Prebaked anode grinding production control system based on feedback control

Abstract

The present application relates to the technical field of production process control, in particular to a prebaked anode grinding production control system based on feedback control, the system comprises: a deviation calculation module acquires acoustic vibration, flow and particle size and calculates filling state, output and particle size deviation, a state modeling module constructs an observation vector based on the deviation and performs Kalman filtering, a weighted comprehensive grinding state index is generated, a regulation and decision module performs PID operation on the index to generate static blade and turbine adjustment instructions, and a feedback module generates pulses and voltage to be sent to an actuator. In the present application, an observation vector is constructed and a state estimation value is output by filtering and removing noise, the deviation is weighted and fused to generate a comprehensive index reflecting the load of the mill, accurate driving instructions are obtained according to the index, dynamic closed-loop control is realized to compensate for the real-time fluctuation of the raw material characteristics, and the estimated results are used to cooperatively adjust the classification parameters to ensure that the finished product particle size meets the standards and improve the overall energy efficiency of the grinding operation.

Description

Technical Field

[0001] This invention relates to the field of production process control technology, and in particular to a prebaked anode grinding production control system based on feedback control. Background Technology

[0002] In the field of production process control technology, production process control refers to the general technical scope of using industrial automation instruments, programmable logic controllers, and upper-level computer configuration software to centrally monitor, logically interlock, and coordinate the operating status and process parameters of mechanical equipment in industrial production lines. Traditionally, control systems used in prebaked anode grinding production employ a PLC control cabinet as the core hardware, connecting physical components such as Raymond mills, bucket elevators, electromagnetic vibrating feeders, belt conveyors, and bag filters via electrical cables. In actual production operations, this system controls the speed of the feeding motor to feed the calcined coke raw material into the mill grinding chamber. The material is pulverized by the mechanical extrusion of the grinding rollers and grinding rings. The fine powder after grinding enters the cyclone separator for physical sedimentation and collection with the airflow generated by the fan, while coarse particles that do not meet the particle size standard are intercepted by the classifier impeller and fall back to the bottom of the grinding chamber for secondary grinding.

[0003] Traditional prebaked anode grinding production relies on PLCs for basic logic control of mechanical equipment. This operation mode, based on fixed electrical connections and mechanical extrusion, makes it difficult to deeply perceive the complex grinding load state inside the mill. Relying solely on fan airflow and physical sedimentation for material collection results in poor adaptability of the system to fluctuations in raw material characteristics. The extensive control strategy exhibits significant lag in the response of the actuator when faced with changes in the filling state inside the grinding chamber. The non-adaptive mechanical adjustment method is prone to causing uneven particle size distribution in the finished product and low grinding efficiency, leading to increased production energy consumption and insufficient equipment operation stability. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and propose a prebaked anode grinding production control system based on feedback control.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a prebaked anode grinding production control system based on feedback control includes:

[0006] The deviation calculation module acquires the acoustic vibration voltage signal, real-time mass flow rate value, and particle size distribution percentage value. It calculates the difference between the acoustic vibration voltage signal and the preset reference value to generate the filling state deviation, calculates the difference between the real-time mass flow rate value and the preset target output to generate the output deviation, and calculates the difference between the particle size distribution percentage value and the preset target range to generate the particle size deviation.

[0007] The state modeling module constructs a state observation vector based on the filling state deviation, the production deviation, and the particle size deviation. It performs Kalman filtering on the state observation vector to output a state estimate, and then performs a weighted summation of the state estimate with a preset weight coefficient matrix to generate a comprehensive grinding state index.

[0008] The control and decision module performs PID calculations on the comprehensive grinding state index to obtain proportional response, integral response and derivative response, and generates stationary blade opening adjustment command and turbine speed adjustment command based on the proportional response, integral response and derivative response.

[0009] The execution feedback module generates a stepper motor drive pulse based on the stationary blade opening adjustment command and sends it to the stationary blade actuator, and generates a frequency converter control voltage based on the turbine speed adjustment command and sends it to the turbine blade drive motor.

[0010] As a further aspect of the present invention, the filling state deviation specifically refers to the voltage difference data of the material filling state in the mill relative to the benchmark operating condition; the output deviation specifically refers to the quality difference data of the current production flow rate relative to the preset capacity target; the particle size deviation specifically refers to the percentage difference data of the fineness distribution of the finished powder relative to the qualified range; the comprehensive grinding state index includes the system operating efficiency evaluation value and the quality state evaluation value after weighted fusion; the stationary blade opening adjustment command includes the target opening value and stepping direction for the guide vanes of the cyclone dynamic separator; the turbine speed adjustment command includes the target speed setting value and adjustment range for the turbine classifier impeller; the stepper motor drive pulse is specifically a digital electrical signal sequence with a target frequency and count; and the frequency converter control voltage is specifically an analog voltage signal corresponding to the target output frequency.

