Multi-stage centrifugal control system for heavy alkali filtration

Abstract

The present invention discloses a multi-stage centrifugal control system for heavy alkali filtration, which relates to the technical field of alkali production. The system comprises: a control record analysis module to obtain multiple groups of typical control parameters and corresponding typical filtration performance indicators. A simulation construction module establishes an initial centrifugal filtration simulation model, sets a simulation control series analysis to obtain initial simulation analysis results. A simulation correction module compares typical filtration performance indicators with initial simulation analysis results, and optimizes the initial simulation model. A slurry characteristic acquisition module obtains historical slurry sampling characteristics of a target scene, and determines the incoming material characteristic space. A parameter adjustment and calibration module performs random scatter point distribution in the incoming material characteristic space, obtains a calibrated incoming material characteristic set for simulation and parameter adjustment, and outputs a calibrated control parameter set. A control execution module performs multi-stage centrifugal control based on the calibrated control parameter set. Thereby achieving the technical effects of improving product stability, reducing parameter regulation costs, and improving monitoring feedback capabilities.

Description

Technical Field

[0001] The invention relates to the technical field of alkali production, in particular to a multi-stage centrifugal control system for heavy alkali filtration. Background Art

[0002] In the alkali production process, in order to reduce the salt and moisture in the product, the heavy alkali needs to be filtered. The filtering methods of heavy alkali include vacuum filtration, single-stage centrifugal filtration and multi-stage centrifugal filtration. Among them, multi-stage centrifugal filtration has gradually become the mainstream choice for heavy alkali filtration due to its advantages of high-efficiency separation and continuous operation. However, due to factors such as the variable characteristics of heavy alkali slurry and the complex control parameters of multi-stage centrifugal filtration, the existing multi-stage centrifugal control technology for heavy alkali filtration has technical problems such as unstable product quality, difficulty in parameter regulation, and poor monitoring and feedback capabilities. Summary of the invention

[0003] The present invention provides a multi-stage centrifugal control system for heavy alkali filtration to solve the technical problems of unstable product quality, difficulty in parameter control, and poor monitoring and feedback capabilities in the prior art, thereby achieving the technical effects of improving product stability, reducing parameter control costs, and improving monitoring and feedback capabilities.

[0004] The multi-stage centrifugal control system for heavy alkali filtration provided by the present invention comprises:

[0005] A control record analysis module is used to obtain multiple groups of typical control parameters and corresponding typical filtration performance indicators based on historical centrifugal control records.

[0006] A simulation construction module is used to establish an initial centrifugal filtration simulation model, set a simulation control series based on multiple groups of typical control parameters, call the initial centrifugal filtration simulation model for simulation analysis, and obtain initial simulation analysis results.

[0007] A simulation correction module is used to compare the typical filtration performance index with the initial simulation analysis result, obtain simulation difference information, and optimize and correct the initial centrifugal filtration simulation model with minimization of simulation difference as the optimization goal to obtain a characteristic centrifugal filtration simulation model.

[0008] A slurry characteristic acquisition module is used to obtain historical slurry sampling characteristics of a target scene and determine the incoming material characteristic space, wherein the slurry sampling characteristics include slurry particle size characteristics, slurry solid content, and slurry temperature.

[0009] A parameter adjustment and calibration module is used to perform random scatter point distribution according to the incoming material characteristic space, obtain a calibration incoming material characteristic set, activate the characteristic centrifugal filtration simulation model to simulate and optimize the parameters of the calibrated incoming material characteristic set, output multiple groups of calibration control parameters, and obtain a calibration control parameter set.

[0010] A control execution module is used to perform multi-stage centrifugal control of heavy alkali slurry based on the calibration control parameter set.

