Converter oxygen lance cooling water self-adaptive adjusting method and related equipment

Abstract

This invention discloses an adaptive adjustment method and related equipment for converter oxygen lance cooling water, relating to the field of converter oxygen lance regulation. It constructs a heat load assessment model, utilizing real-time collected furnace gas temperature, oxygen lance body temperature, and their respective time-varying rates to predict the oxygen lance heat load at a preset forward time point. Based on this, it dynamically generates a cooling water flow rate setpoint, achieving forward-looking cooling regulation with furnace gas temperature as the key feedforward variable. Simultaneously, it combines the flow deviation and deviation rate between the actual cooling water flow rate and the setpoint, and introduces the oxygen lance heat load prediction value as a fuzzy control input. Through a predefined fuzzy control rule set, it corrects the PID parameters in real time, forming an adaptive optimization of the flow control loop. This overcomes the lag of traditional pure feedback control while significantly improving the system's response speed, tracking accuracy, and operational stability to sudden changes in heat load and external disturbances.

Description

Technical Field

[0001] This invention relates to the field of converter oxygen lance regulation, and more particularly to an adaptive regulation method and related equipment for converter oxygen lance cooling water. Background Technology

[0002] In the oxygen top-blown converter steelmaking process, the oxygen lance, as a core process equipment, is vertically inserted into the high-temperature molten pool (the furnace temperature can reach over 1600°C) through the furnace opening, and high-speed oxygen is injected to drive violent metallurgical reactions such as decarburization and heating. Due to long-term exposure to strong heat radiation, high-temperature flue gas, and chemical corrosion, the oxygen lance must rely on a forced water cooling system for continuous cooling; otherwise, the lance body is prone to burning, deformation, or even explosion, seriously threatening production safety.

[0003] Currently, industrial sites commonly employ feedback control strategies based on oxygen lance body temperature or cooling water outlet temperature to regulate cooling water flow. However, this method is essentially a lag-response control: a significant rise in lance body temperature usually indicates that the thermal load shock has already acted on the oxygen lance structure. Increasing the cooling water flow at this point often fails to prevent localized instantaneous overheating, posing a significant safety risk. Furthermore, traditional systems rely heavily on manual valve adjustment based on experience, resulting in low control accuracy, slow response, and a tendency for excessive or insufficient cooling water flow (leading to energy waste and pump overload) or insufficient cooling, thereby affecting equipment lifespan and operational stability.

[0004] In recent years, although there have been attempts to introduce automatic control methods, most are still limited to PID control with fixed parameters, which is difficult to adapt to the strong nonlinearity, large time delay, and abrupt changes in heat load during converter blowing (such as the differences in heat release during the start-up, main blowing, and post-blowing periods). Crucially, existing technologies generally neglect furnace gas temperature, a precursor signal highly synchronized with the smelting process: changes in furnace gas temperature precede the oxygen lance body temperature response and can effectively reflect the trend of the heat load to be applied to the oxygen lance. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, such as lagging cooling water control, low precision, high energy consumption, and inability to adapt to dynamic heat load changes in the smelting process, this invention proposes an adaptive adjustment method for converter oxygen lance cooling water, comprising: Real-time acquisition of multi-source parameters reflecting the oxygen lance heat load status and converter smelting process, including at least: oxygen lance body temperature, furnace gas temperature, actual cooling water flow rate, cooling water inlet temperature, and cooling water outlet temperature. Based on the current furnace gas temperature, oxygen lance body temperature, furnace gas temperature change rate over time and oxygen lance body temperature change rate over time, the predicted oxygen lance heat load at a future preset forward time point is evaluated using a heat load assessment model. Based on the predicted oxygen lance heat load, a cooling water flow rate setpoint is dynamically generated. The flow deviation is calculated based on the actual flow rate of the cooling water and the set flow rate of the cooling water, and the rate of change of the deviation is obtained by differential calculation of the flow deviation. Based on the flow deviation, deviation change rate, and oxygen lance heat load prediction, the PID parameter correction amount is obtained using a predefined fuzzy control rule set. The PID parameter correction is applied to the current PID parameters to obtain the proportional coefficient, integral coefficient, and derivative coefficient optimized in real time. Using the real-time optimized proportional coefficient, integral coefficient, and derivative coefficient, and taking the flow deviation as input, a control output signal is generated through a PID control algorithm, which drives the actuator to adjust the cooling water flow rate so that the actual cooling water flow rate follows the dynamically generated cooling water flow rate setpoint.

[0006] Furthermore, the acquisition of the heat load assessment model includes: Collect multivariate time series data corresponding to multiple historical blowing cycles. The multivariates include furnace gas temperature, oxygen lance body temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate. Based on the cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, and specific heat capacity of water, the actual heat load value absorbed by the oxygen lance at each moment is calculated to form a heat load time series; wherein, for each historical blowing cycle, its corresponding multivariate time series and heat load time series are aligned at the same time. A training sample set was constructed using multivariate time series data from multiple historical blowing cycles and heat load time series data. Using the training sample set, a heat load assessment model is obtained through supervised learning algorithm training.