[0011] As a further aspect of the present invention, the specific function of the deviation calculation module is as follows:

[0012] The acoustic signal preprocessing submodule acquires the acoustic vibration voltage signal, uses a preset bandpass filter to remove the ambient background noise frequency band to retain the grinding sound characteristic frequency band, performs a sliding window root mean square operation on the filtered signal to extract the effective energy value, and performs differential calculation between the effective energy value and the preset no-load reference voltage to generate the filling state deviation.

[0013] The flow data standardization submodule acquires real-time quality flow rate values, reads preset rated capacity settings, calculates the algebraic difference between the real-time quality flow rate values ​​and the rated capacity settings, and normalizes the calculated algebraic difference based on the range coefficient of the flow sensor to generate production deviation.

[0014] The particle size distribution analysis submodule obtains the particle size distribution percentage value, identifies the cumulative distribution proportion within the qualified particle size range, calculates the difference between the cumulative distribution proportion and the preset target qualified rate, and generates the particle size deviation.

[0015] As a further aspect of the present invention, the specific functional implementation of the state modeling module is as follows:

[0016] The vector space construction submodule arranges and combines the filling state deviation, the yield deviation, and the granularity deviation at the same sampling time in a preset order to construct the state observation vector at the current time.

[0017] The filtering iterative estimation submodule calls the preset system state transition matrix and observation noise covariance matrix, performs prior state prediction for the current time based on the optimal state estimate of the previous time, and uses the state observation vector to correct and update the prior state prediction result, and outputs the state estimate.

[0018] The weighted index synthesis submodule obtains the weight coefficient matrix corresponding to multiple dimensional components in the state estimate, performs a linear weighted summation operation on the state estimate, and maps the summation result to a preset scoring interval to generate a comprehensive grinding state index.

[0019] As a further aspect of the present invention, the specific function of the control decision module is as follows:

[0020] The error feature extraction submodule calculates the difference between the comprehensive grinding state index and the preset ideal steady-state index to obtain the comprehensive system error, and stores the comprehensive system error based on the time series to calculate the cumulative error and the rate of change error.

[0021] The PID response calculation submodule obtains the preset proportional gain coefficient, integral time constant, and derivative time constant, respectively, and performs multiplication operations on the system comprehensive error, the cumulative error, and the rate of change error to generate the proportional response, integral response, and derivative response.

[0022] The multivariable instruction generation submodule linearly superimposes the proportional response, the integral response, and the derivative response to obtain the total adjustment amount. Based on a preset decoupling control strategy, the total adjustment amount is allocated to the stationary blade control channel and the turbine control channel to generate stationary blade opening adjustment instructions and turbine speed adjustment instructions, respectively.

[0023] As a further aspect of the present invention, the specific function of the execution feedback module is as follows:

[0024] The pulse sequence mapping submodule parses the target angle change in the stationary blade opening adjustment command, calculates the required total number of pulses based on the step angle parameters of the stepper motor, determines the pulse frequency based on the adjustment speed requirements, and generates stepper motor drive pulses.

[0025] The analog voltage conversion submodule parses the target speed value in the turbine speed adjustment command, calculates the corresponding output voltage amplitude based on the linear correspondence between the rated speed of the frequency converter and the control voltage, and generates the frequency converter control voltage through the digital-to-analog conversion circuit.

[0026] As a further aspect of the present invention, the process by which the acoustic signal preprocessing submodule generates the filling state deviation includes:

[0027] Set a time sliding window of length N, and collect N discrete acoustic vibration voltage signal sample points within the window;

[0028] Calculate the arithmetic mean of the sum of squares of the N sample points, and then take the square root of the arithmetic mean to obtain the effective value;

[0029] Obtain the reference voltage RMS value of the mill under no-load operation as pre-calibrated;

[0030] Calculate the difference between the effective value and the effective value of the reference voltage, and define the difference as the fill state deviation.

[0031] As a further aspect of the present invention, the process by which the filtering iterative estimation submodule outputs the state estimate includes:

[0032] The prior estimate for the current time step is calculated based on the posterior estimate of the previous time step and the system state transition matrix.

[0033] Calculate the prediction error covariance matrix at the current time, and combine it with the observation noise covariance matrix to calculate the Kalman gain matrix;

[0034] Calculate the residual between the state observation vector and the predicted observation corresponding to the prior estimate;

[0035] The residuals are weighted using the Kalman gain matrix, and the weighted result is superimposed on the prior estimate to generate the state estimate.

[0036] As a further aspect of the present invention, the calculation process of the PID response calculation submodule in generating the proportional response, the integral response, and the derivative response specifically conforms to the following response function model:

[0037] ;

[0038] in, This represents the total adjustment of the output at time k. This represents the proportional gain coefficient. Represents the integral coefficient. Represents the differential coefficient, The system's overall error at time k is represented by the following. The system composite error at time k-1 is represented by [the error name]. Represents the sampling period of the control system. The item corresponds to the aforementioned proportional response. The term corresponds to the integral response, The term corresponds to the differential response.

