Self-adaptive control method and system for technological parameters of preparation equipment and storage medium

Abstract

The invention relates to the technical field of beneficiation equipment control, and discloses a beneficiation equipment process parameter self-adaptive control method and system and a storage medium. The method comprises the following steps: improving the reliability of state estimation through redundant measurement data fusion, enabling a model to track process characteristic change by adopting online parameter identification, realizing prospective predictive control by utilizing rolling optimization, inhibiting process fluctuation caused by multivariable coupling through hierarchical coordination execution, and improving the reliability of state estimation. The controller parameters are continuously optimized based on gradient adjustment driven by performance indicators. The problems that in the prior art, the reliability of measured data is poor, model parameters are fixed and cannot adapt to ore property changes, control response lags behind, and the multivariable coordination control capacity is insufficient are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of beneficiation equipment control, and in particular to a beneficiation equipment process parameter adaptive control method and system and a storage medium. BACKGROUND

[0002] Beneficiation equipment process parameter control is a key technology to ensure the stable operation of the beneficiation process and improve technical and economic indicators. Traditional beneficiation process parameter control mainly relies on manual adjustment of control parameters such as air supply, stirring intensity and reagent addition amount by operators according to their experience and observation of indicators such as froth layer state and concentrate grade. This manual control method has obvious hysteresis and subjectivity. To improve control effect, existing technologies introduce automatic control systems, use single-point measurement sensors to collect flotation process state parameters, and automatically adjust process parameters according to concentrate grade deviation through PID feedback control algorithm. Some advanced systems apply machine learning methods such as LSTM-DNN neural network to establish a prediction model between process parameters and technical and economic indicators, use cloud computing platforms to process massive historical data for model training, and realize parameter optimization and regulation based on prediction results. These existing technologies reduce the dependence on manual operation to some extent and improve control accuracy.

[0003] However, the existing technologies still have significant deficiencies. First, the state information obtained by single-point measurement is easily affected by local disturbances and measurement noise. When the sensor drifts or fails, there is no effective data correction mechanism, resulting in the control system making wrong decisions based on unreliable data. Second, the control model with fixed parameters cannot adapt to the dynamic changes of ore properties. When the hardness, oxidation rate or mineral composition of the selected ore fluctuates, the pre-trained neural network model parameters remain unchanged, causing the model prediction accuracy to drop sharply and the control effect to deteriorate seriously. Third, the lag feedback control method based on results such as concentrate grade has slow response speed. From process parameter adjustment to grade detection result feedback usually takes 15 to 20 minutes. This control delay causes the flotation process to be in a non-optimal state for a long time when the ore properties suddenly change. SUMMARY

[0004] The present application provides a beneficiation equipment process parameter adaptive control method, system and storage medium, which solves the problems of poor measurement data reliability, fixed model parameters that cannot adapt to changes in ore properties, control response lag and insufficient multivariable coordination control capability in the prior art by constructing a state estimation system based on redundant measurement data fusion, a predictive control model with online parameter identification, a multivariable execution strategy with hierarchical coordination, and a controller parameter adaptive adjustment mechanism driven by performance indicators, thereby improving the accuracy, real-time performance and robustness of beneficiation equipment process parameter control.

[0005] In a first aspect, the application provides a beneficiation equipment process parameter adaptive control method, which comprises: S1 step: obtaining first rheological parameters at the inlet of the flotation tank, second rheological parameters and bubble characteristic parameters at the middle, and third rheological parameters and electrochemical parameters at the outlet; performing redundant fusion processing on the first rheological parameters, the second rheological parameters and the third rheological parameters by a weighted least squares method to obtain a fusion state estimation value; S2 step: constructing a state transition matrix and an input matrix according to the fusion state estimation value, and performing online identification on a parameter vector of the state transition matrix and the input matrix by a recursive least squares algorithm to obtain an updated parameter vector; S3 step: substituting the updated parameter vector into a rolling optimization objective function to solve a control increment sequence by quadratic programming to obtain an optimal control output; S4 step: decomposing the optimal control output into an aeration amount adjustment sequence, a stirring intensity adjustment sequence and a reagent amount adjustment sequence, and performing hierarchical execution on the aeration amount adjustment sequence, the stirring intensity adjustment sequence and the reagent amount adjustment sequence according to a time sequence priority to obtain state response data; S5 step: calculating a tracking deviation index and a control change index according to the state response data, and adjusting output weights and control weights in the rolling optimization objective function by gradient to obtain adjusted weight parameters.

[0006] In a second aspect, the application provides a beneficiation equipment process parameter adaptive control system, which comprises: A fusion module is configured to obtain first rheological parameters at the inlet of the flotation tank, second rheological parameters and bubble characteristic parameters at the middle, and third rheological parameters and electrochemical parameters at the outlet, and perform redundant fusion processing on the first rheological parameters, the second rheological parameters and the third rheological parameters by a weighted least squares method to obtain a fusion state estimation value; An identification module is configured to construct a state transition matrix and an input matrix according to the fusion state estimation value, and perform online identification on a parameter vector of the state transition matrix and the input matrix by a recursive least squares algorithm to obtain an updated parameter vector; A solving module is configured to substitute the updated parameter vector into a rolling optimization objective function to solve a control increment sequence by quadratic programming to obtain an optimal control output; A hierarchical module is configured to decompose the optimal control output into an aeration amount adjustment sequence, a stirring intensity adjustment sequence and a reagent amount adjustment sequence, and perform hierarchical execution on the aeration amount adjustment sequence, the stirring intensity adjustment sequence and the reagent amount adjustment sequence according to a time sequence priority to obtain state response data; An adjusting module is configured to calculate a tracking deviation index and a control change index according to the state response data, and to perform gradient adjustment on output weights and control weights in the rolling optimization objective function to obtain adjusted weight parameters.

[0007] In a third aspect, a beneficiation equipment process parameter adaptive control device is provided, which comprises a memory and at least one processor, and the memory stores instructions; the at least one processor invokes the instructions in the memory to enable the beneficiation equipment process parameter adaptive control device to perform the beneficiation equipment process parameter adaptive control method described above.

[0008] In a fourth aspect, a computer readable storage medium is provided, which stores instructions, and when the instructions are run on a computer, the computer is enabled to perform the beneficiation equipment process parameter adaptive control method described above.

[0009] The technical scheme provided in the application comprises the following steps: obtaining rheological parameters, bubble characteristic parameters and electrochemical parameters at three key positions of an inlet, a middle part and an outlet of a flotation tank, constructing a redundant measurement network, establishing a constraint equation group based on a material conservation relationship, performing fusion processing on multi-measurement-point data by using a weighted least square method, automatically reducing a weight coefficient of an abnormal measurement point and re-solving a corrected fusion value when the abnormal measurement point is detected, significantly improving the reliability of process state estimation, avoiding the defects that single-point measurement is easily affected by local disturbance and sensor failure, constructing a state transition matrix and an input matrix according to the fusion state estimation value, performing online identification on a parameter vector by using a recursive least square algorithm and judging a convergence state, enabling a control model to track dynamic characteristic changes of the flotation process in real time, overcoming the limitation that a fixed parameter model cannot adapt to fluctuations in ore properties, substituting the updated parameter vector into a rolling optimization objective function to perform quadratic programming solving, constructing an optimization objective based on tracking errors of output prediction value sequences in a prediction time domain and a reference trajectory and a cost item of a control increment in a control time domain, solving optimal control outputs under the constraint conditions of air supply amount, stirring intensity and reagent dosage change rate, realizing prospective predictive control, significantly shortening a response time compared with traditional lag feedback control, decomposing the optimal control outputs into three adjustment sequences and performing hierarchical execution according to time sequence priorities, monitoring state change rates to realize safe adjustment by using a three-stage step mode for the air supply amount, monitoring state coupling indexes to avoid excessive coupling by using a slope tracking mode for the stirring intensity, and starting reagent addition operation when the stirring intensity is adjusted to a preset progress, the hierarchical and coordinated execution strategy effectively suppresses process fluctuations caused by simultaneous adjustment of multiple variables through time sequence decoupling and amplitude coordination, calculates tracking deviation indexes and control change indexes according to state response data, constructs an integrated performance index to evaluate control effects, calculates gradients of output weights and control weights on the integrated performance index when the performance index is lower than a threshold, and performs numerical updating to obtain adjusted weight parameters in the gradient direction, and the performance index driven parameter adaptive adjustment mechanism enables the controller to continuously optimize its own parameters according to actual operation effects, thereby maintaining long-term stable control performance.

