A method and system for adaptive optimization control of refractory production parameters

Abstract

The present application belongs to the technical field of general control or regulation system, and particularly relates to a refractory material production parameter adaptive optimization control method and system, which comprises the following steps: obtaining a torque data segment at each sampling time; determining a preliminary violent factor and a preliminary violent degree at the sampling time; determining a jam interference factor and a real violent degree at the sampling time; determining an optimized sliding time window length at the sampling time; and adaptively controlling the stirring speed of the stirring motor. The present application dynamically adjusts the sliding time window length, makes up for the deficiency of the fixed length of the sliding time window length, strips the mechanical jam noise under complex working conditions, avoids the motor overload shutdown risk and the aggregate crushing phenomenon, and improves the mixing uniformity of the special refractory material product.

Description

Technical Field

[0001] This invention relates to the field of general control or regulation systems. More specifically, this invention relates to an adaptive optimization control method and system for refractory material production parameters. Background Technology

[0002] In the production process of special refractory materials, the mixing process is the core link that determines the final physical and chemical properties of the product. During the mixing stage, it is necessary to uniformly mix various refractory aggregates with different particle size distributions, as well as fine powders and binders. With the continuous addition of water and binders, the rheological properties and viscosity of the mud will undergo complex nonlinear changes. If the speed of the stirring motor remains constant, the mixing efficiency will be low when the mud viscosity is low. When the mud viscosity increases sharply, it is easy to overload the stirring motor or even cause the high-speed rotating stirring blades to break the fragile lightweight aggregates.

[0003] To address the issue of the aforementioned inability to dynamically adjust the stirring speed, the local weighted regression adaptive control algorithm can optimize and dynamically adjust the stirring speed in real time. By performing local weighted regression fitting on historical torque data within a sliding time window, it predicts the material viscosity change trend at the next moment and then adaptively issues speed adjustment commands in advance. This effectively avoids motor overload, protects aggregate integrity, and significantly improves the uniformity of mixing.

[0004] Existing local weighted regression adaptive control algorithms typically set the length of the sliding time window to a fixed constant value when predicting and fitting torque trends. This fixed setting cannot adapt to the complex working conditions of variable and nonlinear mud viscosity in the mixing process, causing the controller to fall into a contradiction between sensitivity and lag. Specifically, the fixed sliding time window length may be too short in the initial wet mixing stage after adding water, causing the algorithm to be overly sensitive to conventional mechanical vibrations, resulting in frequent speed fluctuations and increased equipment wear. However, after adding binder, the fixed sliding time window length may be too long, which will lead to serious lag in regression prediction and thus make it impossible to make deceleration decisions in a timely manner based on recent drastic changes.

[0005] In related technologies, Chinese patent document CN114953189B, entitled "Method, Device, Mixing Tank and Working Machinery for Speed ​​Control of Mixing Tank," discloses a speed control method for smooth start-up and shutdown and anti-vibration of mixing equipment. During the control phase of speed reduction or shutdown of the mixing equipment, after receiving a speed stop or speed reduction command, the control system uses a speed control mode to control the mixing tank to initially reduce its speed. Subsequently, when the speed of the mixing tank decreases to a preset speed, mode switching and smooth shutdown control are performed based on the actual output torque conditions. This patent document uses a conditional triggering method based on a single speed and torque logic to adjust and control the mixing speed and prevent vibration. Its triggering mechanism is typically based on a fixed preset speed threshold set according to the smoothness characteristics of conventionally mixed materials. The parameters have the following drawbacks: they do not take into account the complex nonlinear viscosity changes caused by the physical and chemical reactions of the refractory material mixing system under the complex working conditions of adding binders or wet mixing with water, as well as the instantaneous pulse interference caused by the temporary jamming of large hard aggregate particles on the stirring blades. The fixed speed judgment threshold and rigid mode switching mechanism are prone to misinterpreting extremely isolated mechanical torque peaks caused by the jamming of large hard aggregate particles as a surge in the overall viscosity of the refractory material when faced with extremely unstable and non-Gaussian stirring torque fluctuations of the refractory material, thus making incorrect decisions to reduce speed and stop the machine. Or, when the actual material viscosity increases nonlinearly, the fixed parameters have serious response lag, which may mistakenly leave dangerous overload signals as normal operation, leading to motor overload. The adaptive capability is insufficient. Summary of the Invention

