable S = (A, F); Step S103: aggregating the state variables generated by each hot zone at multiple acquisition time points to form a hot zone state matrix containing a time series, and uploading and storing the state matrix to a central control system; Step S104: Obtain the historical hot zone state matrix in the central control system, use the load state sequence and temperature of the historical hot zone state matrix as training data, and construct a time series model for temperature prediction through a sliding window regression model as F t =f(A t-n , A t-n+1 ,...,A t ), the time series prediction model takes the load status of multiple consecutive sampling time points as input and the temperature at the tth sampling time point as output, where F t It is expressed as the temperature at the tth acquisition time point, A t-n , A t-n+1 ,...,A t They are respectively represented as the load status at the tn, t-n+1, ​​..., tth acquisition time points, and n is represented as the window length.

3. The energy consumption control method for an annealing furnace according to claim 2, characterized in that: The step S200 includes the following steps: Step S201: Based on the historical hot zone state matrix of step S104 and in combination with the spatial arrangement relationship of each hot zone in the annealing furnace, a coupling relationship set of adjacent hot zones is determined, and the temperature and load state of each hot zone in the adjacent hot zones are extracted as the input feature set of the coupling strength model; Step S202: For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of the adjacent hot zones. The calculation formula of the coupling strength model is: ； Among them, G 12 It represents the coupling strength of hot zone 2 to hot zone 1, F1 and F2 represent the temperatures of adjacent hot zones 1 and 2 respectively, A1 and A2 represent the load states of adjacent hot zones 1 and 2 respectively, e1 and e2 represent positive constants to avoid division by zero error, and a1 and a2 represent the weights of temperature and load state respectively; Step S203: Summarize the coupling strength of each adjacent hot zone and calculate the average coupling strength to construct a coupling strength standard matrix of adjacent hot zones. The coupling strength standard matrix is ​​expressed as ； Where H represents the coupling strength standard matrix, G ij It is expressed as the average coupling strength of the coupling strength of hot zone j to hot zone i, N is expressed as the total number of hot zones, and the diagonal elements in the coupling strength standard matrix are set to 0, and the elements of non-adjacent hot zones are 0.

4. The energy consumption control method for an annealing furnace according to claim 3, characterized in that: The step S300 includes the following steps: Step S301: Preset a number of consecutive days as a training cycle, obtain the operation records of the annealing furnace, input the operation parameters collected at each collection time point in a certain operation record into the time series model for temperature prediction, and obtain the predicted temperature corresponding to each collection time point; Step S302: Obtain the temperature and load state corresponding to each hot zone in the annealing furnace at different acquisition time points, obtain a hot zone state matrix for each hot zone, calculate the coupling strength of each adjacent hot zone based on the coupling strength model, generate a coupling strength matrix, subtract the coupling strength matrix from the coupling strength standard matrix, obtain the coupling strength difference at each acquisition time point, summarize it, and calculate the average difference and variance of the coupling strength; Step S303: Summarize the predicted temperature and temperature corresponding to a certain hot zone at each acquisition time point, and calculate the average temperature control deviation according to the following formula: Among them, Fp represents the average value of temperature control deviation, F' d and F d They are the predicted temperature and temperature corresponding to the dth acquisition time point, respectively, and b is the total number of acquisition time points; Step S304: obtaining the temperature correction value applied by the central control system to each hot zone at different sampling time points, summing up the temperature correction values ​​at all sampling time points, calculating the average temperature correction value and setting it as the first compensation factor; Step S305: Combine the average difference and variance of the coupling strength of a certain adjacent hot zone with the average temperature control deviation of the target hot zone of the adjacent hot zone to construct a training sample of the hot zone compensation factor. Summarize the training samples of the first compensation factor corresponding to all operation records of the hot zone. Take the training samples as input and the first compensation factor as output to establish the formula of the first compensation factor as follows: ； Among them, Q1 is represented as the first compensation factor, L1 and L2 are represented as the average difference and variance of coupling strength, respectively, k1, k2, and k3 are represented as the weights of the average difference, variance, and average value of temperature control deviation of coupling strength, respectively, and k4 is represented as the deviation value of the first compensation factor formula. Least squares fitting is performed based on the training samples to obtain the values ​​of k1, k2, k3, and k4, and input them into the first compensation factor formula to establish the final first compensation factor formula.