[0010] The innovation of the application in the field of self-adaptive control of process parameters of beneficiation equipment lies in the organic integration of multi-level self-adaptive mechanism. The weighted least squares fusion algorithm realizes the self-adaptation of the measurement layer through redundancy constraint and weight self-adaptation. The recursive least squares identification algorithm realizes the self-adaptation of the model layer through covariance update and convergence judgment. The quadratic programming optimization algorithm realizes the forward-looking optimization of the control layer through the rolling time domain and constraint processing. The gradient descent adjustment algorithm realizes the self-adaptation of the controller layer through performance feedback and parameter update. The four levels of algorithms cooperate with each other to form a complete self-adaptive control system. In particular, the gradient descent algorithm is not used for model training but for online adjustment of controller parameters in the scheme. This innovative application enables the control system to have self-improvement capability, which can automatically optimize the controller weight parameters according to the comprehensive performance index in long-term operation, adapt to the slow drift and seasonal changes of process characteristics. The forgetting factor mechanism of the recursive least squares algorithm makes the model identification pay more attention to recent data and quickly respond to sudden changes in ore properties. The learning rate mechanism of the gradient descent algorithm makes the controller parameter adjustment remain progressive, avoiding oscillation of control performance caused by parameter mutation. The quadratic programming algorithm considers the tracking performance in the prediction time domain and the stability in the control time domain when solving the optimal control output, and ensures the physical feasibility of the control output through constraint conditions. This multi-objective optimization feature makes the control strategy pursue accurate tracking while considering the stability of control action and equipment safety. The state change rate monitoring and state coupling index monitoring in the hierarchical execution strategy are essentially safety protection mechanisms based on real-time feedback. When detecting abnormal process response during adjustment, the subsequent adjustment action is immediately suspended or attenuated. This adaptive protection at the execution level further enhances the robustness of the control system. In summary, the application introduces specific adaptive algorithms in data processing, model identification, optimization control and parameter adjustment, and builds a hierarchical and closed-loop adaptive control architecture, which fundamentally solves the complex control problems of beneficiation process, such as fluctuation of ore properties, measurement noise interference, multi-variable coupling and long-term performance degradation. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0012] Figure 1 An embodiment of the self-adaptive control method of process parameters of beneficiation equipment in the embodiment of the present application is shown in the figure. Figure 2 An embodiment of the parameter identification convergence process in the embodiment of the present application is shown in the figure. Figure 3An embodiment of a beneficiation equipment process parameter adaptive control system in the present application is shown in the figure; Figure 4 An embodiment of a beneficiation equipment process parameter adaptive control system in the present application is shown in the figure; DETAILED DESCRIPTION

[0013] The present application provides a beneficiation equipment process parameter adaptive control method, system and storage medium. The terms "first", "second", "third", "fourth" and the like in the specification and claims of the present application and the above-described drawings (if any) are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0014] For ease of understanding, the specific process of the embodiments of the present application is described below. Please refer to Figure 1 An embodiment of a beneficiation equipment process parameter adaptive control method in the present application includes: S1 step: obtain the first rheological parameter at the inlet of the flotation tank, the second rheological parameter at the middle, the bubble characteristic parameter, the third rheological parameter at the outlet and the electrochemical parameter, and perform redundant fusion processing on the first rheological parameter, the second rheological parameter and the third rheological parameter by weighted least squares method to obtain a fusion state estimation value; S2 step: construct a state transition matrix and an input matrix according to the fusion state estimation value, and perform online identification on the parameter vector of the state transition matrix and the input matrix by recursive least squares algorithm to obtain an updated parameter vector; S3 step: substitute the updated parameter vector into a rolling optimization objective function to solve a quadratic programming for a control increment sequence to obtain an optimal control output; S4 step: decompose the optimal control output into an aeration amount adjustment sequence, a stirring intensity adjustment sequence and a reagent amount adjustment sequence, and perform hierarchical execution on the aeration amount adjustment sequence, the stirring intensity adjustment sequence and the reagent amount adjustment sequence according to the time sequence priority to obtain state response data; S5 step: calculate a tracking deviation index and a control change index according to the state response data, and perform gradient adjustment on the output weight and the control weight in the rolling optimization objective function to obtain an adjusted weight parameter.

[0015] It can be understood that the subject of the present application can be a beneficiation equipment process parameter self-adaptive control system, and can also be a terminal or a server, and the specific embodiments are not limited herein. The server is taken as an example for description of the embodiments of the present application.

[0016] Specifically, the process state data is obtained by deploying measuring devices at three key positions of the flotation tank. The first rheological parameter is collected at the inlet to reflect the initial viscosity state of the ore pulp, the second rheological parameter and the bubble characteristic parameter are collected at the middle part to reflect the dynamic characteristics of the flotation zone, and the third rheological parameter and the electrochemical parameter are collected at the outlet to reflect the properties of the ore pulp when the concentrate overflows. Since there is noise interference and sensor drift in the measurement, the data reliability needs to be improved through redundant fusion processing. Specifically, the weighted least squares method is adopted. Firstly, the constraint relationship among the three rheological parameters is established according to the continuity of the material in the flotation tank, a coefficient matrix is constructed to associate the rheological parameters of the three measuring points, the weight coefficients of each measuring point are set to reflect the measurement confidence, the preliminary fusion value is obtained by minimizing the weighted residual sum of squares, then the residual vector between the actual measurement value and the fusion value is calculated, the standardized processing of the residual eliminates the dimension influence, when the absolute value of the standardized residual of a measuring point exceeds 3 times the standard deviation, it is determined as an abnormal measuring point, the weight of the measuring point is reduced to one fifth of the original value, and the corrected fusion value is obtained by re-executing the weighted least squares solution, finally, the corrected rheological fusion value, the bubble characteristic parameter and the electrochemical parameter are combined in columns to form a state vector as the fusion state estimation value, which integrates the information of multiple measuring points and eliminates abnormal interference.

[0017] Based on the fusion state estimation value, the mathematical description of the process is established, and the flotation process is abstracted into a state space form. The state transition matrix describes how the current state evolves to the next state, and the input matrix describes the influence degree of the control input on the state change. The elements of these two matrices constitute the parameter vector. Since the flotation process characteristics will change with the properties of the ore, the parameter vector needs to be identified online. The recursive least squares algorithm is adopted to update the parameter estimation by continuously collecting input and output data. The specific process is as follows: at each sampling time, the control input sequence and the actual output sequence are collected, the historical state value and the input value are combined into a regression vector, and a covariance matrix is maintained to reflect the uncertainty of the parameter estimation. The regression vector and the covariance matrix are substituted into the recursive formula to calculate the identified parameter vector, which is the current estimation of the true parameter. To determine whether the parameter converges, the Euclidean distance between the current identified parameter vector and the parameter vector at the previous time is divided by the modulus of the parameter vector to obtain the parameter change rate. When the change rate is less than 2% for 20 consecutive sampling periods, it is determined that the parameter has converged to a stable value. At this time, the identified parameter vector is output as the updated parameter vector for subsequent control, and the covariance matrix is updated according to the recursive formula for the next time identification. The whole identification process makes the process model parameters in the controller continuously approach the true system characteristics.

[0018] The optimal control strategy is calculated with the updated parameter vector. Firstly, the parameter vector is split by dimension to reconstruct the state transition matrix and the input matrix. Based on these two matrices, a prediction mechanism is established to predict the output value at each time in the future prediction horizon from the current time. The prediction is calculated by recursion. The state prediction value at each step is equal to the state transition matrix multiplied by the state of the previous step plus the input matrix multiplied by the control input at the corresponding time. The output prediction value sequence is obtained by mapping the state to the output. Then, the sequence is subtracted point by point from the expected reference trajectory to obtain the tracking error at each time. The tracking error term is formed by squaring all the errors, multiplying by the output weight and summing. The larger the value of this term, the worse the tracking performance. At the same time, the control increment at each time in the control horizon is squared, multiplied by the control weight and summed to obtain the control cost term. This term reflects the degree of control action. The two terms are added to form the rolling optimization objective function. To ensure the physical feasibility of the control output, the change rate of the air volume is set to not more than 0.15 cubic meters per minute, the change rate of the stirring intensity is set to not more than 15 revolutions per minute, and the change rate of the medicament amount is set to not more than 10 grams per ton as the constraint condition. These constraints are converted into inequality form. The quadratic programming is solved for the rolling optimization objective function with constraints to obtain the optimal control increment sequence. This sequence contains the control increment at each time in the control horizon, but only the first element in the sequence is executed. The element is added to the control amount at the previous time to obtain the optimal control output at the current time.

[0019] The optimal control output is separated into three control channels, the air volume control command is used to adjust the air supply system at the bottom of the flotation tank, the stirring intensity control command is used to adjust the stirring motor speed, and the reagent dosage control command is used to adjust the flow of the reagent pump. Each command is subtracted by the corresponding control amount at the previous moment to obtain the actual adjustment amplitude, forming an air volume adjustment sequence, a stirring intensity adjustment sequence, and a reagent dosage adjustment sequence. Due to the different response speeds and coupling effects of the three control channels, a hierarchical execution strategy is adopted. First, the air volume adjustment is executed, which is fast in response. The air volume adjustment sequence is divided into three step amounts in the proportions of 40%, 35%, and 25%. The first segment is executed at the current moment, the second segment is executed after an interval of 1 second, and the third segment is executed after another interval of 1 second. The fusion state estimate value is collected immediately after each segment is executed. The difference between the adjacent two state values divided by the time interval is obtained to get the state change rate. When the rate exceeds the threshold value of 0.08 per second, the remaining steps are suspended to avoid process fluctuations. After the air volume adjustment is completed, the stirring intensity adjustment is started. Instead of using the step method, the adjustment amount is divided equally into 6 time steps. The sixth of the adjustment amount is increased every second to achieve ramp tracking. During the adjustment process, the product of the state sensitivity to the air volume and the air volume change amplitude is calculated in real time, and the product of the state sensitivity to the stirring intensity and the stirring intensity change amplitude is added to obtain the state coupling index. When the index exceeds the upper limit of 0.15, the subsequent adjustment amplitude is attenuated to 70% of the original to prevent overcoupling. When the stirring intensity adjustment reaches 50% progress, the reagent dosage adjustment is started simultaneously. According to the adjustment sequence and the feed flow, the volume of reagent to be added is calculated. The reagent pump is controlled to complete the addition in 180 seconds at a constant flow. The control output value and the corresponding fusion state estimate value at each moment are recorded to form time series data as state response data.