[0007] Furthermore, the construction of a training sample set using multivariate time-series data from multiple historical blowing cycles and heat load time series data includes: For the time series data of each historical blowing cycle, iterate through each collection point; For the currently traversed acquisition time, extract the input features, which include: The furnace gas temperature at the current data collection moment; The rate of change of furnace gas temperature over time is calculated from the furnace gas temperature time series data with the current acquisition time as the endpoint and a preset time window as the length. The oxygen lance body temperature at the current sampling moment; The rate of change of oxygen lance body temperature over time is calculated from the time series data of oxygen lance body temperature with the current acquisition time as the endpoint and the length as a preset time window. Determine the training label corresponding to the current acquisition time. The training label is the oxygen lance heat load value corresponding to a preset look-ahead time point after the current acquisition time and within the same blowing cycle. The preset look-ahead point is determined based on the converter blowing thermal response characteristics. The input features are combined with the corresponding training labels to form a training sample; A training sample set is constructed by using multiple training samples.

[0008] Furthermore, the calculation of the heat load value includes: Calculate the difference between the cooling water outlet temperature and the cooling water inlet temperature to obtain the cooling water temperature rise; Multiply the cooling water flow rate by the density of water to obtain the cooling water mass flow rate; The actual heat load absorbed by the oxygen lance is obtained by multiplying the cooling water mass flow rate, the specific heat capacity of the water, and the cooling water temperature rise.

[0009] Furthermore, the heat load assessment model is a multilayer perceptron, a long short-term memory network, or a support vector regression machine.

[0010] Further, the step of dynamically generating a cooling water flow rate setpoint based on the predicted oxygen lance heat load includes: Based on the predicted oxygen lance heat load, the corresponding cooling water flow rate setting value is obtained using a preset mapping table. The mapping table is pre-established based on the oxygen lance heat load value and cooling water flow rate setting value corresponding to the oxygen lance body temperature not exceeding the preset safety threshold in historical blowing data.

[0011] Furthermore, the PID parameter correction includes proportional coefficient correction, integral coefficient correction, and derivative coefficient correction; The step of obtaining PID parameter correction values ​​based on the flow deviation, deviation change rate, and oxygen lance heat load prediction value using a predefined fuzzy control rule set includes: The flow deviation, deviation change rate, and oxygen lance heat load prediction value are used as three input variables of the fuzzy controller, and fuzzification is performed according to the preset membership function to obtain their respective fuzzy language values. From a predefined set of fuzzy control rules, select association rules that match the fuzzy linguistic values ​​of the flow deviation, the deviation change rate, and the oxygen lance heat load prediction value in the premise part, wherein the premise part is composed of the fuzzy linguistic values ​​corresponding to the three input variables. Based on the conclusion of the selected association rule, generate output fuzzy sets corresponding to the proportional coefficient correction, integral coefficient correction, and differential coefficient correction, respectively. Defuzzification operations are performed on each output fuzzy set to obtain real-time, numerical correction values ​​for the proportional coefficient, integral coefficient, and differential coefficient.

[0012] To address the aforementioned problems, another aspect of this invention provides an electronic device, including: a processor and a memory storing a program, the program including instructions that, when executed by the processor, cause the processor to perform the aforementioned adaptive adjustment method for converter oxygen lance cooling water.

[0013] To address the aforementioned problems, in another aspect of this invention, a non-transitory machine-readable medium storing computer instructions is provided, the computer instructions being used to cause the computer to execute the aforementioned adaptive adjustment method for converter oxygen lance cooling water.

[0014] To address the aforementioned problems, in another aspect of this invention, a computer program product is provided, comprising a computer program / instruction, which, when executed by a processor, implements the adaptive adjustment method for converter oxygen lance cooling water described above.

[0015] Compared with the prior art, the present invention has at least the following beneficial effects: (1) This invention constructs a heat load assessment model and uses real-time collected furnace gas temperature, oxygen lance body temperature and their respective time change rates to predict the oxygen lance heat load forecast value at a preset forward time point in advance, and dynamically generates a cooling water flow set value accordingly, realizing forward cooling regulation with furnace gas temperature as the key feedforward variable; at the same time, it combines the flow deviation and deviation change rate formed by the actual cooling water flow rate and the set value, and introduces the oxygen lance heat load forecast value as a fuzzy control input, and corrects the PID parameters in real time through a predefined fuzzy control rule set to form an adaptive optimization of the flow control loop, thereby overcoming the lag of traditional pure feedback control and significantly improving the system's response speed, tracking accuracy and operation stability to sudden changes in heat load and external disturbances.

[0016] (2) The present invention constructs a training sample set by using a multivariate time series based on historical blowing cycles and a calculated heat load time series, and uses a supervised learning algorithm to train a heat load assessment model, so that the model can accurately capture the dynamic evolution law of heat load in the converter smelting process, and provide a reliable basis for the forward-looking adjustment of cooling water flow rate set value.