[0006] To address the technical problem that the fixed sliding time window length cannot adapt to complex working conditions, the present invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an adaptive optimization control method for refractory material production parameters, comprising: collecting torque data of a stirring motor, preprocessing it to obtain torque data segments at each sampling time; determining an initial severity factor at the sampling time based on the average absolute value of the difference between the value of each data point in the torque data segment at the sampling time and the mean value of all data points in the torque data segment; performing maximum-minimum normalization calculation on the initial severity factor to determine the initial severity level at the sampling time; determining a stagnation interference factor at the sampling time based on the difference between the total number of data points in the torque data segment at the sampling time and the total number of intervals between data points exhibiting a positive growth gradient in the torque data segment; performing maximum-minimum normalization calculation on the stagnation interference factor; determining the actual severity level at the sampling time based on the stagnation interference factor after maximum-minimum normalization calculation and the initial severity level; determining the optimized sliding time window length at the sampling time based on the actual severity level and the upper and lower limits of the sliding time window length; and using the optimized sliding time window length and a local weighted regression algorithm to adaptively control the stirring speed of the stirring motor.

[0008] This invention analyzes the intensity of viscosity fluctuations in slurry by calculating the initial severity factor and the initial severity level. Simultaneously, it introduces a jamming interference factor to effectively eliminate isolated noise interference caused by sudden mechanical jamming. Combining these two factors yields the true severity level, enabling the identification of the actual viscosity surge during mixing. Based on this, the sliding time window length is dynamically optimized according to the true severity level, allowing the local weighted regression algorithm to adaptively match the optimal data span under complex and variable working conditions. This overcomes the shortcomings of fixed time windows, such as oversensitivity to high-frequency mechanical vibrations and delayed response to viscosity changes. Finally, based on the optimized sliding time window length, local weighted regression fitting is performed to predict future torque change trends. This enables the early issuance of speed adjustment commands to avoid motor overload and protect the integrity of fragile aggregates when facing a sharp increase in viscosity, while maintaining the rated speed within the safe torque range to ensure excellent mixing uniformity. This improves the control accuracy of the mixing process and the safety of equipment operation in refractory material production.

[0009] Preferably, the initial drastic factor satisfies the expression: In the formula, For the first The initial drastic factor at each sampling time, The total number of data points in the torque data segment at each sampling time. For the first In the torque data segment at the sampling time, the first The value of each data point. For the first The average value of all data points within the torque data segment at each sampling time.

[0010] This invention constructs an average value calculation model that includes the sum of the absolute values ​​of the differences between the values ​​of each data point in the torque data segment and the mean value. This model can intuitively extract the overall discrete characteristics of the torque data, making the initial drastic factor calculated for stages with stronger local discreteness larger. This lays a reliable data foundation for accurately reflecting the intensity of drastic viscosity fluctuations caused by changes in the internal resistance of the mud.

[0011] Preferably, the initial intensity satisfies the expression: In the formula, For the first The initial intensity at each sampling moment, For the first The initial drastic factor at each sampling time, and The first The minimum and maximum values ​​of the initial severity factor at each sampling time for all data points within the torque data segment at each sampling time. To prevent the initial intensity from being a constant with a value of 0.

[0012] This invention maps the initial drastic factor to a relatively standardized numerical range by introducing a maximum-minimum normalization calculation mechanism. This effectively clarifies the relative magnitude of the torque fluctuation intensity at the current sampling moment within the global range of recent local historical data. This allows the initial drastic factor with a higher relative ranking to be given a greater confidence weight, thereby significantly improving the sensitivity and reliability of the initial drastic factor as an early warning signal for initial viscosity changes.

[0013] Preferably, the lag interference factor satisfies the expression: In the formula, For the first The lag interference factor at each sampling time. The total number of data points in the torque data segment at each sampling time. For the first The total number of data points with a positive increasing gradient among all adjacent time points within a sampling time segment of torque data. To prevent constants with a denominator of 0.