[0020] The control performance is evaluated according to the state response data, and the controller parameters are adjusted, the output variable time sequence is extracted from the state response data, the sequence reflects the changes of the quality index and the recovery rate index over time, the output sequence is time-aligned with the set reference trajectory, the difference between the output value and the reference value is calculated at each time point, the sum of the squares of all the differences is divided by the number of time points to obtain the average tracking error, which is used as a tracking deviation index to quantify the set value tracking performance, the execution trajectories of the three control channels are extracted from the state response data, the difference between the control amounts at adjacent time points is calculated for each channel, the sum of the squares of the differences at each time point of the three channels is accumulated and divided by the total number of time points to obtain the average control change rate, which is used as a control change index to reflect the smoothness of the control action, then the tracking deviation index is divided by the target tracking deviation to obtain the tracking performance ratio, the control change index is divided by the target control change to obtain the control stability ratio, the squares of the two ratios are multiplied by the weight coefficients and added to obtain a comprehensive performance index, the index is compared with the performance threshold 0.90, when the index is less than the threshold, it indicates that the control performance is not up to standard and the controller parameters need to be adjusted, the relative size relationship between the tracking deviation index and the control change index is analyzed to determine the adjustment direction, if the tracking deviation is more than 1.2 times the target value and the control change is less than the target value, it indicates that the tracking is insufficient and the output weight should be increased and the control weight should be reduced, if the tracking deviation is less than the target value but the control change exceeds 1.3 times the target value, it indicates that the control is too active and the output weight should be reduced and the control weight should be increased, the output weight and the control weight in the rolling optimization objective function are expanded into a weight parameter vector, the partial derivatives of the comprehensive performance index with respect to each element in the vector are calculated to form a gradient vector, the gradient direction indicates the parameter change direction that makes the performance index increase the fastest, the gradient vector is multiplied by the learning rate 0.05 to obtain a weight adjustment amount vector, the adjustment amount vector is added to the current output weight and the control weight to complete parameter update to obtain the adjusted weight parameters, and the updated weight parameters will be used in the rolling optimization of the next control period, so that the controller gradually adapts to the changes of process characteristics and performance requirements.

[0021] In a specific embodiment, the S1 step comprises: obtaining a first rheological parameter at the inlet of the flotation tank, a second rheological parameter at the middle, and a third rheological parameter at the outlet, constructing a constraint equation set based on the material conservation relationship, substituting the first rheological parameter, the second rheological parameter, and the third rheological parameter into the constraint equation set for weighted least squares solution to obtain a preliminary fusion value; calculating a measurement residual vector of the constraint equation set, and performing standardization processing on the measurement residual vector to obtain a standardized residual of each measuring point; determining an abnormal measuring point according to the standardized residual, and performing weighted least squares solution again after reducing the weight coefficient corresponding to the abnormal measuring point to obtain a corrected fusion value; combining the corrected fusion value with bubble characteristic parameters and electrochemical parameters to obtain a fusion state estimation value.

[0022] Specifically, the first rheological parameter at the inlet of the flotation tank is measured by a rotary viscosity sensor, the second rheological parameter at the middle of the flotation tank is measured by a same type of sensor installed at the middle of the tank, and the third rheological parameter at the outlet of the flotation tank is measured by a sensor at the overflow weir. The three rheological parameters are physically related because the pulp flows continuously in the flotation tank and follows the law of conservation of mass. A set of constraint equations is established based on the conservation of mass. The principle of establishing the constraint equations is that although the viscosity of the pulp changes due to the addition of bubbles and the separation of minerals during the flow from the inlet to the middle and then to the outlet, the changes follow a certain physical law. Specifically, the first constraint equation describes the relationship between the inlet and the middle. The first rheological parameter multiplied by the first transfer coefficient should be equal to the theoretical value of the second rheological parameter. The first transfer coefficient reflects the proportion of viscosity change due to the addition of bubbles during the flow from the inlet to the middle. The second constraint equation describes the relationship between the middle and the outlet. The second rheological parameter multiplied by the second transfer coefficient should be equal to the theoretical value of the third rheological parameter. The second transfer coefficient reflects the proportion of viscosity change due to the separation of concentrate during the flow from the middle to the outlet. The two transfer coefficients are obtained by analyzing a large amount of measurement data during the normal operation period. The first transfer coefficient is 0.92, indicating that the viscosity of the pulp generally decreases to 92% of the original value after flowing through the flotation reaction zone. The second transfer coefficient is 0.88, indicating that the viscosity of the concentrate overflow further decreases to 88% of the previous value. The constraint equation set links the three independently measured rheological parameters through the physical constraint of mass conservation to form a mutually restraining mathematical relationship. The actual measured values of the first, second, and third rheological parameters are substituted into the constraint equation set. Due to factors such as random noise in the measurement process, sensor drift, and local disturbance of the pulp, the actual measured values often cannot accurately satisfy the theoretical relationship of the constraint equation set. The weighted least squares method is used to solve a set of fusion values that are as close as possible to the measured values of each measuring point and as close as possible to the physical constraint relationship of the constraint equation set. The weights reflect the differences in measurement reliability of different measuring points. The inlet measuring point is located in the relatively stable flow of the feed pipe, so the weight is set to 0.25. The middle measuring point is located in the key flotation reaction zone, so the weight is set to 0.30. The outlet measuring point is located at the overflow weir, which is greatly affected by the foam layer fluctuations, so the weight is set to 0.20. The weighted least squares objective function is the square of the difference between the measured value and the fusion value of the first rheological parameter multiplied by the weight 0.25, plus the square of the difference between the measured value and the fusion value of the second rheological parameter multiplied by the weight 0.30, plus the square of the difference between the measured value and the fusion value of the third rheological parameter multiplied by the weight 0.20. At the same time, the fusion value is required to satisfy the constraint equation set or to minimize the violation of the constraint equation set. The partial derivative of the weighted objective function with respect to the fusion value is taken and set to zero to obtain the optimal fusion value that minimizes the objective function, which is the preliminary fusion value. The preliminary fusion value integrates the information of the three measuring points and considers the constraint of mass conservation.

[0023] The measurement residual vector of the constraint equation set is a key step to evaluate the measurement deviation of each measuring point, and the calculation method of the measurement residual vector is to substitute the preliminary fusion value back into the constraint equation set to check the satisfaction degree of the constraint equation set. For the first constraint equation, the preliminary fusion value is multiplied by the first transfer coefficient 0.92 to obtain the theoretical prediction value of the middle measuring point, and the actual measurement value of the second rheological parameter is subtracted from the theoretical prediction value to obtain the first residual component, which reflects the deviation between the actual measurement value of the middle measuring point and the theoretical value calculated based on the inlet measuring point and the material conservation. For the second constraint equation, the preliminary fusion value is multiplied by the first transfer coefficient 0.92 and then multiplied by the second transfer coefficient 0.88 to obtain the theoretical prediction value of the outlet measuring point, and the actual measurement value of the third rheological parameter is subtracted from the theoretical prediction value to obtain the second residual component, which reflects the deviation between the actual measurement value of the outlet measuring point and the theoretical value calculated based on the inlet measuring point and the two-stage material conservation. For the inlet measuring point itself, the first rheological parameter actual measurement value is subtracted from the preliminary fusion value to obtain the third residual component, which reflects the direct deviation between the actual measurement value of the inlet measuring point and the fusion value. The three residual components constitute the measurement residual vector, and each element of the residual vector corresponds to the measurement deviation degree of a measuring point. The greater the absolute value of the residual, the worse the consistency of the measurement value of the measuring point with the overall data. The residual vector contains the distribution information of the measurement errors of the three measuring points.

[0024] The standardization processing of the measurement residual vector is to convert the residual with dimension and order of magnitude of different measuring points into dimensionless standardized residual for the convenience of setting a unified abnormal judgment threshold. The specific method of standardization processing is to first calculate the statistical characteristics of the residual vector, including the residual mean and the residual standard deviation. The residual mean is calculated by adding the three elements of the residual vector and dividing by 3 to reflect the central tendency of the residual. The residual standard deviation is calculated by calculating the square sum of the difference between each residual and the mean value and then dividing by 3 to obtain the dispersion degree of the residual. Then, a standardization transformation is performed on each element of the residual vector. The standardization transformation formula is to subtract the residual mean from the residual value of the element and then divide by the residual standard deviation. The standardized residual eliminates the dimension and has a mean of zero and a standard deviation of one, which follows a standard normal distribution. The standardized residual of each measuring point corresponds to the deviation degree of the inlet measuring point, the middle measuring point and the outlet measuring point under the unified standard, respectively. The greater the absolute value of the standardized residual, the farther the measuring point deviates from the normal distribution.

[0025] According to the standardized residual, the 3-sigma criterion is used as the statistical test standard to determine abnormal measurement points. The 3-sigma criterion is based on the normal distribution theory, which considers that the probability of normal data falling within the range of 3 times the standard deviation of the mean is 99.73%. When the absolute value of the standardized residual of a measurement point is greater than 3, the data of the measurement point falls in the extreme area of the normal distribution tail, and it is determined to be an abnormal measurement point. The causes of abnormal measurement points include sensor surface fouling leading to measurement lag, electrical interference introducing pulse noise, and local pulp property mutation causing measurement values to be unrepresentative. The weight coefficient corresponding to the abnormal measurement point is adjusted from the original value to one-fifth of the original value. The reduction in weight means that the influence of the abnormal measurement point on the fusion result is greatly weakened when the weighted least squares is solved again. However, the abnormal measurement point data is still retained in the calculation rather than being completely excluded because the measurement value may be biased but still contains some real process information. The weighted least squares objective function is reconstructed using the adjusted weight matrix. The weight coefficient of the abnormal measurement point in the objective function has been updated to a reduced value. The normal measurement point weight remains unchanged. The partial derivative of the updated objective function with respect to the fusion value is taken again and set to zero to solve the corrected fusion value considering abnormal suppression. The corrected fusion value is less disturbed by abnormal measurement points than the preliminary fusion value and more accurately reflects the true rheological state of the pulp in the flotation tank. The abnormality detection and weight adjustment mechanism enables the redundant fusion algorithm to have self-adaptive anti-interference capability.