5. The energy consumption control method for an annealing furnace according to claim 4, characterized in that: The step S400 includes the following steps: Step S401: Count all adjacent hot zone sets of a target hot zone, collect the coupling strength of each adjacent hot zone to the target hot zone in the coupling strength standard matrix, and calculate the thermal disturbance influence weight of the adjacent hot zone to the target hot zone according to the following formula: ； Among them, q r It is expressed as the thermal disturbance influence weight of the adjacent hot zone r on the target hot zone g, G rg It is expressed as the coupling strength of the adjacent hot zone r to the target hot zone g, G ng It is expressed as the coupling strength of the adjacent hot zone n to the target hot zone g, and m is the total number of adjacent hot zones; Step S402: Obtain the temperature change value of each adjacent hot zone between two consecutive acquisition time points, and calculate the second compensation factor of the target hot zone according to the following second compensation factor formula: ； Among them, Q2 represents the second compensation factor, q ng It is expressed as the thermal disturbance influence weight of the adjacent hot zone n on the target hot zone g, Expressed as the temperature change value of adjacent hot zone n.

6. The energy consumption control method for an annealing furnace according to claim 5, characterized in that: The step S500 includes the following steps: Step S501: Acquire the real-time operating parameters of the annealing furnace, input them into the time series model of temperature prediction, calculate the real-time predicted temperature corresponding to each acquisition time point, collect the real-time temperature at the acquisition time point, calculate the real-time temperature control deviation average value, calculate the real-time average difference and real-time variance of the coupling strength, and input them into the first compensation factor formula to calculate the first compensation factor; Step S502: obtaining the real-time temperature change value of each adjacent hot zone between two consecutive acquisition time points, and inputting it into the second compensation factor formula to calculate the second compensation factor; Step S503: Obtain the target temperature setting value preset by the central control system as F s , and calculate the temperature setpoint executed by the central control system according to the following formula: ； Among them, F' s It represents the temperature setting value executed by the central control system. The temperature setting value executed by the central control system is sent to the annealing furnace controller through the temperature control instruction interface, and the heater power output of the corresponding hot zone is adjusted according to the temperature setting value.

7. An energy consumption control system for an annealing furnace, used to implement the energy consumption control method for an annealing furnace according to any one of claims 1 to 6, characterized in that: The system includes a hot zone state module, a coupling strength module, a first compensation factor module, a second compensation factor module and a real-time control module; The hot zone status module: obtains the internal structure diagram of the annealing furnace, determines the location information of several hot zones and marks them, and arranges multiple sensors inside each hot zone; According to the preset collection time points, the operating parameters of each hot zone are collected and the hot zone state matrix is ​​constructed; The coupling strength module: based on the hot zone state matrix and the spatial arrangement of the hot zones, constructs an adjacent hot zone coupling strength model, calculates the coupling strength between each adjacent hot zone, and constructs a coupling strength standard matrix; The first compensation factor module: compares the predicted temperature of each hot zone with the temperature in the historical operation record, calculates the average value of the temperature control deviation, combines the average difference and variance of the coupling strength of adjacent hot zones, constructs a training sample and fits the first compensation factor formula based on the least squares method; The second compensation factor module: counts the adjacent hot zone sets of each target hot zone, calculates the thermal disturbance influence weight according to the coupling strength, collects the temperature change values ​​of the adjacent hot zones, and constructs the second compensation factor; The real-time control module obtains the real-time operating parameters of the annealing furnace and inputs them into the temperature prediction model, calculates the real-time predicted temperature, combines the real-time temperature control deviation average value, the real-time average difference of the coupling strength and the real-time variance parameter, inputs them into the first compensation factor and the second compensation factor formula respectively, calculates the first compensation factor and the second compensation factor, and calculates the temperature control set value executed by the central control system, sends the temperature control set value to the annealing furnace controller through the command interface, and adjusts the heater power output of the corresponding hot zone.