[0017] (3) This invention uses the flow deviation, deviation change rate and oxygen lance heat load prediction value as inputs to the fuzzy controller, and uses a predefined fuzzy control rule set to correct the PID parameters online, so that the control system can adaptively adjust the control intensity according to the heat load trend and flow tracking status, and still maintain good dynamic response performance and robustness when the blowing conditions change suddenly or there are external disturbances.

[0018] (4) This invention establishes a mapping relationship table between the predicted heat load value and the set value of cooling water flow rate based on the working condition that the oxygen lance body temperature does not exceed the safety threshold in the historical blowing data. This ensures that the flow rate is accurately matched under the premise of meeting cooling safety, avoids energy waste caused by excessive supply of cooling water, and reduces water pump energy consumption and system operating costs. Attached Figure Description

[0019] Figure 1 A flowchart of an adaptive adjustment method for oxygen lance cooling water in a converter; Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] The following are specific embodiments of the present invention, which are described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0021] To overcome the shortcomings of existing technologies, such as lag in cooling water control, low precision, high energy consumption, and inability to adapt to dynamic heat load changes in the smelting process, such as... Figure 1 As shown, this invention proposes an adaptive adjustment method for converter oxygen lance cooling water, comprising: Real-time acquisition of multi-source parameters reflecting the oxygen lance heat load status and converter smelting process, including at least: oxygen lance body temperature, furnace gas temperature, actual cooling water flow rate, cooling water inlet temperature, and cooling water outlet temperature. Based on the current furnace gas temperature, oxygen lance body temperature, furnace gas temperature change rate over time and oxygen lance body temperature change rate over time, the predicted oxygen lance heat load at a future preset forward time point is evaluated using a heat load assessment model. The acquisition of the heat load assessment model includes: Collect multivariate time series data corresponding to multiple historical blowing cycles. The multivariates include furnace gas temperature, oxygen lance body temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate. Based on the cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, and specific heat capacity of water, the actual heat load value absorbed by the oxygen lance at each moment is calculated to form a heat load time series; wherein, for each historical blowing cycle, its corresponding multivariate time series and heat load time series are aligned at the same time. A training sample set was constructed using multivariate time series data from multiple historical blowing cycles and heat load time series data. The construction of a training sample set using multivariate time series data from multiple historical blowing cycles and heat load time series includes: For the time series data of each historical blowing cycle, iterate through each collection point; For the currently traversed acquisition time, extract the input features, which include: The furnace gas temperature at the current data collection moment; The rate of change of furnace gas temperature over time is calculated from the furnace gas temperature time series data with the current acquisition time as the endpoint and a preset time window as the length. The oxygen lance body temperature at the current sampling moment; The rate of change of oxygen lance body temperature over time is calculated from the time series data of oxygen lance body temperature with the current acquisition time as the endpoint and the length as a preset time window. Determine the training label corresponding to the current acquisition time. The training label is the oxygen lance heat load value corresponding to a preset look-ahead time point after the current acquisition time and within the same blowing cycle. The preset look-ahead point is determined based on the converter blowing thermal response characteristics. In one embodiment of the present invention, the preset look-ahead time point is determined based on the thermal response characteristics during the converter blowing process. Thermal response characteristics refer to the response speed and lag time of changes in relevant thermal environment parameters (such as furnace gas temperature, oxygen lance body temperature, cooling water temperature rise, etc.) of the converter system after being subjected to operational disturbances such as oxygen supply and charging. Due to the large thermal inertia and complex heat transfer process of the converter system, abrupt changes in the furnace thermal state (such as furnace gas temperature or reaction intensity) do not immediately reflect in the oxygen lance heat load, but rather exhibit a significant time lag—typically several seconds to tens of seconds—due to the inertia of heat transfer and flow field response. To overcome this lag, the present invention analyzes historical blowing data to identify how long after the observed furnace thermal state (such as furnace gas temperature) will have a major impact on the oxygen lance heat load, i.e., the typical lag time between the two. The future time point corresponding to this lag time is set as the "preset look-ahead time point". By incorporating furnace gas temperature, oxygen lance body temperature, and their rate of change over time as inputs, the heat load assessment model can effectively capture the dynamic evolution trend of the thermal environment during converter blowing, thereby accurately predicting the oxygen lance heat load level in the near future.

[0022] In this embodiment, the preset look-ahead time point is typically offset from the current time by 5 to 15 seconds. This time offset is sufficient to cover the dynamic delay process from the change in the furnace thermal state to the oxygen lance heat load reaching its peak. For example, when a violent reaction occurs in the furnace causing a rapid increase in the furnace gas temperature, the oxygen lance heat load usually reaches its peak after 5 to 10 seconds. Therefore, setting the preset look-ahead time point within this time range can effectively achieve early prediction of the heat load peak, thereby improving the timeliness and foresight of cooling regulation.

[0023] The calculation of the heat load value includes: Calculate the difference between the cooling water outlet temperature and the cooling water inlet temperature to obtain the cooling water temperature rise; Multiply the cooling water flow rate by the density of water to obtain the cooling water mass flow rate; The actual heat load absorbed by the oxygen lance is obtained by multiplying the cooling water mass flow rate, the specific heat capacity of the water, and the cooling water temperature rise.