[0026] The corrected fusion value, bubble characteristic parameters, and electrochemical parameters are combined to form a complete process state description. The bubble characteristic parameters are the bubble Sauter mean diameter, which is measured by an ultrasonic detection device to reflect the bubble size distribution in the flotation process. The Sauter mean diameter is defined as 6 times the ratio of the volume to the surface area of the bubble group, representing the statistical average characteristic of the bubble size. The electrochemical parameter is the pulp conductivity, which is measured by an electromagnetic probe to reflect the ion concentration and chemical environment state in the pulp. High conductivity indicates high soluble ion concentration, which affects reagent adsorption and mineral surface electrical properties. The three types of parameters describe the flotation process characteristics from different physical dimensions. The corrected fusion value describes the rheological properties of the pulp, reflecting the pulp viscosity resistance and flow characteristics. The bubble characteristic parameters describe the gas-liquid interface characteristics, reflecting the geometric conditions of bubble and mineral particle collision and adhesion. The electrochemical parameter describes the chemical environment state, reflecting the chemical background of reagent action. The corrected fusion value is taken as the first element of the state vector, the bubble characteristic parameters are taken as the second element of the state vector, and the electrochemical parameter is taken as the third element of the state vector. The three elements are arranged in column vector form to obtain a fusion state estimation value vector containing three numerical elements. This vector comprehensively represents the comprehensive running state of the flotation process in the rheological characteristics, bubble morphology, and chemical environment dimensions at the current time. The fusion state estimation value vector serves as the basic data source for the state feedback signal of the control system.

[0027] In a specific embodiment, the S2 step includes: According to the fusion state estimation value, a state space expression is constructed, and a state transition matrix and an input matrix in the state space expression are unfolded into a parameter vector; A control input sequence and an actual output sequence are collected, a regression vector and a covariance matrix are constructed, the regression vector and the covariance matrix are substituted into a recursive least square formula for parameter identification, and an identification parameter vector is obtained; A parameter change rate of the identification parameter vector and a previous time parameter vector is calculated, a parameter convergence state is determined according to the parameter change rate, and a convergence determination result is obtained; When the convergence determination result meets a convergence condition, the identification parameter vector is output as an updated parameter vector, and the covariance matrix is updated.

[0028] Specifically, the state space expression includes two parts of a state equation and an output equation, the state equation describes how a next time state vector is jointly determined by a current time state vector and a control input vector, and the output equation describes how an output vector is mapped to be obtained from the state vector and the control input vector, a general form of the state equation is that the next time state vector is equal to a state transition matrix multiplied by the current time state vector plus an input matrix multiplied by the current time control input vector, the state transition matrix describes a dynamic evolution law of the process itself and reflects a mutual influence relationship between state variables, the input matrix describes an action strength of the control input on state change and reflects how the air volume, the stirring intensity and the reagent amount change the flotation process state, a form of the output equation is that the output vector is equal to an output matrix multiplied by the state vector plus a direct transfer matrix multiplied by the control input vector, the output vector includes a concentrate quality index and a recovery rate index, the fusion state estimation value is substituted into the state space expression as a real-time value of the state vector, and unfolding the state transition matrix and the input matrix in the state space expression into the parameter vector means arranging all elements of the matrix in a long vector in a row or in a column, the state transition matrix is 5 rows and 5 columns and includes 25 elements, the input matrix is 5 rows and 3 columns and includes 15 elements, the output matrix is 2 rows and 5 columns and includes 10 elements, and the direct transfer matrix is 2 rows and 3 columns and includes 6 elements, all elements of the four matrices are spliced to form a parameter vector including 56 elements, each element of the parameter vector corresponds to a to-be-identified parameter of the state space model, and these parameters reflect dynamic characteristics of the flotation process and change with the ore property and need to be identified online.

[0029] The collection control input sequence and actual output sequence are the data basis for parameter identification, the control input sequence contains the historical record values of the aeration amount, stirring intensity and medicament amount at continuous multiple sampling time points, the actual output sequence contains the measured values of the concentrate quality index and recovery rate index at the corresponding time points, the data of the last 10 sampling time points are selected to form an identification data window, the regression vector is constructed by combining the historical state values and historical input values in a specific manner, the construction of the regression vector is based on the state equation to concatenate the state vector at the previous time point and the control input vector at the previous time point by column, the dimension of the regression vector is equal to the dimension of the state vector plus the dimension of the control input vector, that is, 5 plus 3 is equal to 8, the covariance matrix describes the uncertainty of parameter estimation and the correlation between parameters, the covariance matrix is a 56*56 symmetric positive definite matrix initialized as a unit matrix multiplied by a larger constant 100, indicating that the initial parameters are completely unknown, the regression vector and the covariance matrix are substituted into the recursive least squares formula to perform parameter identification, the recursive least squares formula includes three sub-steps of gain calculation, parameter update and covariance update, the gain calculation is the covariance matrix multiplied by the regression vector divided by the forgetting factor plus the regression vector transposed multiplied by the covariance matrix multiplied by the regression vector, the forgetting factor is 0.98, indicating that the old data is applied to the decay weight to make the identification pay more attention to the recent data, the parameter update is the previous time point parameter vector plus the gain multiplied by the prediction error, the prediction error is equal to the actual measured output minus the predicted output calculated based on the previous time point parameter and the current input, the covariance update is the previous time point covariance matrix minus the gain multiplied by the regression vector transposed multiplied by the covariance matrix and then divided by the forgetting factor, the recursive least squares algorithm continuously corrects the parameter estimation by rolling calculation at each time point to make the parameter vector approach the true value, and the identification parameter vector is obtained as the latest estimation of the process model parameters at the current time point, and the identification parameter vector is continuously updated to track the process characteristic changes with the addition of new data.

[0030] The parameter change rate of the identification parameter vector and the parameter vector of the previous moment is calculated to determine whether the parameter identification converges to a stable value. The parameter change rate calculation method is to subtract the parameter vector of the previous moment from the current identification parameter vector to obtain a parameter increment vector, to calculate the Euclidean norm of the parameter increment vector to obtain the total amplitude of the parameter change, and to normalize the Euclidean norm of the parameter vector of the previous moment to obtain the relative change rate. The Euclidean norm is the square root of the sum of squares of all elements of the vector, reflecting the overall size of the vector. The parameter change rate is a dimensionless quantity reflecting the degree of parameter update. According to the parameter change rate, when the parameter change rate is less than a set threshold value 0.02, that is, the relative change of the parameter is less than two percent, it is considered that the parameter identification enters the convergence state. However, a single condition is not enough to determine convergence. It needs to be continuously met for multiple sampling periods to confirm. The number of continuous meeting threshold is set to 20. Whether the parameter change rate is less than the threshold value at each moment is recorded to form a judgment sequence. Whether all the judgment sequences of the last 20 moments are true is checked. If all are true, the convergence determination result is that it has converged. If any one is false, the convergence determination result is that it has not converged. The convergence determination mechanism prevents temporary parameter fluctuations from being misjudged as converged to ensure the stability and reliability of the identified parameters.

[0031] When the convergence determination result meets the convergence condition, that is, it is determined to have converged, the parameter output operation is performed. The identification parameter vector is output as an updated parameter vector for subsequent use of the controller. The updated parameter vector is an estimated value of the parameter that has been sufficiently identified and converged, accurately reflecting the true dynamic characteristics of the current flotation process. The updated parameter vector is reconstructed into the form of a state transition matrix and an input matrix according to the inverse operation of the parameter vector expansion. The first 25 elements are arranged in 5 rows and 5 columns to restore the state transition matrix. The 26th to 40th elements are arranged in 5 rows and 3 columns to restore the input matrix. The reconstructed matrix is used for subsequent predictive control calculation. The covariance matrix is updated for the next identification. The covariance matrix update is performed according to the covariance update formula of the recursive least squares. The updated covariance matrix reflects the uncertainty of the current parameter estimation. The smaller the diagonal elements of the covariance matrix, the more accurate the corresponding parameter estimation. The parameter identification process continues to perform recursive calculation at each sampling moment. When the properties of the ore change, causing the process characteristics to change, the parameter vector deviates from the convergence state and the change rate exceeds the threshold value. The identification algorithm automatically detects the parameter change and restarts the convergence determination process. Through online parameter identification, the internal model of the controller always tracks the actual process characteristics, solving the problem that the fixed parameter model in the prior art cannot adapt to fluctuations in ore properties.

[0032] Referring to Figure 2 , Figure 2 is a schematic diagram of the parameter identification convergence process in the embodiments of the present application. Figure 2The parameter change rate in the recursive least square algorithm online parameter identification process is shown to change with the sampling period. The horizontal axis is the sampling period, and the vertical axis is the parameter change rate (%). The dotted line represents the 2% convergence threshold. As can be seen from the figure, the parameter change rate rapidly decreases from about 15% at the initial stage, and the parameter change rate decreases to below the threshold and remains stable within 30-70 cycles. When the parameter change rate is less than 2% for 20 consecutive cycles, the system determines that the parameter has converged. Near the 70th cycle, the parameter change rate rises to about 6% due to changes in ore properties. The identification algorithm detects changes in process characteristics and restarts the convergence determination. The parameter change rate rapidly decreases again and tends to be stable within 85-100 cycles, verifying that the recursive least square algorithm has the ability to adaptively track changes in process characteristics. This mechanism enables the controller's internal model to track changes in the dynamic characteristics of the flotation process in real time, solving the problem of fixed parameter models that cannot adapt to fluctuations in ore properties.