[0011] In a feasible implementation, the initial centrifugal filtration simulation model is established, and a simulation control series is set based on multiple groups of typical control parameters, the initial centrifugal filtration simulation model is called for simulation analysis, and the initial simulation analysis results are obtained. The execution steps include:

[0012] Establish a centrifugal filtration numerical model, wherein the numerical model includes a centrifugal force field model, a sedimentation dynamics model, and a control conversion model; obtain the structural parameters and performance parameters of the target centrifuge, and construct a simplified centrifugal filtration model based on simulation software, wherein the simplified centrifugal filtration model at least includes a multi-stage drum, a material flow path, and a material discharge path; fit the centrifugal filtration numerical model and the simplified centrifugal filtration model to construct the initial centrifugal filtration simulation model, and use multiple groups of typical control parameters as simulation control parameters, and the slurry characteristics corresponding to the multiple groups of typical control parameters as slurry simulation characteristics, and set the simulation control series; configure the initial centrifugal filtration simulation model based on the simulation control series to perform simulation analysis, record simulation analysis data, and store them as the initial simulation analysis results.

[0013] In a feasible implementation, the typical filtration performance index is compared with the initial simulation analysis result to obtain simulation difference information, and the initial centrifugal filtration simulation model is tuned and corrected with minimization of simulation difference as the optimization goal to obtain a characteristic centrifugal filtration simulation model, and the execution steps include:

[0014] Based on the filtering performance requirements of the target scenario, performance indicator weights are allocated and a difference information evaluation function is constructed; the control coefficient of the numerical model in the initial centrifugal filtration simulation model is used as an independent variable to construct a control coefficient space, and a control particle group is randomly generated; the simulation control series is used as multiple target dimensions, and the control particle group is iteratively optimized in combination with the difference information evaluation function to obtain the characteristic centrifugal filtration simulation model.

[0015] In a feasible implementation, the multi-stage centrifugal control of heavy alkali slurry based on the calibration control parameter set includes the following steps:

[0016] Acquire the incoming sample characteristics of heavy alkali slurry in real time, and acquire the real-time slurry characteristics; map the real-time slurry characteristics to the incoming material characteristic set for similarity evaluation, and acquire the slurry characteristic similarity sequence; based on the slurry characteristic similarity sequence, perform parameter call in the calibration control parameter set, and acquire the multi-stage centrifugal control parameters to perform multi-stage centrifugal control of the heavy alkali slurry.

[0017] In a feasible implementation, the multi-stage centrifugal control of the heavy alkali slurry based on the calibration control parameter set further includes:

[0018] The calibrated incoming material characteristic set and the calibrated control parameter set are used as sample data to construct a decision training set and a decision verification set; a centrifugal filtration baseline model based on a neural network is constructed, and supervised training is performed through the decision training set, and the centrifugal filtration baseline model is verified with the decision verification set; the real-time slurry characteristics are input into the centrifugal filtration baseline model to make a centrifugal filtration decision, and multi-stage centrifugal control of the heavy alkali slurry is performed according to the decision result.

[0019] In a feasible implementation, the execution steps of the system further include:

[0020] Vibration data of the target centrifuge is collected and the vibration spectrum is extracted; operation stability analysis is performed based on the vibration spectrum, and risk disposal is performed based on the stability analysis result.

[0021] In a feasible implementation, the execution steps of the system further include:

[0022] Analyze the historical centrifugal control records to obtain the filter cake crushing thickness and the filter cake thickness of the vibration car, and calculate and determine the filter cake accumulation limit based on the pushing interval, wherein the filter cake accumulation limit includes the minimum filtering accumulation rate and the maximum filter cake accumulation rate; monitor the filter cake thickness growth rate in real time to obtain the real-time filter cake accumulation rate; and perform filtration load judgment and instant filtration load control based on the filter cake accumulation limit.

[0023] In a feasible implementation, the step of performing random scatter point distribution according to the incoming material characteristic space to obtain a calibration incoming material characteristic set further includes:

[0024] Based on the historical slurry sampling characteristics, the probability distribution of the incoming material characteristic space is analyzed and determined; according to the probability distribution, random scattered point allocation is performed to obtain the calibrated incoming material characteristic set.