[0014] This invention, through in-depth analysis of the characteristic differences between isolated torque pulses generated by large hard aggregates accidentally getting stuck in mechanical gaps and the sustained monotonically increasing gradient of the viscosity surge of real materials, cleverly constructs a jamming interference assessment model with the total number of data points with positively increasing gradients as the inverse correlation term. This makes the jamming interference factor calculated by the data segment with more isolated spike interference larger, thereby achieving effective identification of false overload risks.

[0015] Preferably, the actual intensity satisfies the expression: In the formula, For the first The true intensity at each sampling moment For the first The initial intensity at each sampling moment, For the first The lag interference factor at each sampling time. and The first The minimum and maximum values ​​of the jamming interference factor at each sampling time for all data points within the torque data segment at each sampling time.

[0016] This invention constructs a composite function that includes a positive correlation term for the initial severity and a normalized negative correlation term for the jamming interference factor. It integrates an index reflecting the overall dispersion of torque data with a constraint index that eliminates sudden mechanical interference. This allows the true viscosity surge state, which is less affected by isolated jamming spikes and exhibits continuous and severe fluctuations, to obtain a greater true severity. This solves the problem that existing locally weighted regression prediction fitting curves are easily skewed by high-frequency noise.

[0017] Preferably, the optimized sliding time window length satisfies the expression: In the formula, For the first The optimized sliding time window length at each sampling time point For the first The true intensity at each sampling moment and These are the upper and lower limits of the sliding time window length, respectively.

[0018] Preferably, the adaptive control of the stirring speed of the stirring motor includes: using the optimized sliding time window length, combined with a local weighted regression algorithm to perform local weighted regression fitting on the torque data, and solving for the slope of the regression curve of the torque change trend at the sampling time; predicting the torque value within a future preset time based on the slope of the regression curve at the sampling time; in response to the torque value within the future preset time at the sampling time being greater than the safe torque warning threshold, calculating a negative speed compensation amount to drive the frequency converter to reduce the motor speed to resolve the overload crisis; in response to the torque value within the future preset time at the sampling time not being greater than the safe torque warning threshold, maintaining the basic rated speed to ensure optimal mixing uniformity.

[0019] Preferably, the acquisition of torque data of the mixing motor includes: deploying a high-precision dynamic torque sensor on the main shaft motor of the mixer to acquire torque data of the mixing motor during the mixing process.

[0020] Preferably, the preprocessing to obtain the torque data segment for each sampling moment includes: for all torque data points, selecting a number of consecutive sampling moments immediately preceding them to obtain the torque data segment for each sampling moment.

[0021] Secondly, the present invention provides an adaptive optimization control system for refractory material production parameters, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned adaptive optimization control method for refractory material production parameters is implemented.

[0022] By adopting the above technical solution, a computer program is generated from the above-mentioned adaptive optimization control method for refractory material production parameters and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0023] The beneficial effects of this invention are as follows: This invention solves the technical problem that existing local weighted regression adaptive control algorithms are prone to contradictions when faced with complex working conditions in the mixing process where the viscosity of mud is variable and nonlinear, and are easily limited by the fixed time window length, resulting in excessive sensitivity to high-frequency mechanical vibration and lag in response to viscosity changes. This is achieved by introducing an adaptive adjustment mechanism for the length of the sliding time window based on multi-feature fusion.

[0024] This invention analyzes the dispersion of data points within torque data segments and extracts the initial intensity of the disturbance, thus establishing a preliminary mapping relationship between the intensity of fluctuations in the internal resistance of the mud and the overall dispersion characteristics of the torque. Based on this, it further integrates the data point interval characteristics that present a positive growth gradient to construct a jamming interference factor. This distinguishes between the continuously increasing gradient index reflecting the surge in the true viscosity of the material and the isolated spike characteristics of mechanical sudden jamming caused by large aggregate particles, thereby enabling the identification of the true intensity of the disturbance after removing mechanical interference during the mixing process.