[0024] The input features are combined with the corresponding training labels to form a training sample; A training sample set is constructed by using multiple training samples.

[0025] Using the training sample set, a heat load assessment model is obtained through supervised learning algorithm training.

[0026] The heat load assessment model is a multilayer perceptron, a long short-term memory network, or a support vector regression machine.

[0027] This invention constructs a training sample set by using multivariate time series based on historical blowing cycles and calculated heat load time series, and uses a supervised learning algorithm to train a heat load assessment model. This enables the model to accurately capture the dynamic evolution of heat load during the converter smelting process, providing a reliable basis for the forward-looking adjustment of cooling water flow rate setpoints.

[0028] Based on the predicted oxygen lance heat load, a cooling water flow rate setpoint is dynamically generated. The step of dynamically generating a cooling water flow rate setpoint based on the predicted oxygen lance heat load includes: Based on the predicted oxygen lance heat load, the corresponding cooling water flow rate setting value is obtained using a preset mapping table. The mapping table is pre-established based on the oxygen lance heat load value and cooling water flow rate setting value corresponding to the oxygen lance body temperature not exceeding the preset safety threshold in historical blowing data.

[0029] This invention establishes a mapping table between predicted heat load and set cooling water flow rate based on historical blowing data where the oxygen lance body temperature does not exceed a safe threshold. This ensures accurate flow rate matching while meeting cooling safety requirements, avoids energy waste caused by excessive cooling water supply, and reduces pump energy consumption and system operating costs.

[0030] In the oxygen top-blown converter steelmaking process, the oxygen lance is one of the core process equipment. It is inserted vertically into the converter through the furnace opening and blows high-purity oxygen into the molten pool at high speed to drive key metallurgical reactions such as decarburization and heating. Because the oxygen lance is in a harsh environment of high temperature (the furnace temperature can reach over 1600°C), strong radiation, and violent chemical reactions for a long time, a forced water cooling system must be used to continuously cool the lance body to prevent burn-out, deformation, or even explosion accidents.

[0031] Traditional oxygen lance cooling water control systems often employ feedback control strategies based on the oxygen lance body temperature or the cooling water outlet temperature. However, such methods suffer from significant lag: when the lance body temperature rises significantly, it indicates that a thermal load shock has already occurred, and increasing the cooling water flow rate at this point is often too late to avoid the risk of localized overheating.

[0032] This invention recognizes that the intensity of the heat load borne by the oxygen lance is highly coupled with the smelting process inside the converter. At different stages of blowing (such as the initial blowing period, main blowing period, and post-blowing period), the intensity of the carbon-oxygen reaction within the furnace varies, resulting in significant differences in the released heat, which directly leads to significant changes in the furnace flue gas temperature (i.e., furnace gas temperature). Changes in furnace gas temperature precede the response of the oxygen lance body temperature and can serve as a precursor signal to changes in heat load.

[0033] Therefore, this invention proposes to simultaneously acquire the furnace gas temperature signal while acquiring the oxygen lance body temperature in real time, and to construct a heat load assessment model based on both. This model can predict the heat load level in the near future and adjust the cooling water flow rate setpoint accordingly, achieving feedforward compensation.

[0034] However, due to the strong nonlinearity and sudden changes in operating conditions during the converter smelting process, simply adjusting the setpoint is insufficient to guarantee the dynamic performance of flow control. Therefore, this invention further introduces a PID parameter adaptive mechanism based on fuzzy rules. The flow deviation, deviation change rate, and the aforementioned predicted heat load are used as inputs to the fuzzy controller, optimizing the proportional, integral, and derivative coefficients of the PID controller in real time. This improves the system's response speed and stability to changes in setpoints and external disturbances.

[0035] In summary, furnace gas temperature serves as a key feedforward variable, enabling the system to predict performance; while the dual-layer collaborative control (setpoint feedforward + PID parameter adaptive control) effectively overcomes the lag defect of pure feedback control, significantly improving the safety and cooling efficiency of oxygen lance operation.

[0036] The flow deviation is calculated based on the actual flow rate of the cooling water and the set flow rate of the cooling water, and the rate of change of the deviation is obtained by differential calculation of the flow deviation. In one embodiment of the present invention, the deviation change rate is calculated based on the flow deviation at multiple consecutive sampling times. Specifically, the control system samples the actual flow rate of cooling water at a fixed period and compares the current actual flow rate with the corresponding flow rate setpoint at each sampling time to obtain the flow deviation at the current time. Simultaneously, the system caches the flow deviation calculated at the previous sampling time. Based on this, by comparing the change magnitude between the flow deviation at the current time and the flow deviation at the previous sampling time, and combining this with the system's preset sampling period, the rate of change of the flow deviation, i.e., the deviation change rate, can be determined. This deviation change rate reflects the trend of flow deviation changing over time, providing crucial dynamic information for subsequent adaptive adjustment.