[0033] In a specific embodiment, the S3 step includes: The updated parameter vector is decomposed by dimension into a state transition matrix and an input matrix. The output prediction value sequence at each time in the prediction horizon is calculated based on the state transition matrix and the input matrix. The output prediction value sequence is point-by-point difference calculated with the reference trajectory at the corresponding time. The difference at each time is weighted and squared to obtain the tracking error term. The control increments at each time in the control horizon are weighted and squared to obtain the control cost term. The tracking error term and the control cost term are added to form the rolling optimization objective function. The upper limit of the air volume change rate, the upper limit of the stirring intensity change rate, and the upper limit of the reagent dosage change rate are set as constraint conditions. The rolling optimization objective function is solved by quadratic programming to obtain the control increment sequence. The first element of the control increment sequence is added to the control amount at the previous time to obtain the optimal control output.

[0034] Specifically, decomposing the update parameter vector by dimension into a state transition matrix and an input matrix means recovering from a long vector form to a matrix form, the update parameter vector contains 56 elements, the first 25 elements are filled in the form of 5 rows and 5 columns to reconstruct the state transition matrix, the 26th to 40th elements are filled in the form of 5 rows and 3 columns to reconstruct the input matrix, the state transition matrix describes the evolution rule of the state vector from the current time to the next time, and the input matrix describes the driving effect of the control input vector on the state change, the output prediction value sequence at each time in the prediction time domain is calculated according to the state transition matrix and the input matrix, the prediction time domain is set to 10 sampling periods to represent 10 steps of prediction to the future, the prediction calculation adopts a recursive manner, the first step prediction state is equal to the state transition matrix multiplied by the current state vector plus the input matrix multiplied by the first step control input assumed, the second step prediction state is equal to the state transition matrix multiplied by the first step prediction state plus the input matrix multiplied by the second step control input assumed, and the recursive calculation is sequentially performed to the tenth step to obtain 10 prediction state vectors, the prediction state vectors are multiplied by the output matrix to obtain the corresponding output prediction values, the output prediction values include concentrate quality index prediction values and recovery rate index prediction values, and the output prediction values at 10 times are arranged in time sequence to form an output prediction value sequence.

[0035] Point-by-point difference calculation of the output prediction value sequence and the reference trajectory at the corresponding time is the basis for evaluating the tracking performance, the reference trajectory is the expected output change path containing the concentrate quality target value and the recovery rate target value at each time, and the reference trajectory is determined by the production plan and reflects the process index requirements, the difference between the output prediction value and the reference trajectory is calculated at each time in the prediction time domain, the first time difference is the first time output prediction value minus the first time reference value, the second time difference is the second time output prediction value minus the second time reference value, and 10 difference vectors are sequentially calculated, each difference vector contains quality deviation and recovery rate deviation, the weighted square sum of the difference at each time is calculated, the weight reflects the importance difference of different times and different output variables, the quality index weight is set to 100 because it is directly related to product quality, the recovery rate index weight is set to 80 because it affects economic benefits, and the weight of the long-term prediction time is slightly lower than that of the recent time, reflecting that the prediction uncertainty increases with time, and the specific calculation is that the square of the first time quality deviation is multiplied by the weight 100, and the square of the first time recovery rate deviation is multiplied by the weight 80 to obtain the first time weighted error, and the tracking error term is obtained by accumulating the weighted errors of 10 times, and the larger the tracking error term value is, the more serious the deviation between the predicted output and the expected trajectory is.

[0036] The control increments at each time in the control time domain are weighted and squared to obtain a control cost term, the control time domain is set to 3 sampling periods, indicating that only the control actions of the first 3 steps are optimized, the control time domain is shorter than the prediction time domain because the long-term control action is uncertain and will not be executed immediately, the control increment is the difference between the control amounts at adjacent times, reflecting the change amplitude of the control action, the first-time control increment is the first-time control output minus the current control amount, the second-time control increment is the second-time control output minus the first-time control output, and the third-time control increment is the third-time control output minus the second-time control output, each control increment includes three components of the aeration amount increment, the stirring intensity increment and the medicament amount increment, the control increments at the 3 control time domain times are weighted and squared, the aeration amount increment weight is set to 1.5 because the aeration system responds quickly and needs to be limited to change, the stirring intensity increment weight is set to 2.0 because the stirring motor has large power and changes too fast to affect the service life of the equipment, and the medicament amount increment weight is set to 1.0 because the medicament adjustment is relatively moderate, the first-time aeration amount increment is squared, multiplied by 1.5, and added to the stirring intensity increment, which is squared and multiplied by 2.0, and then added to the medicament amount increment, which is squared and multiplied by 1.0, to obtain the first-time control cost, the control costs at the 3 times are accumulated to obtain the control cost term, the larger the control cost term value, the more intense the control action, the tracking error term and the control cost term are added to form a rolling optimization objective function, the objective function considers both tracking performance and control stability requirements, minimizing the objective function means finding the optimal control strategy that can both track the reference trajectory and keep the control action stable.

[0037] The upper limit of the aeration amount change rate, the upper limit of the stirring intensity change rate and the upper limit of the medicament amount change rate are set as constraint conditions to ensure the physical feasibility of the control output, the upper limit of the aeration amount change rate is set to 0.15 cubic meters per minute, indicating that the aeration amount adjustment amplitude per unit time does not exceed this value to prevent the bubble layer from being unstable, the upper limit of the stirring intensity change rate is set to 15 revolutions per minute, indicating that the speed adjustment rate is limited by the motor acceleration and deceleration capability, and the upper limit of the medicament amount change rate is set to 10 grams per ton, indicating that the flow regulation range of the dosing pump is constrained. The three change rate upper limits are converted into inequality constraints on the control increment, the absolute value of the aeration amount increment is less than or equal to 0.15 cubic meters per minute, the absolute value of the stirring intensity increment is less than or equal to 15 revolutions per minute, and the absolute value of the medicament amount increment is less than or equal to 10 grams per ton. There are 9 control increment components in the control time domain, and each component corresponds to a pair of upper and lower bound constraints. The rolling optimization objective function is solved by quadratic programming. Quadratic programming is used to solve optimization problems with quadratic objective functions and linear constraints. The tracking error term and the control cost term in the objective function are both quadratic functions of the control increment. The constraint condition is a linear inequality of the control increment. The quadratic programming solver iteratively calculates the optimal control increment sequence that minimizes the objective function and satisfies all constraints through the efficient set method or the interior point method. The control increment sequence includes 3 time points and 9 control increment components. The first element of the control increment sequence is taken, which includes the aeration amount increment, the stirring intensity increment and the medicament amount increment at the first time point. These three increments are added to the corresponding control amounts at the previous time point. The optimal aeration amount output at the current time point is obtained by adding the aeration amount increment to the aeration amount at the previous time point. The optimal stirring intensity output at the current time point is obtained by adding the stirring intensity increment to the stirring intensity at the previous time point. The optimal medicament amount output at the current time point is obtained by adding the medicament amount increment to the medicament amount at the previous time point. The three output values constitute the optimal control output vector. Rolling optimization is performed once every sampling period. Only the first control increment is implemented each time. The next period is based on the new state measurement to solve the optimization problem again. The rolling optimization mechanism enables the controller to continuously adjust the control strategy based on process state feedback to achieve closed-loop optimal control.

[0038] In a specific embodiment, the S4 step includes: The optimal control output is decomposed into aeration amount control instructions, stirring intensity control instructions and medicament amount control instructions according to the control channel. The difference between each control instruction and the control amount at the previous time point is calculated to obtain an aeration amount adjustment sequence, a stirring intensity adjustment sequence and a medicament amount adjustment sequence. The aeration amount adjustment sequence is divided into three step amounts, and each step amount is executed in turn according to the set time interval. The fusion state estimation value is collected after each step is executed, and the state change rate is calculated. When the state change rate exceeds the threshold value, the subsequent step execution is paused. After the completion of the aeration amount adjustment sequence, the stirring intensity adjustment sequence is converted into a ramp tracking instruction, the stirring intensity is gradually adjusted at a set slope, and the state coupling index is monitored synchronously, and when the state coupling index exceeds the upper limit, the adjustment amplitude is attenuated. When the stirring intensity adjustment sequence is executed to a preset progress, the filling operation of the medicament amount adjustment sequence is started, the execution trajectory of each control channel and the fusion state estimation value at the corresponding time are recorded, and the state response data is obtained.

[0039] Specifically, the decomposition of the optimal control output according to the control channel refers to splitting the control output vector containing three control variables into three independent control instructions. The optimal control output vector contains three elements of aeration amount, stirring intensity and medicament amount. The first element is extracted as the aeration amount control instruction, the second element is extracted as the stirring intensity control instruction, and the third element is extracted as the medicament amount control instruction. The difference between each control instruction and the control amount at the previous time is calculated. The aeration amount control instruction minus the aeration amount at the previous time to obtain the aeration amount adjustment amplitude. The stirring intensity control instruction minus the stirring intensity at the previous time to obtain the stirring intensity adjustment amplitude. The medicament amount control instruction minus the medicament amount at the previous time to obtain the medicament amount adjustment amplitude. The three adjustment amplitudes respectively constitute single-element aeration amount adjustment sequence, stirring intensity adjustment sequence and medicament amount adjustment sequence. The adjustment sequence value is positive, indicating that the control amount is increased. The value is negative, indicating that the control amount is decreased.