[0025] The invention discloses a multi-stage centrifugal control system for heavy alkali filtration, comprising: a control record analysis module: used for obtaining multiple groups of typical control parameters and corresponding typical filtration performance indicators based on historical centrifugal control records; a simulation construction module: used for establishing an initial centrifugal filtration simulation model, and setting a simulation control series based on multiple groups of typical control parameters, calling the initial centrifugal filtration simulation model for simulation analysis, and obtaining initial simulation analysis results; a simulation correction module: used for comparing typical filtration performance indicators with initial simulation analysis results, obtaining simulation difference information, and performing initial centrifugal filtration simulation simulation with the minimization of simulation difference as the optimization goal. The optimization and correction of the type is used to obtain the characteristic centrifugal filtration simulation model; the slurry characteristic acquisition module is used to obtain the historical slurry sampling characteristics of the target scene and determine the incoming material characteristic space, wherein the slurry sampling characteristics include the slurry particle size characteristics, the slurry solid content, and the slurry temperature; the parameter adjustment and calibration module is used to perform random scattered point distribution according to the incoming material characteristic space, obtain the calibrated incoming material characteristic set, activate the characteristic centrifugal filtration simulation model to simulate and optimize the calibrated incoming material characteristic set, output multiple groups of calibration control parameters, and obtain the calibration control parameter set; the control execution module is used to perform multi-stage centrifugal control of heavy alkali slurry based on the calibration control parameter set. The multi-stage centrifugal control system for heavy alkali filtration disclosed in the present invention solves the technical problems of unstable product quality, difficult parameter regulation, and poor monitoring feedback capability, and achieves the technical effects of improving product stability, reducing parameter regulation costs, and improving monitoring feedback capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic structural diagram of a multi-stage centrifugal control system for heavy alkali filtration according to the present invention.

[0027] Figure 2 It is a schematic diagram of the process of obtaining initial simulation analysis results in the multi-stage centrifugal control system for heavy alkali filtration of the present invention.

[0028] Explanation of the accompanying drawings: control record analysis module 11, simulation construction module 12, simulation correction module 13, slurry characteristic acquisition module 14, parameter adjustment and calibration module 15, control execution module 16. DETAILED DESCRIPTION

[0029] The technical solution provided in the embodiments of the present invention is to solve the technical problems of unstable product quality, difficulty in parameter control, and poor monitoring and feedback capability in the prior art. The overall idea adopted is as follows:

[0030] Firstly, the control record analysis module obtains multiple groups of typical control parameters and corresponding typical filtration performance indicators based on historical centrifugal control records.

[0031] Next, the simulation construction module establishes an initial centrifugal filtration simulation model, sets a simulation control series based on multiple groups of typical control parameters, calls the initial centrifugal filtration simulation model for simulation analysis, and obtains initial simulation analysis results.

[0032] Then, the simulation correction module compares the typical filtration performance indicators with the initial simulation analysis results, obtains the simulation difference information, and takes minimizing the simulation difference as the optimization goal to tune and correct the initial centrifugal filtration simulation model to obtain a characteristic centrifugal filtration simulation model.

[0033] Subsequently, the slurry characteristic acquisition module obtains the historical slurry sampling characteristics of the target scene and determines the incoming material characteristic space, wherein the slurry sampling characteristics include slurry particle size characteristics, slurry solid content, and slurry temperature.

[0034] Next, the parameter adjustment and calibration module performs random scatter distribution according to the incoming material characteristic space to obtain a calibration incoming material characteristic set, activates the characteristic centrifugal filtration simulation model to simulate and optimize the parameters of the calibration incoming material characteristic set, outputs multiple groups of calibration control parameters, and obtains a calibration control parameter set.

[0035] Finally, the control execution module performs multi-stage centrifugal control of the heavy alkali slurry based on the calibrated control parameter set.

[0036] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0037] Example

[0038] Figure 1 It is a schematic structural diagram of a multi-stage centrifugal control system for heavy alkali filtration of the present invention, wherein the system includes.

[0039] The control record analysis module 11 is used to obtain multiple groups of typical control parameters and corresponding typical filtration performance indicators based on historical centrifugal control records.

[0040] Specifically, a data acquisition tool is used to export historical records from the control system of the centrifuge to obtain historical centrifugal control records, which refer to local centrifugal control records of the multi-stage centrifuge in the target centrifugal filtration scenario, including centrifuge operating parameters: such as centrifuge speed, feed speed, centrifugal force, etc. Centrifuge filtration performance indicators: such as centrifuge filtration efficiency, product quality, product moisture content, etc. Centrifuge material characteristics: such as feed particle size, density, moisture content, etc.