[0039] As a further aspect of the present invention, the process by which the pulse sequence mapping submodule generates the stepper motor drive pulse includes:

[0040] Obtain the target adjustment angle included in the stationary blade opening adjustment command;

[0041] Read the inherent step angle and transmission reduction ratio of the stepper motor in the stationary blade actuator;

[0042] The required pulse count value is obtained by dividing the target adjustment angle by the inherent step angle, multiplying it by the transmission reduction ratio, and rounding down.

[0043] The trajectory of pulse frequency change is determined based on a preset acceleration / deceleration curve model, and the stepper motor drive pulse, including direction signal and clock signal, is generated by combining the pulse count value.

[0044] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0045] In this invention, real-time deviations of multi-dimensional parameters relative to benchmark values ​​are calculated by collecting acoustic vibration and flow data. A state observation vector is constructed, and Kalman filtering is used to remove signal noise and output accurate state estimates. Multi-source deviation information is weighted and fused to generate a comprehensive index reflecting the internal load of the mill. Based on this index, proportional-integral-differential operations are performed to obtain precise drive commands for the stationary blades and turbine components. Dynamic logic closed-loop control realizes real-time compensation for fluctuations in raw material characteristics. Based on the state estimation results, the grading parameters are coordinated to ensure that the particle size distribution of the finished product meets the standards and improve the overall energy efficiency of the grinding operation. Attached Figure Description

[0046] Figure 1 This is a block diagram illustrating the principle of the prebaked anode grinding production control system of the present invention.

[0047] Figure 2 This is a flowchart illustrating the internal operating logic of the deviation calculation module of the present invention.

[0048] Figure 3 This is a flowchart of the internal data processing of the state modeling module of the present invention;

[0049] Figure 4 This is a flowchart of the PID control logic inside the control decision module of the present invention;

[0050] Figure 5 This is a flowchart illustrating the internal signal conversion process of the feedback module in this invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the software-based technical solution is described in detail below with reference to system architecture diagrams and embodiments. It should be understood that the specific embodiments described herein are only for explaining the technical solutions of this invention and do not constitute a limitation on the scope of protection.

[0052] In the description of this invention, the system architecture relationships or data processing flows indicated by terms such as "layer," "module," "interface," "data flow," "client," and "server" are all defined based on the architecture diagram or flowchart corresponding to the embodiments. This way of describing is only used to clearly illustrate the logical relationships between the elements in the technical solution, and not to limit the physical deployment form. The term "multiple" includes two or more technical units, including but not limited to multiple data nodes, processing threads, service instances, or functional components and other scalable elements. The specific number is determined according to the actual business scenario and needs to be specifically specified.

[0053] Please see Figure 1 and Figure 2 This invention provides a technical solution: a prebaked anode grinding production control system based on feedback control includes:

[0054] The deviation calculation module acquires the acoustic vibration voltage signal, real-time mass flow rate value, and particle size distribution percentage value. It calculates the difference between the acoustic vibration voltage signal and the preset reference value to generate the filling state deviation, calculates the difference between the real-time mass flow rate value and the preset target output to generate the output deviation, and calculates the difference between the particle size distribution percentage value and the preset target range to generate the particle size deviation.

[0055] The filling state deviation specifically refers to the voltage difference data of the material filling state in the mill relative to the reference operating condition;

[0056] Production deviation specifically refers to the quality difference data between the current production flow and the preset capacity target.

[0057] Particle size deviation specifically refers to the percentage difference in the fineness distribution of the finished powder relative to the acceptable range;

[0058] The specific functions of the deviation calculation module are as follows:

[0059] The acoustic signal preprocessing submodule acquires the acoustic vibration voltage signal, uses a preset bandpass filter to remove the ambient background noise frequency band to retain the grinding sound characteristic frequency band, performs a sliding window root mean square operation on the filtered signal to extract the effective energy value, and performs differential calculation between the effective energy value and the preset no-load reference voltage to generate the filling state deviation.

[0060] The process of generating fill state bias in the acoustic signal preprocessing submodule includes:

[0061] Set a time sliding window of length N, and collect N discrete acoustic vibration voltage signal sample points within the window;

[0062] Calculate the arithmetic mean of the sum of squares of N sample points, and then take the square root of the arithmetic mean to obtain the effective value;

[0063] Obtain the reference voltage RMS value of the mill under no-load operation as pre-calibrated;

[0064] Calculate the difference between the effective value and the effective value of the reference voltage, and define the difference as the fill state deviation;

[0065] The flow data standardization submodule acquires real-time quality flow rate values, reads preset rated capacity settings, calculates the algebraic difference between real-time quality flow rate values ​​and rated capacity settings, and normalizes the calculated algebraic difference based on the range coefficient of the flow sensor to generate production deviation.

[0066] The particle size distribution analysis submodule obtains the particle size distribution percentage value, identifies the cumulative distribution proportion within the qualified particle size range, calculates the difference between the cumulative distribution proportion and the preset target qualified rate, and generates the particle size deviation.