[0025] This invention can calculate an optimized sliding time window length that matches the actual viscosity fluctuation state of the slurry at each sampling moment. By dynamically adjusting the historical data backtracking span of the local weighted regression algorithm, a deep fit between the prediction model and the actual working conditions is achieved. For sampling moments with real and drastic viscosity changes, a smaller data span is assigned to quickly capture the dangerous trend and prevent motor overload. For sampling moments that are stable or subject to normal mechanical vibration, a larger time window is maintained to filter high-frequency noise. Ultimately, this invention improves the accuracy and noise resistance of the adaptive control of the stirring speed, ensuring the optimal mixing uniformity of special refractory materials and the safe operation of production equipment. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating an adaptive optimization control method for refractory material production parameters according to the present invention; Figure 2 This is a comparison chart of the torque values ​​predicted by existing technologies and the torque values ​​predicted by this invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] This invention discloses an adaptive optimization control method for refractory material production parameters, referring to... Figure 1 This includes steps S001-S005: S001: Obtain the torque data segment at each sampling time.

[0030] Specifically, a high-precision dynamic torque sensor is deployed on the main shaft motor of the mixer to collect torque data of the stirring motor during the mixing process. For the sampling time corresponding to all torque data points, several consecutive sampling times immediately preceding it are selected to obtain the torque data segment for each sampling time. In this embodiment, the torque data acquisition frequency is set to 20 times per second, and the number of sampling times selected when constructing the torque data segment is set to 100. In other embodiments, the implementer can set the torque data acquisition frequency and the torque data segment construction method according to the specific implementation situation. To prevent the situation where the number of sampling times immediately preceding the sampling time is less than 100 at the beginning of mixing, the first 100 data points are only used for the construction of torque data segments when the torque data of the stirring motor is collected.

[0031] S002: Determine the initial intensity factor and initial intensity level at the sampling time.

[0032] It should be noted that, due to the difficulty of adapting to the complex working conditions of variable and nonlinear mud viscosity in the mixing process when the existing local weighted regression algorithm predicts the trend of torque change, the preset fixed sliding time window length is difficult to adapt to the speed controller when facing sudden viscosity changes, resulting in a serious lag. This reduces the accuracy and safety of the adaptive adjustment of the stirring speed. Based on the characteristic that changes in the internal resistance of the mud will directly cause changes in the overall dispersion of the stirring motor torque data, this invention combines the average of the absolute values ​​of the differences between each data point value in the torque data segment and the mean value of all data points in the torque data segment to determine the initial severity factor and initial severity at the sampling time, which is used to characterize the fluctuation intensity of the mud viscosity at the sampling time.

[0033] Specifically, the initial drastic factor satisfies the expression: ; In the formula, For the first The initial drastic factor at each sampling time, The total number of data points in the torque data segment at each sampling time. , For the first In the torque data segment at the sampling time, the first The value of each data point. For the first The average value of all data points within the torque data segment at each sampling time.

[0034] In the formula, The larger the value, the higher the value. The greater the dispersion of torque data in the torque data segment at each sampling time, the more drastic the viscosity fluctuation of the mud during this stage. Therefore, the... The larger the initial intensity factor at each sampling time, the greater the initial intensity factor.

[0035] Specifically, the initial intensity satisfies the expression: ; In the formula, For the first The initial intensity at each sampling moment, For the first The initial drastic factor at each sampling time, and The first The minimum and maximum values ​​of the initial severity factor at each sampling time for all data points within the torque data segment at each sampling time. To prevent the initial severity level from being a constant of 0, in this embodiment, the constant for preventing the initial severity level from being a constant of 0 is set to 0.01. In other embodiments, implementers can set the constant for preventing the initial severity level from being a constant of 0 according to the specific implementation situation.

[0036] In the formula, The larger it is, the more likely it is to be the first The initial intensity factor value at the sampling time is [value missing]. The larger the relative magnitude of the initial severity factor at the corresponding sampling time for all data points within the torque data segment at the sampling time, the stronger the initial severity factor at the sampling time. The larger the initial drastic factor at the sampling time, the greater the reliability. The greater the initial intensity at each sampling moment.

[0037] S003: Determine the lag interference factor and the actual intensity at the sampling time.