[0041] Specifically, typical control parameters and corresponding typical filtration performance indicators refer to the most representative control parameters and filtration performance indicators in historical centrifugal control records. Exemplarily, a data analysis tool is used to perform statistical analysis on historical data to extract the top groups of control parameters and corresponding filtration performance indicators that appear most frequently in historical centrifugal control records, and the output is multiple groups of typical control parameters and corresponding typical filtration performance indicators.

[0042] The simulation construction module 12 is used to establish an initial centrifugal filtration simulation model, set a simulation control series based on multiple groups of typical control parameters, call the initial centrifugal filtration simulation model to perform simulation analysis, and obtain initial simulation analysis results.

[0043] Optionally, a physics-based and data-driven initial centrifugal filtration simulation model is constructed so that the centrifugal filtration simulation model has good basic accuracy and reliability.

[0044] Specifically, multiple groups of typical control parameters are selected, different simulation control schemes are set, and a simulation control series is formed, wherein each group of typical control parameters corresponds to a simulation control scheme. The simulation control series is set based on multiple groups of typical control parameters for simulation analysis, and the actual simulation accuracy and simulation performance of the obtained initial centrifugal filtration simulation model can be verified.

[0045] Optionally, during the simulation process, key data such as filtration efficiency, separation purity, processing volume, energy consumption, etc. are collected, sorted and analyzed to ensure the integrity and consistency of the data, and multiple groups of simulated key data corresponding to the simulation control series are output as initial simulation analysis results.

[0046] In some embodiments, Figure 2 As shown, an initial centrifugal filtration simulation model is established, and a simulation control series is set based on multiple groups of typical control parameters, and the initial centrifugal filtration simulation model is called for simulation analysis. The execution steps of obtaining the initial simulation analysis results include:

[0047] A centrifugal filtration numerical model is established, wherein the numerical model includes a centrifugal force field model, a sedimentation dynamics model, and a control conversion model.

[0048] The structural parameters and performance parameters of the target centrifuge are obtained, and a simplified centrifugal filtration model is constructed based on simulation software, wherein the simplified centrifugal filtration model at least includes a multi-stage drum, a material flow path, and a material discharge path.

[0049] The centrifugal filtration numerical model and the centrifugal filtration simplified model are fitted to construct the initial centrifugal filtration simulation model, and the simulation control series is set by using multiple groups of typical control parameters as simulation control parameters and the slurry characteristics corresponding to the multiple groups of typical control parameters as slurry simulation characteristics.

[0050] The initial centrifugal filtration simulation model is configured based on the simulation control series to perform simulation analysis, and simulation analysis data is recorded and stored as the initial simulation analysis result.

[0051] Specifically, firstly, a centrifugal filtration numerical model is established, which includes multiple sub-models: such as a centrifugal force field model, a sedimentation dynamics model, and a control conversion model.

[0052] Specifically, the centrifugal force field model describes the centrifugal force field in centrifugal filtration. Centrifugal force is the main driving force in centrifugal filtration, which determines the direction and speed of movement of solid particles in liquid. For example, the slurry as feed is subjected to the centripetal force and centrifugal inertial force in centrifugal filtration, and the two are equal and opposite:

[0053] .

[0054] Among them, F0 is the centripetal force, F C is the centrifugal inertial force, m is the mass of the grain, ω is the angular velocity of rotation, and r is the rotation radius of the grain.

[0055] Specifically, the sedimentation kinetics model describes the sedimentation process of crystals in a centrifugal field. The speed and efficiency of the sedimentation process directly affect the performance of centrifugal filtration. The sedimentation kinetics model describes the sedimentation characteristics of crystals in a slurry by considering the interaction between solid particles and liquid and the interaction between particles.

[0056] For example, assuming that the particles are uniformly distributed and the interaction between the particles and the turbulent effect of the fluid are ignored, the sedimentation of the grains in the slurry belongs to the case of small particles and low Reynolds number. Since centrifugal filtration applies the treated water-containing material to the porous screen surface of the drum through centrifugal action, it can be regarded as the grains being affected by a weight field in the same direction as the centripetal acceleration generated by the centripetal force, and settling toward the centrifugal rotation axis relative to the liquid phase in the slurry.

[0057] The sedimentation velocity can be calculated using Stokes' law, which is expressed as:

[0058] .