[0067] The acoustic signal preprocessing submodule, along with the flow data standardization and particle size distribution analysis submodules, operates in parallel during execution. They acquire raw data and convert it into normalized deviation values ​​through a specific hardware interface protocol. First, for the acoustic signal preprocessing submodule, the system continuously acquires analog voltage signals at a sampling frequency of 10kHz using piezoelectric accelerometers installed at specific locations on the outside of the ball mill cylinder (typically three key monitoring points distributed axially along the middle of the cylinder). This acoustic vibration voltage signal refers to the continuous analog voltage waveform data output after the mechanical vibration of the ball mill cylinder wall caused by the impact of the grinding media is converted into a proportional amount of charge by the piezoelectric accelerometer, and then conditioned by a charge amplifier. The amplitude characteristics of this signal directly reflect the energy intensity of the steel balls impacting the liner inside the mill. Due to the presence of low-frequency noise (approximately 50-500Hz) from the meshing of the reducer gears and high-frequency airflow noise from the surrounding fans, the submodule first invokes a fourth-order Butterworth bandpass filter algorithm, setting the lower cutoff frequency to 800Hz and the upper cutoff frequency to 4000Hz. This frequency band concentrates the characteristic frequencies of the steel ball impacting the liner and the material. The filtered digital signal sequence then enters the length... A time-sliding window is used. Within the window, the processor processes 2048 discrete voltage sample points. Perform squaring operations one by one, sum them up, and then divide by... Then, the square root of the result is taken to obtain the current root mean square (RMS) effective value. .

[0068] At the same time, the system calls the pre-calibrated no-load reference voltage in the storage unit. The calibration process for this benchmark value is as follows: With the ball mill stopped feeding and all material discharged, leaving only steel balls and running at rated speed, run continuously for 30 minutes, recording the RMS value every minute. The maximum and minimum values ​​are removed, and the arithmetic mean is taken. For example, after calibration... Set to 4.50V. If the currently calculated... If the voltage is 3.20V (indicating that the sound is attenuated due to material filling), the processor will perform a subtraction operation. This difference is the filling state deviation.

[0069] For the flow data standardization submodule, the system reads real-time mass flow values ​​via a Coriolis mass flow meter using the RS485 communication protocol. (Unit: t / h). Set the preset rated capacity setting value. The optimal operating point of the system, for example, 45.0 t / h. The full-scale coefficient of the flow meter. The speed is set to 100.0 t / h. The module first calculates the algebraic difference. Then divide by Perform normalization. For example, if the real-time flow rate is 42.5 t / h, then... Production deviation .

[0070] For the particle size distribution analysis submodule, the system connects to an online laser particle size analyzer and acquires a particle size distribution data packet every 5 minutes. The module parses this data packet and extracts the cumulative distribution percentage below a specific particle size (e.g., 75 μm, corresponding to 200 mesh). Read the preset target pass rate required by the process. (e.g., 85.0%).

[0071] Calculate particle size deviation .

[0072] If the current detected particles smaller than 75μm account for 82.5%, then (i.e. -2.5%).

[0073] Finally, the deviation calculation module synchronously transmits the calculated filling state deviation (-1.30), yield deviation (-0.025), and particle size deviation (-2.5) to the next-level state modeling module.

[0074] Please see Figure 1 and Figure 3 The state modeling module constructs a state observation vector based on filling state deviation, yield deviation, and particle size deviation. It performs Kalman filtering on the state observation vector to output state estimates, and then performs weighted summation on the state estimates and the preset weight coefficient matrix to generate a comprehensive grinding state index.

[0075] The comprehensive grinding status index includes the weighted and fused system operating efficiency assessment value and quality status assessment value;

[0076] The specific functional implementation of the state modeling module is as follows:

[0077] The vector space construction submodule arranges and combines the filling state deviation, yield deviation, and granularity deviation at the same sampling time in a preset order to construct the state observation vector at the current time.

[0078] The filtering iterative estimation submodule calls the preset system state transition matrix and observation noise covariance matrix, performs prior state prediction for the current time based on the optimal state estimate of the previous time, and uses the state observation vector to correct and update the prior state prediction result, and outputs the state estimate.

[0079] The process of the filtering iterative estimation submodule outputting the state estimate includes:

[0080] The prior estimate for the current time step is calculated based on the posterior estimate of the previous time step and the system state transition matrix.

[0081] Calculate the prediction error covariance matrix at the current time, and combine it with the observation noise covariance matrix to calculate the Kalman gain matrix;

[0082] Calculate the residual between the state observation vector and the predicted observation corresponding to the prior estimate;

[0083] The residuals are weighted using the Kalman gain matrix, and the weighted result is superimposed on the prior estimate to generate the state estimate.