[0040] The three-step amount is used to avoid the instability of the bubble layer caused by the sudden change of the inflation amount. The decomposition ratio is set to 40%, 35%, and 25%. The first-step amount is the inflation amount adjustment sequence value multiplied by 0.4. The second-step amount is the inflation amount adjustment sequence value multiplied by 0.35. The third-step amount is the inflation amount adjustment sequence value multiplied by 0.25. The sum of the three-step amounts is equal to the original adjustment sequence value. Each step amount is executed in turn according to the set time interval. The time interval is set to 1 second. The first-step amount is executed at the current time to increase the inflation amount from the previous time value by the first-step amount. After an interval of 1 second, the second-step amount is executed to continue to increase the first-step amount by the second-step amount. After another interval of 1 second, the third-step amount is executed to complete the entire adjustment. The fusion state estimation value is collected after each step is executed. The fusion state estimation value collected after the first step is executed is recorded as the first collection value. The fusion state estimation value collected after the second step is executed is recorded as the second collection value. The state change rate is calculated to evaluate the dynamic response of the process. The state change rate is the current collection fusion state estimation value minus the previous collection fusion state estimation value divided by the time interval of 1 second. The state change rate after the first step is executed is the first collection value minus the state value before execution divided by 1 second. When the state change rate exceeds the threshold value of 0.08 per second, it is determined that the process response is too fast and there is a risk of instability. The subsequent step execution is suspended, which means that the second-step and third-step amounts are no longer executed and the current inflation amount remains unchanged. The suspension mechanism prevents the inflation amount adjustment from causing the collapse of the foam layer or the flooding of the tank.

[0041] After the completion of the aeration amount adjustment sequence, it is determined whether all three-step jumps have been completed or if the state change rate has exceeded the limit and the adjustment is suspended. The end time of the aeration amount adjustment is taken as the start time of the stirring intensity adjustment, and the stirring intensity adjustment sequence is converted into a slope tracking instruction. The slope tracking instruction refers to a control quantity that changes linearly at a constant rate rather than a step mutation. The slope is set to the stirring intensity adjustment sequence value divided by the slope time of 6 seconds. During the slope time, the stirring intensity increases by one sixth of the adjustment amplitude per second. The stirring intensity is gradually adjusted according to the set slope. At the completion time of the aeration amount adjustment, the stirring intensity remains the value at the previous time. After 1 second, the stirring intensity increases by one slope unit. After 2 seconds, the stirring intensity increases by two slope units. After 6 seconds, the stirring intensity accumulatively increases by six slope units to reach the target value and complete the adjustment. The state coupling index is monitored synchronously. The state coupling index describes the combined influence of the aeration amount change and the stirring intensity change on the process state. The state coupling index is calculated as the state sensitivity to the aeration amount multiplied by the aeration amount change amplitude plus the state sensitivity to the stirring intensity multiplied by the stirring intensity change amplitude. The sensitivity is extracted from the identified input matrix to reflect the state change caused by a unit change in the control quantity. The state coupling index is calculated every second during the stirring intensity slope adjustment. When the state coupling index exceeds the upper limit of 0.15, it is determined that the combined effect of the aeration amount and the stirring intensity is too strong, and there is a risk of coupled oscillation. The adjustment amplitude is attenuated. The attenuation operation is to reduce the stirring intensity increment at the subsequent time to 70% of the original value. The slope is adjusted to 0.7 times the original slope. After attenuation, the stirring intensity continues to adjust at a slower rate until the adjustment is completed or the attenuation is triggered again.

[0042] The medicament dosage adjustment sequence is started when the stirring intensity adjustment sequence is executed to a preset progress, the preset progress is set to 50% to indicate that the medicament addition is started when the stirring intensity adjustment is performed to half, the total length of the stirring intensity adjustment is 6 seconds, and the progress of 50% is reached when the stirring intensity adjustment is performed to 3 seconds, at which time the filling operation of the medicament dosage adjustment sequence is started, the medicament dosage adjustment sequence is a single numerical value indicating the total amount of medicament that needs to be increased or decreased, the filling operation is realized by controlling the medicament pump, the medicament volume to be filled is calculated according to the medicament dosage adjustment sequence and the ore supply flow, the medicament volume is equal to the medicament dosage adjustment sequence multiplied by the ore pulp volume of the flotation tank divided by the medicament density, the filling time window is set to 180 seconds, the flow of the medicament pump is controlled to be the medicament volume divided by 180 seconds to realize constant-speed filling, the medicament pump is started at the moment when the stirring intensity adjustment is to 50% to calculate the flow and continuously fill for 180 seconds to complete the medicament addition, the execution trajectories of the control channels are recorded, the aeration amount execution trajectory records the aeration amount values at each moment in the aeration amount adjustment process, including the values at the three-stage jump moments, the stirring intensity execution trajectory records the stirring intensity values per second in the slope adjustment process, the medicament dosage execution trajectory records the cumulative medicament dosage values in the filling process, the fusion state estimation values corresponding to the moments are recorded, the fusion state estimation values are collected at regular time intervals after each stage of the aeration amount, every second in the stirring intensity adjustment process and the medicament filling process, the execution trajectory data and the state estimation value data are associated according to the time stamps to form a time series data set, the data set includes a time column, an aeration amount column, a stirring intensity column, a medicament dosage column, a rheological state column, a bubble characteristic column and an electrochemical state column, state response data is obtained, and the state response data completely records the control action execution process and the process state change trajectory to reflect the control effect.

[0043] In a specific embodiment, the S5 step includes: The output sequence is extracted from the state response data, the output sequence is time-aligned with the reference trajectory, and the difference values are calculated point by point, the sum of squares of the difference values is calculated and the average value is taken to obtain the tracking deviation index; The execution trajectories of the control channels are extracted from the state response data, the difference values of the control amounts at adjacent moments of the control channels are calculated, the sum of squares of the difference values is calculated and the average value is taken to obtain the control change index; The tracking deviation index is calculated by ratio calculation with the target tracking deviation, the control change index is calculated by ratio calculation with the target control change, and the two ratios are weighted and summed to obtain the comprehensive performance index; The performance state is determined according to the comprehensive performance index, when the comprehensive performance index is lower than a threshold value, the gradients of the output weight and the control weight in the rolling optimization target function with respect to the comprehensive performance index are calculated, the output weight and the control weight are updated in the gradient direction, and the adjusted weight parameters are obtained.

[0044] Specifically, extracting the output sequence from the state response data refers to screening the time history of the output variable from the recorded time series data, the state response data containing multiple columns of data, among which the output variable column corresponds to the concentrate quality index and the recovery rate index, extracting these two columns of data to form an output sequence in chronological order, the output sequence being a two-dimensional array, each row corresponding to a time point and each row containing two elements, namely the quality value and the recovery rate value. The output sequence is time-aligned with the reference trajectory, which is a preset expected output value time sequence also containing target quality and target recovery rate at each time point. Time alignment ensures that each time point of the output sequence and the reference trajectory corresponds one-to-one. After alignment, the difference value is calculated point by point, the first time point quality difference value being the first time point quality value of the output sequence minus the first time point target quality of the reference trajectory, the first time point recovery rate difference value being the first time point recovery rate value of the output sequence minus the first time point target recovery rate of the reference trajectory, and the difference values at all time points are calculated in turn to obtain a difference sequence. The square sum of the difference values is calculated, the square sum of the quality difference values at each time point is accumulated to obtain the quality error square sum, the square sum of the recovery rate difference values at each time point is accumulated to obtain the recovery rate error square sum, and the two square sums are added to obtain the total error square sum. The average error square is obtained by dividing the total error square sum by the total number of time points, and the average error square is used as the tracking deviation index to quantify the accuracy of the output tracking reference trajectory.

[0045] Extracting the execution trajectory of each control channel from the state response data refers to extracting the time series data of the air volume column, the stirring intensity column and the reagent amount column respectively. The air volume execution trajectory records the air volume value at each time point, the stirring intensity execution trajectory records the stirring intensity value at each time point, and the reagent amount execution trajectory records the cumulative reagent amount value at each time point. The square sum of the difference values of the control amount of adjacent time points of each control channel is calculated. The first air volume change value is obtained by subtracting the first time point air volume from the second time point air volume, the second air volume change value is obtained by subtracting the second time point air volume from the third time point air volume, and the difference values at all adjacent time points are calculated in turn. The square sum of all air volume change values is obtained by squaring and accumulating the air volume change values. The stirring intensity channel and the reagent amount channel use the same method to calculate their respective change square sums. The change square sums of the three control channels are added to obtain the total control change square sum. The average control change is obtained by dividing the total control change square sum by the total number of difference values. The average control change is used as the control change index to quantify the intensity of the control action.

[0046] The tracking deviation index is calculated by ratio with a target tracking deviation, the target tracking deviation is a preset acceptable tracking error level, the tracking deviation index is divided by the target tracking deviation to obtain a tracking performance ratio, a ratio equal to 1 indicates that the tracking performance reaches the target level, a ratio greater than 1 indicates that the tracking performance is insufficient, and a ratio less than 1 indicates that the tracking performance exceeds the target. The control change index is calculated by ratio with a target control change, the target control change is a preset acceptable control fluctuation level, the control change index is divided by the target control change to obtain a control stability ratio, a ratio equal to 1 indicates that the control stability reaches the target, a ratio greater than 1 indicates that the control is too intense, and a ratio less than 1 indicates that the control is sufficiently stable. The two ratios are weighted and squared to obtain a comprehensive performance index, the tracking performance ratio is squared and multiplied by a weight factor 0.55, and the control stability ratio is squared and multiplied by a weight factor 0.45. The weight factor allocation reflects the control target that tracking performance is prior to control stability. The comprehensive performance index is a dimensionless scalar that comprehensively evaluates the overall performance of the control system.