[0059] in, v is the settling velocity, d is the particle diameter, ρ p and ρ f are the densities of the particle and fluid, respectively, g is the acceleration due to gravity, μ is the dynamic viscosity of the fluid.

[0060] Since in the centrifugal filtration process, the crystals are mainly subjected to the centripetal acceleration provided by the drum and the centrifugal acceleration generated by the rotation, and the two are equal and opposite, the gravitational acceleration g in Stokes' law is replaced by the centrifugal acceleration a accordingly. c The sedimentation velocity model suitable for centrifugal filtration is obtained, which is expressed as:

[0061] .

[0062] And, the centrifugal acceleration:

[0063] .

[0064] Specifically, the control conversion model is designed according to the actual control strategy and system characteristics to realize the conversion between the ideal control parameters and the control parameters applicable to the actual multi-stage centrifugal filter. Exemplarily, the control conversion model based on the PID controller is as follows:

[0065] .

[0066] in, It is the positive output of the control transformation model, that is, the control signal. is the error signal, defined as the difference between the target value and the actual value. K p , K i and K d are the parameters of the control conversion model, representing the proportional gain, integral gain and differential gain respectively, and are used to adjust and optimize the performance of the control conversion model. is the integral of the error signal and represents the accumulated deviation of the system. is the derivative of the error signal, representing the rate of change of the system deviation.

[0067] Furthermore, necessary structural parameters and performance parameters are obtained from the target centrifuge, illustratively including structural parameters (such as the size, shape, material, opening size and shape, porosity, etc. of the multi-stage drum) and performance parameters (speed range, maximum processing capacity, operating temperature, pusher interval range, etc.), and then a simplified centrifugal filtration model is constructed using simulation software such as MATLAB, ANSYS, COMSOL, etc.

[0068] Specifically, the simplified model includes at least a multi-stage drum, a material flow path, a material discharge path, etc. Among them, the drum is the main component of the centrifuge, which generates centrifugal force by high-speed rotation. The multi-stage drum means that each drum can perform different processing steps, thereby improving the processing efficiency. The material flow path is the flow path of the material in the centrifuge, which determines the material processing process and time. The material discharge path is the discharge path of the material in the centrifuge, which determines the material discharge method and time.

[0069] Specifically, the numerical model is fitted with the simplified model to construct an initial centrifugal filtration simulation model. In other words, the movement and sedimentation process of the slurry in the simplified model is described and defined by the numerical model, so as to obtain the initial centrifugal filtration simulation model. Then, different simulation control schemes are set according to multiple groups of typical control parameters to form a simulation control series. The simulation control series includes a series of simulation control parameters and corresponding slurry characteristics. Exemplary, including the speed of the centrifuge, the flow rate of the material, the pushing cycle, etc. The corresponding slurry characteristics include the density, viscosity, particle size distribution, water content, temperature, etc. of the slurry. By changing these parameters, different operating conditions and material characteristics can be simulated, thereby affecting the centrifugal filtration performance.

[0070] Furthermore, the initial centrifugal filtration simulation model is used to perform simulation analysis on each group of simulation control schemes. During the simulation process, key data such as the filtration efficiency, separation purity, processing capacity, energy consumption, etc. of the centrifuge are recorded and stored as initial simulation result data, which is used to reflect and optimize the simulation performance of the initial centrifugal filtration simulation model.

[0071] The simulation correction module 13 is used to compare the typical filtration performance index with the initial simulation analysis result, obtain simulation difference information, and optimize and correct the initial centrifugal filtration simulation model with minimization of simulation difference as the optimization goal to obtain a characteristic centrifugal filtration simulation model.

[0072] Specifically, by comparing typical filtration performance indicators with initial simulation analysis results, the differences between the initial model and the actual situation can be found. After obtaining the simulation difference information, the model is tuned and corrected with the goal of minimizing the simulation difference, which can ensure that the centrifugal filtration simulation model more accurately reflects the performance in the actual filtration process. The characteristic centrifugal filtration simulation model finally obtained can better predict filtration performance, improve filtration efficiency and effect, and reduce uncertainty and risk in the test and production process.