[0038] It should be noted that in actual refractory material mixing scenarios, large, hard aggregate particles may accidentally become stuck in the mechanical gaps of the mixer, causing sudden mechanical jamming and generating severe but isolated torque pulse interference. This special case, along with the actual material viscosity surge caused by the addition of binders, will cause significant torque deviations in the intuitive data. This can lead to the local weighted regression fitting curve being easily skewed by this isolated noise, causing the speed controller to issue incorrect deceleration commands, thus affecting mixing efficiency. Since the actual material viscosity surge has obvious temporal continuity, i.e., it presents a continuously monotonically increasing gradient, while the torque peak generated by sudden mechanical jamming is relatively isolated, i.e., it increases instantaneously and then drops sharply, without the characteristic of a continuously increasing gradient, this invention combines the difference between the total number of data points in the torque data segment and the total number of data points in the torque data segment showing a positive growth gradient to optimize the initial severity, determine the jamming interference factor and the actual severity at the sampling time, and use this to characterize the actual viscosity surge state after removing the sudden mechanical interference.

[0039] Specifically, the lag interference factor satisfies the expression: ; In the formula, For the first The lag interference factor at each sampling time. The total number of data points in the torque data segment at each sampling time. For the first The total number of data points with a positive increasing gradient among all adjacent time points within a sampling time segment of torque data. To prevent the denominator from being zero, in this embodiment, the constant for preventing the denominator from being zero is set to 0.001. In other embodiments, implementers can set the constant for preventing the denominator from being zero according to the specific implementation situation.

[0040] In the formula, The smaller, the more likely it is to be the first The fewer adjacent sampling points exhibiting a positive growth gradient in the torque data segment at the nth sampling time, the better the torque data segment. The greater the likelihood that the torque fluctuations in the torque data segment at each sampling time are caused by a few isolated jamming spikes, the more likely the first... The larger the lag interference factor at each sampling time, the greater the lag interference factor.

[0041] Specifically, the actual intensity satisfies the expression: ; In the formula, For the first The true intensity at each sampling moment For the first The initial intensity at each sampling moment, For the first The lag interference factor at each sampling time. and The first The minimum and maximum values ​​of the jamming interference factor at each sampling time for all data points within the torque data segment at each sampling time.

[0042] In the formula, The larger the value, the higher the value. The more drastic the viscosity fluctuation in the torque data segment at the sampling time, the more it indicates that the... The greater the actual intensity at each sampling moment, the better. In the formula, The smaller the value, the better. The value of the lag interference factor at the sampling time is as follows: The smaller the relative magnitude of the lag interference factor at all data points within the torque data segment at the sampling time, the better. The less likely the torque fluctuations in the torque data segment at each sampling time are caused by a few isolated jamming spikes, the lower the probability that they are due to the first sampling time. The greater the actual intensity at each sampling moment.

[0043] S004: Determine the length of the optimized sliding time window at the sampling time.

[0044] It should be noted that after obtaining the actual intensity, this invention will calculate the optimized sliding time window length based on the actual intensity and the upper and lower limits of the sliding time window length. This ensures that the optimized sliding time window length can match the actual mud viscosity fluctuation at the sampling time, thereby enhancing the robustness of the algorithm under complex working conditions.

[0045] Specifically, the optimized sliding time window length satisfies the expression: ; In the formula, For the first The optimized sliding time window length at each sampling time point For the first The true intensity at each sampling moment and These are the upper and lower limits of the sliding time window length. In this embodiment, the upper and lower limits are set to 150 and 30, respectively. In other embodiments, implementers can set the upper and lower limits of the sliding time window length according to specific implementation conditions. For example, when the conventional electromagnetic or mechanical vibration noise in the mixing environment is large and the system's noise resistance and smoothness requirements are high, the upper limit of the sliding time window length can be appropriately increased to improve the filtering effect of high-frequency noise. When the computational efficiency of the algorithm's daily operation is high, the upper limit of the sliding time window length can be appropriately decreased to improve the algorithm's efficiency. For example, when the response speed of the stirring motor to overload protection is relatively strict, the lower limit of the sliding time window length can be appropriately decreased to improve the algorithm's ability to capture viscosity change trends at high speed and prevent overload. When the mixing equipment has high requirements for smooth speed transition and needs to avoid frequent and violent shaking of the bottom speed, the lower limit of the sliding time window length can be appropriately increased to improve the stability of speed control.