[0037] Based on the flow deviation, deviation change rate, and oxygen lance heat load prediction, the PID parameter correction amount is obtained using a predefined fuzzy control rule set. The PID parameter corrections include proportional coefficient corrections, integral coefficient corrections, and derivative coefficient corrections; The step of obtaining PID parameter correction values ​​based on the flow deviation, deviation change rate, and oxygen lance heat load prediction value using a predefined fuzzy control rule set includes: The flow deviation, deviation change rate, and oxygen lance heat load prediction value are used as three input variables of the fuzzy controller, and fuzzification is performed according to the preset membership function to obtain their respective fuzzy language values. From a predefined set of fuzzy control rules, select association rules that match the fuzzy linguistic values ​​of the flow deviation, the deviation change rate, and the oxygen lance heat load prediction value in the premise part, wherein the premise part is composed of the fuzzy linguistic values ​​corresponding to the three input variables. Based on the conclusion of the selected association rule, generate output fuzzy sets corresponding to the proportional coefficient correction, integral coefficient correction, and differential coefficient correction, respectively. Defuzzification operations are performed on each output fuzzy set to obtain real-time, numerical correction values ​​for the proportional coefficient, integral coefficient, and differential coefficient.

[0038] In a preferred embodiment of the present invention, the process of obtaining the PID parameter correction amount based on the flow deviation, the deviation change rate, and the oxygen lance heat load prediction value using a predefined fuzzy control rule set specifically includes the following steps: First, the flow deviation e, the deviation change rate ec, and the oxygen lance heat load prediction value Q are used as three input variables of the fuzzy controller, and fuzzification processing is performed on each of them.

[0039] Here, both the flow deviation e and the rate of change of deviation ec are physical quantities with positive and negative values, and their corresponding fuzzy linguistic value sets are defined as follows: {NB (Negative Large), NM (Negative Medium), NS (Negative Small), ZO (Zero), PS (Positive Small), PM (Positive Medium), PB (Positive Large)}; The predicted oxygen lance heat load value Q is a non-negative quantity that represents the current heating intensity of the oxygen lance. Its fuzzy linguistic value set is independently defined as: {VL (very low), L (low), M (medium), H (high), VH (very high)}.

[0040] For each input variable, a triangular or trapezoidal membership function is pre-configured; during fuzzification, the precise input value is substituted into the corresponding membership function to calculate its membership degree on each fuzzy linguistic value, which serves as the basis for subsequent inference.

[0041] For example, when the flow deviation e=1.8, it may have a membership degree of 0.6 to PS and a membership degree of 0.4 to PM, indicating that the value is partly "positive small" and partly "positive medium".

[0042] Secondly, based on the above fuzzification results, association rules that match the current input state are selected from the predefined fuzzy control rule set.

[0043] Each association rule adopts the standard "if-then" form. Its premise consists of the fuzzy linguistic values ​​of the three input variables, while the conclusion specifies the fuzzy linguistic values ​​of the proportional coefficient correction ΔKp, integral coefficient correction ΔKi, and differential coefficient correction ΔKd (using the same set of linguistic values ​​{NB, NM, NS, ZO, PS, PM, PB} as the flow deviation e and deviation change rate ec, but each corresponding to an independent output universe of discourse and membership function). Since fuzzification typically activates multiple linguistic values, multiple rules will be matched simultaneously. For example, if e activates both PS and PM, ec activates both NS and ZO, and Q activates H, then a maximum of 2 × 2 × 1 = 4 rules can be activated.

[0044] For each activated rule, the minimum membership degree of each linguistic value in its premise is taken as the activation strength of the rule. The Mamdani inference method is then used to truncate (i.e., clip the top) the output fuzzy set of the rule's conclusion, forming a local output fuzzy set. Specifically: In a specific embodiment of the present invention, after the fuzzification process is completed and the activated fuzzy rules are determined, inference operations need to be performed on each activated rule to generate its corresponding local output. Specifically, for any activated fuzzy rule, its premise part is composed of multiple input linguistic variables connected by a logical AND; the firing strength of this rule is determined by taking the minimum value of the membership degree corresponding to each linguistic value in the premise part. For example, if the premise of a rule is "if the flow deviation e is PM and the deviation change rate ec is NS and the oxygen lance heat load prediction value Q is H", and the current input, after fuzzification, yields: e has a membership degree of 0.6 to PM, ec has a membership degree of 0.8 to NS, and Q has a membership degree of 0.7 to H, then the firing strength of this rule is: ; Subsequently, the Mamdani fuzzy inference method is used to perform implication operations on the output fuzzy set specified in the conclusion part of the rule (such as the fuzzy set "PB" of the scaling factor correction ΔKp). The specific implementation of this operation is as follows: the original membership function of the conclusion fuzzy set is truncated along the vertical axis to the height corresponding to the activation intensity α, that is, all membership values ​​μoriginal(z) of the conclusion fuzzy set are transformed as follows: ; Where Z is the universe of discourse of the output variables (ΔKp, ΔKi, or ΔKd). This operation is also called "clipping," which means that since the rule premise is only satisfied at a degree α, the contribution of the rule to the output should not exceed this confidence level. The fuzzy set obtained after this processing... This refers to the local output fuzzy set generated by the rule for each output variable (ΔKp, ΔKi, or ΔKd) in the conclusion. Therefore, each activated rule will correspond to three local output fuzzy sets, which are used for subsequent aggregation processing.