[0073] Optionally, the obtained simulation difference information refers to the difference between the simulation result and the actual index, including average difference, maximum difference, difference distribution, etc. By adjusting the model parameters, the simulation difference is minimized, so that the simulation result is as close as possible to the actual filtering performance index.

[0074] In some embodiments, the typical filtration performance index is compared with the initial simulation analysis result to obtain simulation difference information, and the initial centrifugal filtration simulation model is tuned and corrected with minimization of the simulation difference as the optimization goal to obtain the characteristic centrifugal filtration simulation model. The execution steps include:

[0075] Based on the filtering performance requirements of the target scenario, performance indicator weights are assigned and a difference information evaluation function is constructed.

[0076] The control coefficient of the numerical model in the initial centrifugal filtration simulation model is used as an independent variable to construct a control coefficient space, and randomly generate a control particle group.

[0077] The simulation control series is used as multiple target dimensions, and the difference information evaluation function is combined to perform iterative optimization of the regulated particle group to obtain the characteristic centrifugal filtration simulation model.

[0078] Specifically, in practical applications, the performance requirements of centrifugal filtration vary depending on the target scenario. For example, some scenarios may focus more on separation efficiency, while other scenarios may focus more on energy consumption or processing speed. Therefore, different weights are assigned to each performance indicator according to the requirements of the target scenario, and the difference information evaluation function is constructed accordingly.

[0079] Specifically, first, determine the performance indicators of centrifugal filtration that the target scenario is concerned about, including separation efficiency, sedimentation velocity, energy consumption, processing speed, etc. Then, assign weights to each performance indicator according to the needs of the target scenario. The value of the weight should reflect the importance of the indicator. If the target scenario is more concerned with separation efficiency, a higher weight is assigned to separation efficiency. Finally, based on the assigned multiple weights, a difference information evaluation function is constructed. The difference information evaluation function calculates a total evaluation value based on the simulation differences and weights of each indicator to compare and select different models or parameters.

[0080] Furthermore, the independent variables in the optimization of the initial centrifugal filtration simulation model are determined. Specifically, the independent variables are the control coefficients of the numerical model in the centrifugal filtration simulation model, including the parameters of the centrifugal force model, the parameters of the sedimentation dynamics model, the parameters of the PID controller, etc. Then, a control coefficient space is constructed, which includes the values ​​of all possible control coefficients. Finally, a control particle swarm is randomly generated in the control coefficient space. Each particle in the control particle swarm represents a possible combination of control coefficients, and the number of particles is set according to the requirements of the target scene and the computing resources. Through the above steps, the initialization of iterative optimization is achieved based on the particle swarm optimization method.

[0081] Specifically, the control coefficient of each particle is applied to the initial centrifugal filtration simulation model, and multiple groups of simulation control parameters in the simulation control series are simulated to obtain a series of simulation results. Then, a series of simulation results and corresponding target results are compared, and the values ​​of multiple difference information evaluation functions are calculated. Next, the statistical analysis indicators (such as mean, standard deviation, median, quartile, etc.) of the values ​​of multiple difference information evaluation functions are calculated as the evaluation value of each particle, and then the speed and position of the particle are updated according to the individual best position of the particle (i.e., the position of the best evaluation value obtained in the history of the particle) and the global best position (i.e., the position of the best evaluation value obtained in the history of all particles). Finally, the above evaluation and update steps are repeated until a certain termination condition is met, and the best individual particle is output as the optimal control coefficient. The termination condition includes reaching the maximum number of iterations or the global best evaluation value has reached a preset threshold.

[0082] Specifically, based on the optimal control coefficient, the control coefficient of the initial centrifugal filtration simulation model is updated to obtain a characteristic centrifugal filtration simulation model. The characteristic centrifugal filtration simulation model can more accurately simulate the centrifugal filtration process and meet the needs of the target scene.

[0083] The slurry characteristic acquisition module 14 is used to obtain the historical slurry sampling characteristics of the target scene and determine the incoming material characteristic space, wherein the slurry sampling characteristics include slurry particle size characteristics, slurry solid content, and slurry temperature.