[0084] The weighted index synthesis submodule obtains the weight coefficient matrix corresponding to multiple dimensional components in the state estimate, performs a linear weighted summation operation on the state estimate, and maps the summation result to a preset scoring interval to generate a comprehensive grinding state index.

[0085] After receiving the deviation data, the state modeling module constructs a state observation vector through preset matrix operation logic and performs Kalman filter iteration to finally synthesize a comprehensive grinding state index.

[0086] In the vector space construction submodule, the processor will use the same sampling time... Filling state deviation below Production deviation and particle size deviation according to The order of arrangement, the construction dimension is Current state observation vector .

[0087] The process then proceeds to the filtering iterative estimation submodule. The Kalman filtering described above is an algorithm that uses the state equations of a linear system and the system's input and output observation data to optimally estimate the system state. Its core lies in utilizing feedback control principles, employing two steps—"prediction" and "update"—to correct errors caused by noise, thereby reconstructing the true system state from noisy measurement data. The system has a preset state transition matrix. for Identity matrix (assuming the system state maintains inertia during extremely short sampling periods), observation noise covariance matrix It is a diagonal matrix, and the diagonal elements are respectively These correspond to the variance characteristics of acoustic, flow, and particle size measurements, respectively. System state noise covariance matrix. Set as A diagonal matrix of magnitude [number missing]. The specific iterative process is as follows:

[0088] First, utilize the previous moment posterior state estimate Predict the prior estimate at the current time. ,Right now .

[0089] Secondly, calculate the prior error covariance matrix. .

[0090] Next, the Kalman gain matrix is ​​calculated. The observation matrix It is an identity matrix.

[0091] Processor calculates residual And use gain-corrected prior estimation: Assume that after filtering iterations, the output state estimate vector... This step eliminates interference from instantaneous measurement noise.

[0092] The weighted index synthesis submodule calls the weight coefficient matrix. The weights are set based on the degree of influence of each parameter on the quality of the electrolytic aluminum anode:

[0093] Weighting of fill status (affecting energy consumption and wear) (Normalization coefficient), output (affecting capacity) is assigned a weight. (The value needs to be increased due to its small size), granularity (affecting product qualification rate) is assigned a weight. Perform a linear weighted summation operation:

[0094] ;

[0095] .

[0096] The system maps the calculated result 47.1 to a preset scoring range of 0-100 to generate a comprehensive grinding state index. If the index is lower than the preset ideal steady-state value (e.g., 85 points), it indicates that the system is in suboptimal operating condition.

[0097] Please see Figure 1 and Figure 4 The control and decision module performs PID calculations on the comprehensive grinding state index to obtain the proportional response, integral response and derivative response, and generates the stationary blade opening adjustment command and turbine speed adjustment command based on the proportional response, integral response and derivative response.

[0098] The stationary blade opening adjustment command includes the target opening value and step direction for the guide vanes of the cyclone dynamic separator.

[0099] The turbine speed regulation command includes the target speed setting value and adjustment range for the turbine staged impeller;

[0100] The specific functions of the regulation and decision-making module are as follows:

[0101] The error feature extraction submodule calculates the difference between the comprehensive grinding state index and the preset ideal steady-state index to obtain the comprehensive system error, and stores the comprehensive system error based on the time series to calculate the cumulative error and the rate of change error.

[0102] The PID response calculation submodule obtains the preset proportional gain coefficient, integral time constant, and derivative time constant, respectively, and performs multiplication operations on the system comprehensive error, cumulative error, and rate of change error to generate the proportional response, integral response, and derivative response.

[0103] The PID response calculation submodule generates proportional, integral, and derivative responses according to the following response function model:

[0104] ;

[0105] in, This represents the total adjustment of the output at time k. Represents the proportional gain coefficient. Represents the integral coefficient. Represents the differential coefficient, This represents the overall system error at time k. This represents the overall system error at time k-1. Represents the sampling period of the control system. Item corresponds to proportional response, The term corresponds to the integral response. The term corresponds to the differential response;

[0106] The multivariable command generation submodule linearly superimposes the proportional response, integral response, and derivative response to obtain the total adjustment amount. Based on the preset decoupling control strategy, the total adjustment amount is allocated to the stationary blade control channel and the turbine control channel to generate stationary blade opening adjustment command and turbine speed adjustment command, respectively.

[0107] The control decision module calculates the control response using a PID algorithm based on the difference between the comprehensive grinding state index and the ideal target, and decouples and generates instructions for different actuators.

[0108] The error feature extraction submodule first reads the preset ideal steady-state index. (Represents the optimal balance between energy efficiency and mass). Get the current moment. Comprehensive grinding condition index Calculate the overall error of the system. The module maintains a historical error queue of length 100 in memory.

[0109] Calculate cumulative error This is the sum of errors over the past 100 sampling periods. Assume the current cumulative error is... .

[0110] Error in calculating rate of change .

[0111] Let the error at the previous time be... Sampling period ,but .