[0045] The above process is performed independently for each rule activated by the current input.

[0046] For the proportional coefficient correction ΔKp, integral coefficient correction ΔKi, and derivative coefficient correction ΔKd, perform the following aggregation and deblurring operations respectively: Aggregate all local output fuzzy sets for the same output variable (e.g., ΔKp): Take the maximum membership degree at each point in the universe of discourse to generate the comprehensive output fuzzy set of that variable; then use the Center of Gravity (COG) method to defuzzify it, obtaining the numerical proportional coefficient correction. The processing of integral and differential coefficients is similar.

[0047] This invention uses flow deviation, deviation change rate, and oxygen lance heat load prediction as inputs to a fuzzy controller, and uses a predefined set of fuzzy control rules to correct PID parameters online. This enables the control system to adaptively adjust the control intensity according to the heat load trend and flow tracking status, and maintain good dynamic response performance and robustness even when there are sudden changes in the blowing conditions or external disturbances.

[0048] The PID parameter correction is applied to the current PID parameters to obtain the real-time optimized proportional coefficient, integral coefficient, and derivative coefficient; specifically including: The proportional coefficient correction, integral coefficient correction, and derivative coefficient correction are respectively superimposed on the proportional coefficient, integral coefficient, and derivative coefficient in the current PID parameters to obtain the real-time optimized proportional coefficient, integral coefficient, and derivative coefficient. It should be noted that the PID parameters have initial values. These initial values ​​include the proportional coefficient, integral coefficient, and derivative coefficient, which are preset before the system is put into operation, based on the dynamic characteristics of the cooling water flow control process, using engineering tuning methods (such as the Ziegler-Nichols method or trial and error method), and serve as a reference for adaptive adjustment.

[0049] In one embodiment, the response curve of the cooling water flow rate can be recorded by applying a step opening command to the cooling water regulating valve, the process gain and time constant can be extracted, and the initial PID parameters can be calculated according to the Ziegler-Nichols step response formula.

[0050] Using the real-time optimized proportional coefficient, integral coefficient, and derivative coefficient, and taking the flow deviation as input, a control output signal is generated through a PID control algorithm, which drives the actuator to adjust the cooling water flow rate so that the actual cooling water flow rate follows the dynamically generated cooling water flow rate setpoint.

[0051] The core of the adaptive adjustment method for converter oxygen lance cooling water provided by this invention lies in constructing a two-layer adaptive control architecture of "setpoint feedforward adjustment + online optimization of control parameters" to cope with complex working conditions such as drastic fluctuations in heat load, strong nonlinearity and large time delay during converter smelting.

[0052] Specifically, the first layer is a dynamic generation mechanism for the cooling water flow rate setpoint. Using a heat load assessment model, based on the current furnace gas temperature, oxygen lance body temperature, and their rate of change over time, the system predicts the oxygen lance heat load value at a preset forward time point. This predicted value reflects the intensity of the impending thermal shock. The system then dynamically adjusts the cooling water flow rate setpoint accordingly, for example, by calculating the required flow rate using a lookup table or heat balance equation. This preemptively increases cooling capacity before the actual heat load increases, achieving feedforward compensation and ensuring the system has sufficient thermal margin.

[0053] However, relying solely on dynamic adjustments to the cooling water flow rate setpoint is insufficient to guarantee control performance under all operating conditions. The reason is: The converter smelting process is characterized by strong nonlinearity and suddenness. Under abnormal conditions such as back-drying, violent carbon-oxygen reactions, or splashing, the heat flux density inside the furnace may increase sharply, leading to a rapid increase in the heat load on the oxygen lance surface, which in turn requires a significant increase in cooling water flow to maintain a safe temperature. However, industrial cooling water systems themselves have factors such as nonlinearity of actuators (e.g., dead zone of regulating valves, pump response delay) and time-varying pipe flow resistance. If a PID controller with fixed parameters is used, it is often difficult to balance response speed and stability: problems such as slow regulation, excessive overshoot, or continuous oscillation may occur. Even if the final flow rate reaches the set value, an excessively long dynamic adjustment process can cause the oxygen lance to be subjected to excessive thermal stress during the adjustment period.

[0054] To address this, the present invention introduces a second-layer adaptive mechanism: online optimization of PID parameters based on fuzzy rules. This mechanism uses the following three types of information as input to the fuzzy controller: (1) Flow deviation (the difference between the actual flow rate of cooling water and the dynamic set value) reflects the current tracking error; (2) The rate of change of deviation reflects the development trend of error; (3) The predicted value of oxygen lance heat load is used as a feedforward index of the intensity of external disturbance.