[0046] In the formula, The larger it is, the more likely it is to be the first The greater the likelihood that the torque fluctuations in the torque data segment at the sampling time point represent actual mud viscosity fluctuations, the more important it is to ensure that the local weighted regression algorithm can quickly capture the [missing information]. To capture the latest and most dangerous trends at each sampling time and prevent motor overload, the first sampling time is assigned... The data span is smaller at each sampling time, so the first sampling time is smaller. The smaller the optimized sliding time window length at each sampling time, the better.

[0047] It should be noted that after obtaining the optimized sliding time window length, for the first... The optimized sliding time window length at each sampling time point is rounded down to ensure that the output time window length is an integer.

[0048] S005: Adaptive control of the stirring speed of the stirring motor.

[0049] Specifically, adaptive control includes: By utilizing the optimized sliding time window length and combining it with a local weighted regression algorithm, the torque data is fitted with a local weighted regression to obtain the slope of the regression curve of the torque change trend at the sampling time. The torque value within a preset time period at the sampling time is predicted based on the slope of the regression curve; If the torque value within a preset time period after the sampling time exceeds the safe torque warning threshold, a negative speed compensation is calculated to drive the frequency converter to reduce the motor speed and resolve the overload crisis. If the torque value within a preset time period after the sampling time does not exceed the safe torque warning threshold, the basic rated speed is maintained to ensure optimal mixing uniformity.

[0050] In this embodiment, the bandwidth parameter in the local weighted regression algorithm is set to 0.8, the base rated speed is set to 45 (r / min), and the safe torque warning threshold is set to 85% of the rated torque. In other embodiments, implementers can set the bandwidth parameter, base rated speed, and safe torque warning threshold according to the specific implementation situation. For example, when the sensitivity requirement for capturing torque change trends is high, the bandwidth parameter in the local weighted regression algorithm can be appropriately reduced to improve the overload response speed. When there is a large amount of random high-frequency noise in the mixing environment, the bandwidth parameter can be appropriately increased to improve the smoothness and stability of the prediction. For example, when the mixing period is tight and the material is not easy to break, the base rated speed can be appropriately increased to improve production efficiency. When there are many lightweight aggregates in the material and they are easy to break, the base rated speed can be appropriately reduced to protect the integrity of the aggregates. For example, when the motor's overload resistance is weak or the equipment safety requirements are strict, the safe torque warning threshold can be appropriately reduced to trigger speed compensation in advance. When the motor performance redundancy is large and it is necessary to reduce the impact of frequent speed reduction on the mixing uniformity, the safe torque warning threshold can be appropriately increased to ensure the continuity of the mixing process.

[0051] like Figure 2As shown in the figure, this diagram illustrates the difference in predictive capabilities between existing technologies and the present invention for future torque trends under complex operating conditions involving jamming noise and a rise in actual viscosity. The background data curve in the figure presents the actual torque fluctuation trajectory affected by multiple factors, and the horizontal dashed line marks the safe torque warning threshold. When the system encounters isolated spike interference caused by large hard aggregate particles accidentally getting stuck in the gap, the existing technology, unable to identify the characteristics, has its prediction curve severely skewed by the instantaneous extreme values, producing a false overload peak. In contrast, the method of the present invention successfully removes such sudden noise through a pre-applied jamming interference factor, maintaining a stable prediction trajectory without misjudgment. Subsequently, when the addition of binder to the mud caused a real and continuous increase in the internal resistance of the material, the existing technology, due to the drag of fixed and lengthy historical data, showed a serious slow response in predicting the trajectory and failed to reach the warning line for a long time. In contrast, the adaptive method of the present invention, by dynamically shrinking the time window span, keenly matched the steep real climbing slope and exceeded the safe torque warning threshold in time. This comparative result confirms that the present invention overcomes the defects of oversensitivity to high-frequency mechanical vibration and lag in response to viscosity change, endows the speed controller with anti-interference robustness and trend capture sensitivity, and provides accurate basis for equipment overload prevention.