[0112] The PID response calculation submodule calls the preset control parameters. The parameter tuning process is as follows:

[0113] proportional gain coefficient :

[0114] The gain is achieved by tuning using the critical proportional gain method, which produces constant-amplitude oscillations in the system. ,Pick , set as .

[0115] Integral coefficient : Set as It is used to eliminate steady-state error.

[0116] Differential coefficients : Set as It is used to suppress overshoot.

[0117] The PID response calculation submodule generates proportional, integral, and derivative responses according to the following response function model:

[0118] ;

[0119] in, The total adjustment amount represents the output at time k. Its physical meaning is the comprehensive correction force required by the system to eliminate the current state deviation. It is usually a dimensionless control signal value. This represents the proportional gain coefficient, used to adjust the system's response speed to the current error; This represents the integral coefficient, used to adjust the system's ability to eliminate steady-state errors; Represents the differential coefficient, used to reflect the trend of error change in order to provide anticipatory adjustment; The system comprehensive error at time k is obtained by subtracting the target value from the actual value. This represents the system synthesis error at time k-1. Represents the sampling period of the control system; This represents the cumulative error term from the initial time to the current time. The term representing the rate of change of error over time; Item corresponds to proportional response; The item corresponds to the integral response; The term corresponds to the differential response.

[0120] Substitute the numerical values ​​into the calculation:

[0121] Proportional term: ;

[0122] Integral term: ;

[0123] Differential term: ;

[0124] Total adjustment .

[0125] The multivariate instruction generation submodule receives a total adjustment amount of 139.73 and allocates it according to the decoupling strategy. The decoupling strategy is based on the principle of "granularity priority, load coordination":

[0126] If particle size deviation is the dominant factor (e.g., in this example, the particle size is coarse, resulting in a negative deviation), then priority is given to allocating to the turbine speed channel. Set the turbine allocation coefficient. static blade distribution coefficient .

[0127] Turbine speed adjustment range (Unit of speed increment).

[0128] Static blade opening adjustment range (In units of 0.1 degrees, i.e., 4.19 degrees).

[0129] Finally, the turbine speed adjustment command (target speed = current speed + 98 rpm) and the stationary blade opening adjustment command (target opening = current opening - 4.2 degrees, note that the direction is to close it to improve powder selection efficiency) are generated.

[0130] Please see Figure 1 and Figure 5 The execution feedback module generates stepper motor drive pulses based on the stationary blade opening adjustment command and sends them to the stationary blade actuator, and generates inverter control voltage based on the turbine speed adjustment command and sends it to the turbine blade drive motor.

[0131] The stepper motor drive pulse is specifically a digital electrical signal sequence with a target frequency and a count.

[0132] The inverter control voltage is specifically an analog voltage signal corresponding to the target output frequency;

[0133] The specific functions of the execution feedback module are as follows:

[0134] The pulse sequence mapping submodule parses the target angle change in the stationary blade opening adjustment command, calculates the required total number of pulses based on the stepper motor's step angle parameters, determines the pulse frequency based on the adjustment speed requirements, and generates stepper motor drive pulses.

[0135] The process of generating stepper motor drive pulses by the pulse sequence mapping submodule includes:

[0136] Obtain the target adjustment angle included in the stationary blade opening adjustment command;

[0137] Read the inherent step angle and transmission reduction ratio of the stepper motor in the stationary blade actuator;

[0138] The required pulse count value is obtained by dividing the target adjustment angle by the inherent step angle, multiplying it by the transmission reduction ratio, and rounding it down.

[0139] The pulse frequency change trajectory is determined based on the preset acceleration and deceleration curve model, and a stepper motor drive pulse including direction signal and clock signal is generated by combining the pulse count value.

[0140] The analog voltage conversion submodule analyzes the target speed value in the turbine speed adjustment command, calculates the corresponding output voltage amplitude based on the linear correspondence between the inverter's rated speed and the control voltage, and generates the inverter control voltage through the digital-to-analog conversion circuit.

[0141] The execution feedback module converts digital adjustment commands into physical electrical signals to drive the field equipment to operate.

[0142] The pulse sequence mapping submodule first parses the stationary blade opening adjustment command, aiming to reduce the guide vane angle by 4.2 degrees (moving in the closing direction). It then reads the stationary blade actuator parameters: the stepper motor's inherent step angle. Transmission mechanism reduction ratio Calculate the number of pulses required for the output shaft to rotate 1 degree: Pulses per degree. Calculate the total pulse count. One pulse. The module sets the pulse frequency. Based on the preset S-shaped acceleration / deceleration curve, the initial frequency is set to 200Hz, and the maximum frequency is set to 1000Hz. The processor generates a sequence of 117 pulses, sets the direction signal level to "low" (defined as turning off the direction), and sends it to the stepper driver via the GPIO port. The aforementioned stepper motor drive pulses refer to a series of discrete square wave signals sent by the controller to the stepper motor driver. Each pulse signal controls the stepper motor to rotate by a fixed step angle. By controlling the frequency of the pulses, the motor speed can be adjusted; by controlling the number of pulses, the angular displacement of the motor can be precisely controlled, thereby achieving precise digital control of the mechanical position.