8. The energy consumption control system for an annealing furnace according to claim 7, characterized in that: The hot zone state module includes a hot zone state variable unit and a temperature prediction model unit: The hot zone state variable unit obtains an internal structural diagram of the annealing furnace, determines the location information of each hot zone based on the internal structural diagram and marks it, arranges a number of sensors inside each hot zone, and monitors the operating parameters of each hot zone according to preset collection time points; Obtaining historical operation records of each hot zone, collecting operating parameters corresponding to each collection time point in the historical operation records, calculating the load state at the collection time point, combining the temperature and load state, and constructing a hot zone state variable; The temperature prediction model unit: aggregates the state variables generated by each hot zone at multiple acquisition time points to form a hot zone state matrix containing a time series, and uploads and stores the state matrix to the central control system; obtains the historical hot zone state matrix in the central control system, uses the load state sequence and temperature of the historical hot zone state matrix as training data, and constructs a time series model for temperature prediction through a sliding window regression model. The time series prediction model uses the load state at multiple consecutive sampling time points as input and the temperature at the acquisition time point as output.

9. The energy consumption control system for an annealing furnace according to claim 7, characterized in that: The coupling strength module includes a coupling strength calculation unit and a coupling strength standard matrix establishment unit: The coupling strength calculation unit determines the coupling relationship set of adjacent hot zones based on the historical hot zone state matrix and the spatial arrangement relationship of each hot zone in the annealing furnace, and extracts the temperature and load state of each hot zone in the adjacent hot zones as the input feature set of the coupling strength model; For each group of adjacent hot zones, a coupling strength model is constructed based on the temperature and load status to calculate the coupling strength of adjacent hot zones; The coupling strength standard matrix unit is established: the coupling strength of each adjacent hot zone is summarized and the coupling strength average is calculated to construct a coupling strength standard matrix of adjacent hot zones, wherein the diagonal elements in the coupling strength standard matrix are set to 0, and the elements of non-adjacent hot zones are 0.

10. The energy consumption control system for an annealing furnace according to claim 7, characterized in that: The first compensation factor module includes a training sample construction unit and a first compensation factor formula establishment unit: The training sample unit is constructed by presetting a number of consecutive days as a training cycle, obtaining the operation record of the annealing furnace, and inputting the operation parameters collected at each collection time point in a certain operation record into the time series model of temperature prediction to obtain the predicted temperature corresponding to each collection time point; Obtain the temperature and load state corresponding to each hot zone in the annealing furnace at different acquisition time points, obtain the hot zone state matrix of each hot zone, calculate the coupling strength of each adjacent hot zone based on the coupling strength model, generate a coupling strength matrix, subtract the coupling strength matrix from the coupling strength standard matrix, obtain the coupling strength difference at each acquisition time point, summarize it, and calculate the average difference and variance of the coupling strength; The predicted temperature and temperature corresponding to a certain hot zone at each sampling time point are summarized to calculate the average temperature control deviation; the temperature correction value applied by the central control system to each hot zone at different sampling time points is obtained, the temperature correction values ​​of all sampling time points are summarized, and the average temperature correction value is calculated and set as the first compensation factor; the average difference and variance of the coupling strength of a certain adjacent hot zone are combined with the average temperature control deviation of the target hot zone of the adjacent hot zone to construct the training sample of the hot zone compensation factor; The first compensation factor formula establishment unit: summarizes the training samples of the first compensation factor corresponding to all the operation records of the hot zone, takes the training samples as input and the first compensation factor as output, establishes the first compensation factor formula, performs least squares fitting based on the training samples, obtains the values ​​of k1, k2, k3, and k4, and inputs them into the first compensation factor formula to establish the final first compensation factor formula.

Images