[0047] According to the comprehensive performance index, the performance state is determined, and a performance threshold of 0.90 is set as a determination reference. When the comprehensive performance index is greater than or equal to 0.90, it is determined that the performance meets the standard, and the controller parameters do not need to be adjusted. When the comprehensive performance index is less than 0.90, it is determined that the performance does not meet the standard, and the controller parameters need to be adjusted. The gradients of the output weight and the control weight in the rolling optimization objective function with respect to the comprehensive performance index are calculated. The gradient describes the rate of change of the comprehensive performance index caused by a small change in the weight parameter. The output weight gradient is the partial derivative of the comprehensive performance index with respect to the output weight, and the control weight gradient is the partial derivative of the comprehensive performance index with respect to the control weight. The gradient calculation uses a numerical differentiation method. The output weight is increased by a small amount, and the comprehensive performance index is recalculated. The difference between the new index and the original index is divided by the weight increment to obtain an approximate value of the output weight gradient. The control weight gradient is calculated in the same way. The output weight and the control weight are numerically updated in the gradient direction. The gradient direction indicates the parameter change direction that increases the comprehensive performance index. The output weight update value is equal to the current output weight plus a learning rate multiplied by the output weight gradient. The control weight update value is equal to the current control weight plus a learning rate multiplied by the control weight gradient. The learning rate is set to 0.05 to control the parameter adjustment step. The updated output weight and control weight are used as the adjusted weight parameters in the rolling optimization of the next control period. The adaptive adjustment of the weight parameters continuously optimizes the performance of the controller.

[0048] In a specific embodiment, according to the comprehensive performance index, when the comprehensive performance index is less than the threshold, the gradients of the output weight and the control weight in the rolling optimization objective function with respect to the comprehensive performance index are calculated, and the output weight and the control weight are numerically updated in the gradient direction to obtain the adjusted weight parameters, including: The comprehensive performance index is compared with the performance threshold value, and when the comprehensive performance index is less than the performance threshold value, it is determined that the performance is not up to standard, and the relative deviation amount of the tracking deviation index and the control change index is extracted; According to the deviation relationship between the tracking deviation index and the target tracking deviation, and the deviation relationship between the control change index and the target control change, the output weight adjustment direction and the control weight adjustment direction are determined; The output weight and the control weight in the rolling optimization objective function are combined into a weight parameter vector, and the partial derivative of the comprehensive performance index with respect to the weight parameter vector is obtained, and a gradient vector is obtained; The gradient vector is multiplied by the learning rate to obtain a weight adjustment amount, and the weight adjustment amount is added to the current output weight and the current control weight to obtain an adjusted weight parameter.

[0049] Specifically, the performance threshold value is set to 0.90 as the minimum standard for the performance of the control system, and when the comprehensive performance index value is greater than or equal to 0.90, it indicates that the comprehensive performance of the tracking performance and the control stability of the control system meets the expectation and does not need to adjust the controller parameters, and when the comprehensive performance index is less than 0.90, it is determined that the performance is not up to standard and the controller parameters need to be adjusted to improve the performance, and the relative deviation amount of the tracking deviation index and the control change index is extracted, the relative deviation amount describes the deviation degree of the actual index from the target index, the tracking deviation relative deviation amount is the tracking deviation index minus the target tracking deviation divided by the target tracking deviation, and the positive value indicates that the tracking error exceeds the target and needs to be improved, and the negative value indicates that the tracking error is less than the target and the tracking performance is good, the control change relative deviation amount is the control change index minus the target control change divided by the target control change, and the positive value indicates that the control fluctuation exceeds the target and needs to be improved, and the negative value indicates that the control fluctuation is less than the target and the control is stable enough, and the two relative deviation amounts provide diagnostic information for adjusting the controller parameters.

[0050] According to the deviation relationship between the tracking deviation index and the target tracking deviation, and the deviation relationship between the control change index and the target control change, the output weight adjustment direction and the control weight adjustment direction are determined, the output weight adjustment direction depends on the tracking performance requirement, when the tracking deviation index is greater than 1.2 times the target tracking deviation, it indicates that the tracking error is too large and the control change index is less than the target control change, which indicates that the control action is not active enough, and it is determined that the output weight should be increased to make the controller pay more attention to the tracking performance and the control weight should be reduced to allow the control action to be more active, when the tracking deviation index is less than the target tracking deviation but the control change index is greater than 1.3 times the target control change, it indicates that the tracking performance is good but the control is too violent, and it is determined that the output weight should be reduced to reduce the tracking requirement and the control weight should be increased to suppress the control fluctuation, the output weight adjustment direction is to increase or decrease, and the control weight adjustment direction is opposite to the output weight, which reflects the trade-off relationship between the tracking performance and the control stability, and after the adjustment direction is determined, the specific adjustment amplitude needs to be calculated.

[0051] The output weight and the control weight in the rolling optimization objective function are combined into a weight parameter vector, the output weight includes two elements of the quality index weight and the recovery rate index weight, the control weight includes three elements of the aeration amount increment weight, the stirring intensity increment weight and the medicament amount increment weight, the five weight parameters are arranged in sequence to form the weight parameter vector, and a partial derivative of the comprehensive performance index with respect to the weight parameter vector is calculated. Since there is a complex nonlinear relationship between the comprehensive performance index and the weight parameter, it is difficult to obtain an analytical partial derivative expression, and a numerical differentiation method is used to calculate the partial derivative. A small disturbance amount is added to the first element of the weight parameter vector, for example, the quality index weight is increased by 1% of the current value, and other weight parameters remain unchanged, and a control cycle is re-executed to calculate a new comprehensive performance index. The new index is subtracted from the original index to obtain an index increment, and the index increment is divided by the weight disturbance amount to obtain the partial derivative of the quality index weight. The corresponding partial derivatives are calculated by sequentially adding disturbances to the five weight parameters, and the five partial derivatives are arranged in sequence to form a gradient vector. The gradient vector indicates which direction the weight parameter changes in to make the comprehensive performance index increase the fastest.

[0052] The gradient vector is multiplied by a learning rate to obtain a weight adjustment amount, the learning rate is a scalar parameter controlling the step size of weight adjustment, the learning rate is set to 0.05, which means that each adjustment is 5% in the gradient direction. If the learning rate is too large, the adjustment will be excessive, causing oscillation. If the learning rate is too small, the convergence will be slow. Each element of the gradient vector is multiplied by the learning rate to obtain the adjustment amount of the corresponding weight parameter. The quality index weight adjustment amount is the product of the quality index weight gradient and 0.05. The recovery rate index weight adjustment amount is the product of the recovery rate index weight gradient and 0.05. The three control increment weight adjustment amounts are calculated in the same way. The five adjustment amounts constitute a weight adjustment amount vector. The weight adjustment amount is added to the current output weight and the current control weight. The updated quality index weight is obtained by adding the corresponding adjustment amount to the current quality index weight. The updated recovery rate index weight is obtained by adding the corresponding adjustment amount to the current recovery rate index weight. The three control increment weights are also updated by adding their respective adjustment amounts. The updated five weight parameters are used as the adjusted weight parameters in the rolling optimization of the next control period. The gradient descent method gradually adjusts the controller parameters in the direction of performance improvement.

[0053] The beneficiation equipment process parameter adaptive control method in the embodiments of the present application is described above, and the beneficiation equipment process parameter adaptive control system in the embodiments of the present application is described below. Please refer to Figure 3 The beneficiation equipment process parameter adaptive control system in the embodiments of the present application includes one embodiment: The fusion module is used for acquiring a first rheological parameter at an inlet of the flotation tank, a second rheological parameter at a middle part, a bubble characteristic parameter, a third rheological parameter at an outlet and an electrochemical parameter, performing redundant fusion processing on the first rheological parameter, the second rheological parameter and the third rheological parameter by a weighted least square method, and obtaining a fusion state estimation value; The identification module is used for constructing a state transition matrix and an input matrix according to the fusion state estimation value, performing online identification on a parameter vector of the state transition matrix and the input matrix by a recursive least square algorithm, and obtaining an updated parameter vector; The solving module is used for substituting the updated parameter vector into a rolling optimization objective function, performing quadratic programming solving on a control increment sequence, and obtaining an optimal control output; The hierarchical module is used for decomposing the optimal control output into an aeration amount adjustment sequence, a stirring intensity adjustment sequence and a reagent amount adjustment sequence, performing hierarchical execution on the aeration amount adjustment sequence, the stirring intensity adjustment sequence and the reagent amount adjustment sequence according to time sequence priorities, and obtaining state response data; The adjustment module is used for calculating a tracking deviation index and a control change index according to the state response data, performing gradient adjustment on output weights and control weights in the rolling optimization objective function, and obtaining adjusted weight parameters.

[0054] The above Figure 3 The mineral processing equipment process parameter adaptive control system in the embodiment of the application is described in detail from the perspective of a modular functional entity, and the mineral processing equipment process parameter adaptive control device in the embodiment of the application is described in detail from the perspective of hardware processing.

[0055] Referring Figure 4 , the embodiment of the application further provides a mineral processing equipment process parameter adaptive control device, which can be a server, and the internal structure thereof can be as shown in Figure 4 The mineral processing equipment process parameter adaptive control device comprises a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. The processor of the computer is used to provide calculation and control capabilities. The memory of the mineral processing equipment process parameter adaptive control device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the mineral processing equipment process parameter adaptive control device is used to store the corresponding data in the embodiment. The network interface of the mineral processing equipment process parameter adaptive control device is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement the above method.