[0143] The analog voltage conversion submodule parses the turbine speed adjustment command. Assuming the current turbine speed is 800 rpm and the target increment is +98 rpm, then the target speed... rpm. Read inverter parameters: rated speed rpm corresponds to the full-scale control voltage. V.

[0144] Calculate the linear correspondence:

[0145] .

[0146] Substitute the values: V.

[0147] The module uses a 12-bit DAC (digital-to-analog converter) chip to convert digital signals into 5.987V analog voltage signals, which are then sent to the inverter's AI (analog input) terminal.

[0148] Table 1 Parameter Mapping Table for Execution Feedback Module

[0149] Parameter name Physical meaning Set value / Calculated value unit Remark Stepper motor step angle 1.8 degree ( ) Two-phase hybrid stepper motor Reduction ratio 50 - worm gear reducer Target pulse count 117 indivual Corresponding to a 4.2 degree adjustment amount Turbine target speed 898 rpm Grader speed Inverter control voltage 5.987 V Corresponding to 0-10V range

[0150] As shown in Table 1, the system precisely maps abstract control commands to specific hardware drive parameters, ensuring the accuracy of the physical actuator's actions, thereby achieving closed-loop control of the prebaked anode grinding process; through the precise output of the aforementioned voltage and pulse, the classification efficiency of the cyclone dynamic separator is adjusted, thereby correcting the particle size and yield deviations detected in the previous step.

[0151] The above embodiments illustrate preferred embodiments of the present invention. Any equivalent adjustments to the technical solution based on software engineering methods are within the scope of protection, including but not limited to: implementing algorithm logic using different programming languages, refactoring functional modules into services, adjusting data interaction protocols, and optimizing resource scheduling strategies. Any implementation scheme derived from reasonable modifications to the data processing flow, service call chain, or system architecture layer without departing from the core technology of the present invention should be considered within the scope of protection defined by the claims of the present invention.

Claims

1. A prebaked anode grinding production control system based on feedback control, characterized in that, The system includes: The deviation calculation module acquires the acoustic vibration voltage signal, real-time mass flow rate value, and particle size distribution percentage value. It calculates the difference between the acoustic vibration voltage signal and the preset reference value to generate the filling state deviation, calculates the difference between the real-time mass flow rate value and the preset target output to generate the output deviation, and calculates the difference between the particle size distribution percentage value and the preset target range to generate the particle size deviation. The state modeling module constructs a state observation vector based on the filling state deviation, the production deviation, and the particle size deviation. It performs Kalman filtering on the state observation vector to output a state estimate, and then performs a weighted summation of the state estimate with a preset weight coefficient matrix to generate a comprehensive grinding state index. The control and decision module performs PID calculations on the comprehensive grinding state index to obtain proportional response, integral response and derivative response, and generates stationary blade opening adjustment command and turbine speed adjustment command based on the proportional response, integral response and derivative response. The execution feedback module generates a stepper motor drive pulse based on the stationary blade opening adjustment command and sends it to the stationary blade actuator, and generates a frequency converter control voltage based on the turbine speed adjustment command and sends it to the turbine blade drive motor.

2. The prebaked anode grinding production control system based on feedback control according to claim 1, characterized in that, The filling state deviation specifically refers to the voltage difference data of the material filling state in the mill relative to the baseline operating condition; the output deviation specifically refers to the quality difference data of the current production flow rate relative to the preset capacity target; the particle size deviation specifically refers to the percentage difference data of the fineness distribution of the finished powder relative to the qualified range; the comprehensive grinding state index includes the weighted and fused system operating efficiency evaluation value and quality state evaluation value; the stationary blade opening adjustment command includes the target opening value and stepping direction for the guide vanes of the cyclone dynamic separator; the turbine speed adjustment command includes the target speed setting value and adjustment range for the turbine classifier impeller; the stepper motor drive pulse is specifically a digital electrical signal sequence with a target frequency and count; and the frequency converter control voltage is specifically an analog voltage signal corresponding to the target output frequency.

3. The prebaked anode grinding production control system based on feedback control according to claim 2, characterized in that, The specific function of the deviation calculation module is as follows: The acoustic signal preprocessing submodule acquires the acoustic vibration voltage signal, uses a preset bandpass filter to remove the ambient background noise frequency band to retain the grinding sound characteristic frequency band, performs a sliding window root mean square operation on the filtered signal to extract the effective energy value, and performs differential calculation between the effective energy value and the preset no-load reference voltage to generate the filling state deviation. The flow data standardization submodule acquires real-time quality flow rate values, reads preset rated capacity settings, calculates the algebraic difference between the real-time quality flow rate values ​​and the rated capacity settings, and normalizes the calculated algebraic difference based on the range coefficient of the flow sensor to generate production deviation. The particle size distribution analysis submodule obtains the particle size distribution percentage value, identifies the cumulative distribution proportion within the qualified particle size range, calculates the difference between the cumulative distribution proportion and the preset target qualified rate, and generates the particle size deviation.