[0055] Based on a predefined set of fuzzy control rules, the system outputs proportional coefficient, integral coefficient, and derivative coefficient corrections in real time, which are then added to the current PID parameters to generate optimized control parameters. For example: When the flow rate deviation is large and the heat load forecast is high, the proportional gain should be increased appropriately to balance disturbance suppression and response speed. When the deviation tends to stabilize but steady-state error exists, the integral action is enhanced to eliminate steady-state error; When the deviation changes drastically, the differential action is increased to effectively suppress overshoot and oscillation.

[0056] In summary, this invention employs a two-layer collaborative control mechanism: The first layer (dynamically adjustable setpoints) addresses the question of whether the cooling capacity is sufficient, providing proactive protection against thermal shock. The second layer (PID parameter adaptive) addresses the issues of "whether the flow regulation is accurate, fast, and robust," improving the system's adaptability to sudden changes in operating conditions and nonlinear execution.

[0057] The combination of the two not only avoids the performance degradation of traditional fixed setpoint + fixed PID control under strong disturbance conditions, but also overcomes the limitation of only adjusting the setpoint while ignoring the dynamic quality of control, thus realizing high-reliability closed-loop control of the converter oxygen lance cooling water system throughout the entire cycle and under all operating conditions.

[0058] This invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, which, when executed by the at least one processor, causes the electronic device to perform the method of this invention.

[0059] The present invention also provides a non-transitory machine-readable medium storing a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform the method of the present invention.

[0060] This invention also provides a computer program product, including a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform the method of this invention.

[0061] refer to Figure 2 The present invention will now describe a structural block diagram of an electronic device that can serve as a server or client in embodiments of the present invention, which is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0062] like Figure 2 As shown, the electronic device includes a computing unit 401, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. The RAM 403 may also store various programs and data required for the operation of the electronic device. The computing unit 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0063] Multiple components in the electronic device are connected to I / O interface 405, including: input unit 406, output unit 407, storage unit 408, and communication unit 409. Input unit 406 can be any type of device capable of inputting information into the electronic device. Input unit 406 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of the electronic device. Output unit 407 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 408 may include, but is not limited to, disks and optical discs. Communication unit 409 allows the electronic device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0064] The computing unit 401 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, CPUs, graphics processing units (GPUs), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the methods and processes described above. For example, in some embodiments, the method embodiments of the present invention may be implemented as a computer program tangibly contained in a machine-readable medium, such as storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on an electronic device via ROM 402 and / or communication unit 409. In some embodiments, the computing unit 401 may be configured to perform the methods described above by any other suitable means (e.g., by means of firmware).

[0065] Computer programs for implementing the methods of embodiments of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0066] In the context of embodiments of the present invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable signal medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0067] This invention constructs a heat load assessment model, utilizing real-time collected furnace gas temperature, oxygen lance body temperature, and their respective rates of change over time to predict the oxygen lance heat load at a preset forward time point. Based on this, a cooling water flow rate setpoint is dynamically generated, achieving forward-looking cooling regulation with furnace gas temperature as the key feedforward variable. Simultaneously, by combining the flow deviation and deviation rate between the actual cooling water flow rate and the setpoint, and introducing the oxygen lance heat load prediction value as a fuzzy control input, the PID parameters are corrected in real time through a predefined fuzzy control rule set, forming an adaptive optimization of the flow control loop. This overcomes the lag of traditional pure feedback control while significantly improving the system's response speed, tracking accuracy, and operational stability to sudden changes in heat load and external disturbances.

[0068] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of the components in a specific posture (as shown in the attached figures). If the specific posture changes, the directional indication will also change accordingly. Furthermore, descriptions involving "first," "second," or "a" in the present invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. In the description of the present invention, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly defined. In the present invention, unless otherwise explicitly specified and defined, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can be a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal communication of two components or the interaction between two components, unless otherwise explicitly defined. For those skilled in the art, the specific meanings of the above terms in this invention can be understood according to the specific circumstances. Furthermore, the technical solutions of the various embodiments of this invention can be combined with each other, but only on the basis that those skilled in the art can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

Claims

1. A method for adaptive adjustment of oxygen lance cooling water in a converter, characterized in that, include: Real-time acquisition of multi-source parameters reflecting the oxygen lance heat load status and converter smelting process, including at least: oxygen lance body temperature, furnace gas temperature, actual cooling water flow rate, cooling water inlet temperature, and cooling water outlet temperature. Based on the current furnace gas temperature, oxygen lance body temperature, furnace gas temperature change rate over time and oxygen lance body temperature change rate over time, the predicted oxygen lance heat load at a future preset forward time point is evaluated using a heat load assessment model. Based on the predicted oxygen lance heat load, a cooling water flow rate setpoint is dynamically generated. The flow deviation is calculated based on the actual flow rate of the cooling water and the set flow rate of the cooling water, and the rate of change of the deviation is obtained by differential calculation of the flow deviation. Based on the flow deviation, deviation change rate, and oxygen lance heat load prediction, the PID parameter correction amount is obtained using a predefined fuzzy control rule set. The PID parameter correction is applied to the current PID parameters to obtain the proportional coefficient, integral coefficient, and derivative coefficient optimized in real time. Using the real-time optimized proportional coefficient, integral coefficient, and derivative coefficient, and taking the flow deviation as input, a control output signal is generated through a PID control algorithm, which drives the actuator to adjust the cooling water flow rate so that the actual cooling water flow rate follows the dynamically generated cooling water flow rate setpoint.