[0052] This invention also discloses an adaptive optimization control system for refractory material production parameters, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an adaptive optimization control method for refractory material production parameters according to the present invention.

[0053] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. An adaptive optimization control method for refractory material production parameters, characterized in that, include: The torque data of the stirring motor is collected, preprocessed, and the torque data segment at each sampling time is obtained; The initial severity factor at the sampling time is determined by averaging the absolute values ​​of the differences between the value of each data point in the torque data segment at the sampling time and the mean value of all data points in the torque data segment. The initial severity factor is then subjected to maximum-minimum normalization calculation to determine the initial severity at the sampling time. Based on the difference between the total number of data points in the torque data segment at the sampling time and the total number of data point intervals showing a positive growth gradient in the torque data segment, the lag interference factor at the sampling time is determined. The lag interference factor is then subjected to maximum and minimum normalization calculation. Based on the lag interference factor after the maximum and minimum normalization calculation, combined with the preliminary severity, the actual severity at the sampling time is determined. Based on the actual intensity of the event, and combined with the upper and lower limits of the sliding time window length, the optimized sliding time window length for the sampling time is determined. Using the optimized sliding time window length and combined with a local weighted regression algorithm, the stirring speed of the stirring motor is adaptively controlled.

2. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The initial drastic factor satisfies the expression: ； In the formula, For the first The initial drastic factor at each sampling time, The total number of data points in the torque data segment at each sampling time. For the first In the torque data segment at the sampling time, the first The value of each data point. For the first The average value of all data points within the torque data segment at each sampling time.

3. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The initial intensity satisfies the expression: ； In the formula, For the first The initial intensity at each sampling moment, For the first The initial drastic factor at each sampling time, and The first The minimum and maximum values ​​of the initial severity factor at each sampling time for all data points within the torque data segment at each sampling time. To prevent the initial intensity from being a constant with a value of 0.

4. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The lag interference factor satisfies the expression: ； In the formula, For the first The lag interference factor at each sampling time. The total number of data points in the torque data segment at each sampling time. For the first The total number of data points with a positive increasing gradient among all adjacent time points within a sampling time segment of torque data. To prevent constants with a denominator of 0.

5. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The actual intensity satisfies the expression: ； In the formula, For the first The true intensity at each sampling moment For the first The initial intensity at each sampling moment, For the first The lag interference factor at each sampling time. and The first The minimum and maximum values ​​of the jamming interference factor at each sampling time for all data points within the torque data segment at each sampling time.

6. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The optimized sliding time window length satisfies the expression: ； In the formula, For the first The optimized sliding time window length at each sampling time point For the first The true intensity at each sampling moment and These are the upper and lower limits of the sliding time window length, respectively.

7. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The adaptive control of the stirring speed of the stirring motor includes: using the optimized sliding time window length, combined with a local weighted regression algorithm to perform local weighted regression fitting on the torque data, and solving for the slope of the regression curve of the torque change trend at the sampling time; predicting the torque value within a future preset time based on the slope of the regression curve at the sampling time; in response to the torque value within the future preset time at the sampling time being greater than the safe torque warning threshold, calculating a negative speed compensation amount to drive the frequency converter to reduce the motor speed to resolve the overload crisis; in response to the torque value within the future preset time at the sampling time not being greater than the safe torque warning threshold, maintaining the basic rated speed to ensure optimal mixing uniformity.

8. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The process of collecting torque data from the mixing motor includes: deploying a high-precision dynamic torque sensor on the main shaft motor of the mixer to collect torque data from the mixing motor during the mixing process.

9. The adaptive optimization control method for refractory material production parameters according to claim 1, characterized in that, The preprocessing to obtain the torque data segment for each sampling moment includes: for all torque data points, selecting several consecutive sampling moments immediately preceding them to obtain the torque data segment for each sampling moment.

10. An adaptive optimization control system for refractory material production parameters, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an adaptive optimization control method for refractory material production parameters according to any one of claims 1-9.

Images