[0084] Specifically, historical slurry sampling characteristics are obtained from the target scenario, including collecting data from past experiments or operations, or obtaining relevant data from literature. Then, based on NLP, keyword extraction, named entity recognition (NER), topic modeling and other technologies, slurry sampling characteristics are extracted, including slurry particle size characteristics, slurry solid content, slurry temperature, etc. The historical slurry sampling characteristics of the target scenario help to understand the various incoming material conditions that may be encountered during the centrifugal filtration process, so as to make more accurate and effective centrifugal control decisions.

[0085] Among them, the slurry particle size characteristics include the average particle size, particle size distribution, maximum and minimum particle size, etc. The solid content of the slurry refers to the content of solid matter in the slurry, expressed as a weight percentage. The temperature of the slurry can affect its viscosity, and thus affect the efficiency of solid-liquid separation.

[0086] Optionally, the slurry sampling characteristics also include slurry viscosity. The viscosity of the slurry will affect the efficiency of centrifugal filtration because liquids with higher viscosity have poorer fluidity.

[0087] Specifically, a multidimensional characteristic space is constructed based on the historical slurry sampling characteristics to form an incoming material characteristic space, wherein each dimension of the incoming material characteristic space represents a slurry characteristic parameter. The multidimensional characteristic space describes the possible range of slurry characteristics. Exemplarily, the incoming material characteristic space should completely include the historical slurry sampling characteristics and be slightly larger than the range of the historical slurry sampling characteristics. The specific range of the incoming material characteristic space is determined based on the filtering task requirements of the target scenario.

[0088] The parameter adjustment and calibration module 15 is used to perform random scatter point distribution according to the incoming material characteristic space, obtain a calibration incoming material characteristic set, activate the characteristic centrifugal filtration simulation model to simulate and optimize the parameters of the calibration incoming material characteristic set, output multiple groups of calibration control parameters, and obtain a calibration control parameter set.

[0089] Specifically, random scattered points are generated in the feature space, and the generated scattered points are used as the calibration incoming material feature set, wherein each scattered point represents a possible incoming material feature combination. Exemplarily, the scattered points are generated using techniques such as Latin Hypercube Sampling (LHS) and Monte Carlo method.

[0090] Furthermore, the characteristic centrifugal filtration simulation model is used to simulate the calibrated incoming material characteristic set, including simulating each scattered point separately, simulating the performance and output results in the actual production process, and adjusting the control parameters according to the simulation results. Specifically, the output results after each simulation are recorded, including production efficiency, quality indicators, energy consumption, etc.

[0091] Specifically, multiple groups of calibration control parameters are extracted based on the simulation results, and the calibration control parameters after all simulations and parameter tuning are combined to form a calibration control parameter set. The calibration control parameter set defines the optimal control parameter set corresponding to multiple groups of incoming material conditions under the calibration incoming material characteristic set, which is used to guide the production process and ensure stability and optimization performance in the production process.

[0092] Through the above steps, the characteristic centrifugal filtration simulation model is used to simulate the incoming material characteristic set of random scattered point calibration, and the parameters are tuned according to the simulation results. It is possible to effectively output multiple sets of calibration control parameters and construct a calibration control parameter set, which helps to quickly call and optimize the process control and equipment adjustment in the actual production process.

[0093] Optionally, the size of the calibration incoming material characteristic set and the calibration control parameter set is determined based on the control needs and computing power level of the target scenario. For high-precision control scenarios, higher control accuracy and finer parameter tuning are required. The more calibration points of the calibration incoming material characteristic set and the corresponding calibration control parameter set, the more computing resources are required. Through the above steps, under reasonable computing power levels and control requirements, the appropriate size of the calibration incoming material characteristic set and calibration control parameter set is determined to ensure the effectiveness and efficiency of simulation and parameter tuning.

[0094] In some embodiments, random scatter point distribution is performed according to the incoming material characteristic space to obtain a calibration incoming material characteristic set, and the execution step further includes:

[0095] Based on the historical slurry sampling characteristics, the probability distribution of the incoming material characteristic space is analyzed and determined.

[0096] According to the probability distribution, random scattered point allocation is performed to obtain the calibrated incoming material characteristic set.