[0056] Those skilled in the art can understand, Figure 4 The structure shown in the figure is only a block diagram of part of the structure related to the technical scheme of the present application, and does not constitute a limitation on the process parameter adaptive control device for the ore dressing equipment to which the technical scheme of the present application is applied.

[0057] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium, and the computer readable storage medium stores instructions, and when the instructions are run on a computer, the computer executes the steps of the process parameter adaptive control method for the ore dressing equipment.

[0058] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, system and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0059] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical scheme of the present application or the part of the prior art that essentially contributes or the whole or part of the technical scheme can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing an ore dressing equipment process parameter adaptive control device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0060] The above embodiments are only used to illustrate the technical scheme of the present application, but not to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical scheme recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical scheme deviate from the spirit and scope of the technical scheme of the present application.

Claims

1. An adaptive control method for process parameters of mineral processing equipment, characterized in that, The method includes: Step S1: Obtain the first rheological parameter at the inlet of the flotation cell, the second rheological parameter and bubble characteristic parameter in the middle, and the third rheological parameter and electrochemical parameter at the outlet. Perform redundant fusion processing on the first rheological parameter, the second rheological parameter and the third rheological parameter using the weighted least squares method to obtain the fusion state estimate. Step S2: Construct a state transition matrix and an input matrix based on the fused state estimate, and use a recursive least squares algorithm to identify the parameter vectors of the state transition matrix and the input matrix online to obtain the updated parameter vector; Step S3: Substitute the updated parameter vector into the rolling optimization objective function, perform quadratic programming on the control increment sequence, and obtain the optimal control output; Step S4: Decompose the optimal control output into an aeration volume adjustment sequence, a stirring intensity adjustment sequence, and a drug dosage adjustment sequence. Execute the aeration volume adjustment sequence, the stirring intensity adjustment sequence, and the drug dosage adjustment sequence in a hierarchical manner according to their time priority to obtain state response data. Step S5: Calculate the tracking deviation index and control change index based on the state response data, and perform gradient adjustment on the output weight and control weight in the rolling optimization objective function to obtain the adjusted weight parameters.

2. The adaptive control method for process parameters of mineral processing equipment according to claim 1, characterized in that, Step S1 includes: The first rheological parameter at the inlet of the flotation cell, the second rheological parameter in the middle, and the third rheological parameter at the outlet are obtained. A set of constraint equations is constructed based on the material conservation relationship. The first rheological parameter, the second rheological parameter, and the third rheological parameter are substituted into the set of constraint equations and weighted least squares solution is performed to obtain the preliminary fusion value. Calculate the measurement residual vector of the constraint equation system, and standardize the measurement residual vector to obtain the standardized residual of each measurement point; Based on the standardized residual, abnormal measurement points are identified. After reducing the weight coefficients corresponding to the abnormal measurement points, a weighted least squares solution is performed again to obtain the corrected fusion value. The corrected fusion value is combined with the bubble characteristic parameters and the electrochemical parameters to obtain the fusion state estimate.

3. The adaptive control method for process parameters of mineral processing equipment according to claim 1, characterized in that, Step S2 includes: A state-space expression is constructed based on the fused state estimate, and the state transition matrix and input matrix in the state-space expression are expanded into a parameter vector. Collect the control input sequence and the actual output sequence, construct the regression vector and the covariance matrix, and substitute the regression vector and the covariance matrix into the recursive least squares formula to identify the parameters and obtain the identified parameter vector. Calculate the rate of change of the identified parameter vector with respect to the parameter vector at the previous moment, determine the convergence state of the parameters based on the rate of change of the parameters, and obtain the convergence determination result; When the convergence determination result satisfies the convergence condition, the identification parameter vector is output as the update parameter vector, and the covariance matrix is ​​updated.

4. The adaptive control method for process parameters of mineral processing equipment according to claim 1, characterized in that, Step S3 includes: The updated parameter vector is decomposed into a state transition matrix and an input matrix according to the dimensions, and the output predicted value sequence at each time point in the prediction time domain is calculated based on the state transition matrix and the input matrix. The output predicted value sequence is compared with the reference trajectory at the corresponding time point by point. The weighted square sum of the differences at each time point is then obtained to obtain the tracking error term. The control increments at each moment in the control time domain are summed by weighted squares to obtain the control cost term. The tracking error term is added to the control cost term to form the rolling optimization objective function. The upper limits of the rate of change of aeration volume, the rate of change of stirring intensity, and the rate of change of drug dosage are set as constraints. The rolling optimization objective function is solved by quadratic programming to obtain the control increment sequence. The first element of the control increment sequence is added to the control quantity at the previous moment to obtain the optimal control output.

5. The adaptive control method for process parameters of mineral processing equipment according to claim 1, characterized in that, Step S4 includes: The optimal control output is decomposed into aeration volume control command, stirring intensity control command and drug dosage control command according to the control channel. The difference between each control command and the control quantity at the previous moment is calculated to obtain the aeration volume adjustment sequence, stirring intensity adjustment sequence and drug dosage adjustment sequence. The inflation volume adjustment sequence is decomposed into three step increments. Each step increment is executed sequentially at a set time interval. After each step is executed, the fusion state estimate is collected, and the state change rate is calculated. When the state change rate exceeds a threshold, the subsequent step increment execution is paused. After the aeration volume adjustment sequence is completed, the stirring intensity adjustment sequence is converted into a ramp tracking command. The stirring intensity is gradually adjusted according to the set slope, and the state coupling index is monitored synchronously. When the state coupling index exceeds the upper limit, the adjustment amplitude is attenuated. When the stirring intensity adjustment sequence reaches the preset progress, the drug dosage adjustment sequence is initiated, the execution trajectory of each control channel and the estimated fusion state value at the corresponding time are recorded to obtain state response data.

6. The adaptive control method for process parameters of mineral processing equipment according to claim 5, characterized in that, Step S5 includes: The output sequence is extracted from the state response data. The output sequence is time-aligned with the reference trajectory, and the difference is calculated point by point. The difference is squared, summed, and averaged to obtain the tracking deviation index. The execution trajectory of each control channel is extracted from the state response data. The control change index is obtained by summing the squares of the control quantity differences between adjacent time points of each control channel and taking the average value. The tracking deviation index is calculated as a ratio to the target tracking deviation, and the control change index is calculated as a ratio to the target control change. The two ratios are then weighted and summed to obtain the comprehensive performance index. The performance status is determined based on the comprehensive performance index. When the comprehensive performance index is lower than the threshold, the gradient of the output weight and control weight in the rolling optimization objective function with respect to the comprehensive performance index is calculated. The output weight and control weight are then updated numerically according to the gradient direction to obtain the adjusted weight parameters.

7. The adaptive control method for process parameters of mineral processing equipment according to claim 6, characterized in that, The process involves determining the performance status based on the comprehensive performance index. When the comprehensive performance index is below a threshold, the gradients of the output weights and control weights in the rolling optimization objective function with respect to the comprehensive performance index are calculated. The output weights and control weights are then numerically updated according to the gradient direction to obtain adjusted weight parameters, including: The comprehensive performance index is compared with the performance threshold. When the comprehensive performance index is less than the performance threshold, it is determined that the performance is substandard. The relative deviation between the tracking deviation index and the control change index is extracted. Based on the deviation relationship between the tracking deviation index and the target tracking deviation, and the deviation relationship between the control change index and the target control change, the direction of output weight adjustment and the direction of control weight adjustment are determined. The output weights and control weights in the rolling optimization objective function are combined into a weight parameter vector. The gradient vector is obtained by taking the partial derivative of the comprehensive performance index with respect to the weight parameter vector. The gradient vector is multiplied by the learning rate to obtain the weight adjustment amount. The weight adjustment amount is then added to the current output weight and the current control weight to obtain the adjusted weight parameters.

8. An adaptive control system for process parameters of mineral processing equipment, characterized in that, For implementing the adaptive control method for process parameters of mineral processing equipment as described in any one of claims 1-7, the adaptive control system for process parameters of mineral processing equipment comprises: The fusion module is used to acquire the first rheological parameter at the inlet of the flotation cell, the second rheological parameter and bubble characteristic parameter in the middle, and the third rheological parameter and electrochemical parameter at the outlet. The first rheological parameter, the second rheological parameter and the third rheological parameter are redundantly fused using the weighted least squares method to obtain the fusion state estimate. The identification module is used to construct a state transition matrix and an input matrix based on the fused state estimate, and to identify the parameter vectors of the state transition matrix and the input matrix online using a recursive least squares algorithm to obtain an updated parameter vector. The solution module is used to substitute the updated parameter vector into the rolling optimization objective function, perform quadratic programming on the control increment sequence, and obtain the optimal control output. The grading module is used to decompose the optimal control output into an aeration volume adjustment sequence, a stirring intensity adjustment sequence, and a drug dosage adjustment sequence, and to execute the aeration volume adjustment sequence, the stirring intensity adjustment sequence, and the drug dosage adjustment sequence in a graded manner according to the time priority to obtain state response data; The adjustment module is used to calculate the tracking deviation index and control change index based on the state response data, and to perform gradient adjustment on the output weights and control weights in the rolling optimization objective function to obtain the adjusted weight parameters.

9. An adaptive control device for process parameters of a mineral processing plant, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the adaptive control method for process parameters of the mineral processing equipment as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the adaptive control method for process parameters of the mineral processing equipment as described in any one of claims 1 to 7.

Images