2. The adaptive adjustment method for converter oxygen lance cooling water according to claim 1, characterized in that, The acquisition of the heat load assessment model includes: Collect multivariate time series data corresponding to multiple historical blowing cycles. The multivariates include furnace gas temperature, oxygen lance body temperature, cooling water inlet temperature, cooling water outlet temperature, and cooling water flow rate. Based on the cooling water inlet temperature, cooling water outlet temperature, cooling water flow rate, and specific heat capacity of water, the actual heat load value absorbed by the oxygen lance at each moment is calculated to form a heat load time series; wherein, for each historical blowing cycle, its corresponding multivariate time series and heat load time series are aligned at the same time. A training sample set was constructed using multivariate time series data from multiple historical blowing cycles and heat load time series data. Using the training sample set, a heat load assessment model is obtained through supervised learning algorithm training.

3. The adaptive adjustment method for converter oxygen lance cooling water according to claim 2, characterized in that, The construction of a training sample set using multivariate time series data from multiple historical blowing cycles and heat load time series includes: For the time series data of each historical blowing cycle, iterate through each collection point; For the currently traversed acquisition time, extract the input features, which include: The furnace gas temperature at the current data collection moment; The rate of change of furnace gas temperature over time is calculated from the furnace gas temperature time series data with the current acquisition time as the endpoint and a preset time window as the length. The oxygen lance body temperature at the current sampling moment; The rate of change of oxygen lance body temperature over time is calculated from the time series data of oxygen lance body temperature with the current acquisition time as the endpoint and the length as a preset time window. Determine the training label corresponding to the current acquisition time. The training label is the oxygen lance heat load value corresponding to a preset look-ahead time point after the current acquisition time and within the same blowing cycle. The preset look-ahead point is determined based on the converter blowing thermal response characteristics. The input features are combined with the corresponding training labels to form a training sample; A training sample set is constructed by using multiple training samples.

4. The adaptive adjustment method for converter oxygen lance cooling water according to claim 2, characterized in that, The calculation of the heat load value includes: Calculate the difference between the cooling water outlet temperature and the cooling water inlet temperature to obtain the cooling water temperature rise; Multiply the cooling water flow rate by the density of water to obtain the cooling water mass flow rate; The actual heat load absorbed by the oxygen lance is obtained by multiplying the cooling water mass flow rate, the specific heat capacity of the water, and the cooling water temperature rise.

5. The adaptive adjustment method for converter oxygen lance cooling water according to claim 2, characterized in that, The heat load assessment model is a multilayer perceptron, a long short-term memory network, or a support vector regression machine.

6. The adaptive adjustment method for converter oxygen lance cooling water according to claim 3, characterized in that, The step of dynamically generating a cooling water flow rate setpoint based on the predicted oxygen lance heat load includes: Based on the predicted oxygen lance heat load, the corresponding cooling water flow rate setting value is obtained using a preset mapping table. The mapping table is pre-established based on the oxygen lance heat load value and cooling water flow rate setting value corresponding to the oxygen lance body temperature not exceeding the preset safety threshold in historical blowing data.

7. The adaptive adjustment method for converter oxygen lance cooling water according to claim 3, characterized in that, The PID parameter corrections include proportional coefficient corrections, integral coefficient corrections, and derivative coefficient corrections; The step of obtaining PID parameter correction values ​​based on the flow deviation, deviation change rate, and oxygen lance heat load prediction value using a predefined fuzzy control rule set includes: The flow deviation, deviation change rate, and oxygen lance heat load prediction value are used as three input variables of the fuzzy controller, and fuzzification is performed according to the preset membership function to obtain their respective fuzzy language values. From a predefined set of fuzzy control rules, an association rule is selected that matches the fuzzy linguistic value of the flow deviation, the fuzzy linguistic value of the deviation change rate, and the fuzzy linguistic value of the oxygen lance heat load prediction value. The premise part is composed of the fuzzy linguistic values ​​corresponding to the three input variables. Based on the conclusion of the selected association rule, generate output fuzzy sets corresponding to the proportional coefficient correction, integral coefficient correction, and differential coefficient correction, respectively. Defuzzification operations are performed on each output fuzzy set to obtain real-time, numerical correction values ​​for the proportional coefficient, integral coefficient, and differential coefficient.

8. An electronic device, comprising: A processor and a memory storing a program, characterized in that the program includes instructions that, when executed by the processor, cause the processor to perform an adaptive adjustment method for converter oxygen lance cooling water according to any one of claims 1 to 7.

9. A non-transitory machine-readable medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute a converter oxygen lance cooling water adaptive adjustment method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the adaptive adjustment method for converter oxygen lance cooling water according to any one of claims 1 to 7.

Images