4. The prebaked anode grinding production control system based on feedback control according to claim 3, characterized in that, The specific functional implementation of the state modeling module is as follows: The vector space construction submodule arranges and combines the filling state deviation, the yield deviation, and the granularity deviation at the same sampling time in a preset order to construct the state observation vector at the current time. The filtering iterative estimation submodule calls the preset system state transition matrix and observation noise covariance matrix, performs prior state prediction for the current time based on the optimal state estimate of the previous time, and uses the state observation vector to correct and update the prior state prediction result, and outputs the state estimate. The weighted index synthesis submodule obtains the weight coefficient matrix corresponding to multiple dimensional components in the state estimate, performs a linear weighted summation operation on the state estimate, and maps the summation result to a preset scoring interval to generate a comprehensive grinding state index.

5. The prebaked anode grinding production control system based on feedback control according to claim 4, characterized in that, The specific functions of the control decision module are as follows: The error feature extraction submodule calculates the difference between the comprehensive grinding state index and the preset ideal steady-state index to obtain the comprehensive system error, and stores the comprehensive system error based on the time series to calculate the cumulative error and the rate of change error. The PID response calculation submodule obtains the preset proportional gain coefficient, integral time constant, and derivative time constant, respectively, and performs multiplication operations on the system comprehensive error, the cumulative error, and the rate of change error to generate the proportional response, integral response, and derivative response. The multivariable instruction generation submodule linearly superimposes the proportional response, the integral response, and the derivative response to obtain the total adjustment amount. Based on a preset decoupling control strategy, the total adjustment amount is allocated to the stationary blade control channel and the turbine control channel to generate stationary blade opening adjustment instructions and turbine speed adjustment instructions, respectively.

6. The prebaked anode grinding production control system based on feedback control according to claim 5, characterized in that, The specific function of the execution feedback module is as follows: The pulse sequence mapping submodule parses the target angle change in the stationary blade opening adjustment command, calculates the required total number of pulses based on the step angle parameters of the stepper motor, determines the pulse frequency based on the adjustment speed requirements, and generates stepper motor drive pulses. The analog voltage conversion submodule parses the target speed value in the turbine speed adjustment command, calculates the corresponding output voltage amplitude based on the linear correspondence between the rated speed of the frequency converter and the control voltage, and generates the frequency converter control voltage through the digital-to-analog conversion circuit.

7. The prebaked anode grinding production control system based on feedback control according to claim 3, characterized in that, The process by which the acoustic signal preprocessing submodule generates the filling state deviation includes: Set a time sliding window of length N, and collect N discrete acoustic vibration voltage signal sample points within the window; Calculate the arithmetic mean of the sum of squares of the N sample points, and then take the square root of the arithmetic mean to obtain the effective value; Obtain the reference voltage RMS value of the mill under no-load operation as pre-calibrated; Calculate the difference between the effective value and the effective value of the reference voltage, and define the difference as the fill state deviation.

8. The prebaked anode grinding production control system based on feedback control according to claim 4, characterized in that, The process by which the filtering iterative estimation submodule outputs the state estimate includes: The prior estimate for the current time step is calculated based on the posterior estimate of the previous time step and the system state transition matrix. Calculate the prediction error covariance matrix at the current time, and combine it with the observation noise covariance matrix to calculate the Kalman gain matrix; Calculate the residual between the state observation vector and the predicted observation corresponding to the prior estimate; The residuals are weighted using the Kalman gain matrix, and the weighted result is superimposed on the prior estimate to generate the state estimate.

9. The prebaked anode grinding production control system based on feedback control according to claim 5, characterized in that, The PID response calculation submodule generates the proportional response, the integral response, and the derivative response through a specific process that conforms to the following response function model: ； in, This represents the total adjustment of the output at time k. This represents the proportional gain coefficient. Represents the integral coefficient. Represents the differential coefficient, The system's overall error at time k is represented by the following. The system composite error at time k-1 is represented by [the error name]. Represents the sampling period of the control system. The item corresponds to the aforementioned proportional response. The term corresponds to the integral response, The term corresponds to the differential response.

10. The prebaked anode grinding production control system based on feedback control according to claim 6, characterized in that, The process by which the pulse sequence mapping submodule generates the stepper motor drive pulse includes: Obtain the target adjustment angle included in the stationary blade opening adjustment command; Read the inherent step angle and transmission reduction ratio of the stepper motor in the stationary blade actuator; The required pulse count value is obtained by dividing the target adjustment angle by the inherent step angle, multiplying it by the transmission reduction ratio, and rounding down. The trajectory of pulse frequency change is determined based on a preset acceleration / deceleration curve model, and the stepper motor drive pulse, including direction signal and clock signal, is generated by combining the pulse count value.

Images