[0097] Specifically, random scattered points are distributed according to the incoming material characteristic space and the calibrated incoming material characteristic set is obtained. First, the sampling data of the historical slurry is collected and analyzed, and the distribution of these characteristic parameters is represented by statistical analysis methods (such as histograms and probability density functions) to determine the probability distribution of each characteristic parameter (such as density, viscosity, particle size, etc.). Then, the determined probability distribution is used to generate random scattered points to ensure that these scattered points can represent the actual distribution in the incoming material characteristic space.

[0098] Through the above steps, random scattered points are generated based on probability distribution to obtain a calibration incoming material characteristic set that represents the actual situation. This process ensures that the basic data for simulation and parameter tuning are representative, thereby improving the availability and efficiency of the obtained calibration control parameter set.

[0099] The control execution module 16 is used to perform multi-stage centrifugal control of the heavy alkali slurry based on the calibration control parameter set.

[0100] In some embodiments, the multi-stage centrifugal control of the heavy alkali slurry is performed based on the calibration control parameter set, and the execution steps include:

[0101] Obtain the incoming sample characteristics of heavy alkali slurry in real time and obtain real-time slurry characteristics.

[0102] The real-time slurry characteristics are mapped to the incoming material characteristic set for similarity evaluation to obtain a slurry characteristic similarity sequence.

[0103] Based on the slurry characteristic similarity sequence, parameter call is performed in the calibration control parameter set to obtain multi-stage centrifugal control parameters for multi-stage centrifugal control of heavy alkali slurry.

[0104] Specifically, first, use sensors and online detection equipment to collect the characteristic parameters of heavy alkali slurry in real time, such as density, viscosity, particle size distribution, etc., to obtain real-time slurry characteristics. Then, map the real-time slurry characteristics to the multidimensional space where the calibration incoming material characteristic set is located, perform similarity evaluation based on Euclidean distance, cosine similarity, etc., calculate the similarity between each calibration incoming material characteristic and the real-time slurry characteristic, and form a slurry characteristic similarity sequence. Finally, according to the slurry characteristic similarity sequence, call the calibration control parameter combination corresponding to the calibration incoming material characteristic that is most similar to the real-time slurry characteristic from the calibration control parameter set as the multi-stage centrifugal control parameter, and use the obtained multi-stage centrifugal control parameters to adjust the target centrifuge operating parameters to perform multi-stage centrifugal control on the heavy alkali slurry.

[0105] In some implementations, the multi-stage centrifugal control of the heavy alkali slurry is performed based on the calibration control parameter set, and the execution step further includes:

[0106] The calibrated incoming material characteristic set and the calibrated control parameter set are used as sample data to construct a decision training set and a decision verification set.

[0107] A centrifugal filtration baseline model based on a neural network is constructed, and supervised training is performed through the decision training set, and the centrifugal filtration baseline model is verified with the decision verification set.

[0108] The real-time slurry characteristics are input into the centrifugal filtration baseline model to make a centrifugal filtration decision, and multi-stage centrifugal control of the heavy alkali slurry is performed according to the decision result.

[0109] Specifically, neural networks are used to make decisions to control the multi-stage centrifugation process of heavy alkali slurry. First, a calibrated incoming material characteristic set and a calibrated control parameter set are used as sample data to construct. Among them, the decision training set is used to train the neural network model, and the decision verification set is used to verify the performance of the model. Then, a centrifugal filtration baseline model based on a neural network is constructed, and the decision training set is used to supervise the training of the neural network model. During the training process, the model learns how to adjust the control parameters according to the slurry characteristics through the decision training set and the decision verification set.

[0110] Exemplarily, the centrifugal filtration baseline model is constructed based on LSTM, multi-layer perceptron, etc., with an input layer, a hidden layer, and an output layer. Among them, the input layer is used to receive slurry characteristic parameters, such as particle size, humidity, etc. The hidden layer includes multiple layers, and each layer has a number of neurons. The output layer is used to output control parameters, such as centrifugal speed, time, etc.

[0111] Furthermore, the real-time slurry characteristics are input into the centrifugal filtration baseline model to make centrifugal filtration decisions. And according to the decision results, multi-stage centrifugal control of the heavy alkali slurry is performed. Through the above steps, the automatic fine adjustment of the control parameters is achieved to adapt to the real-time slurry characteristics, thereby further improving the efficiency and effect of centrifugal filtration.

[0112] Furthermore, the execution steps of the system also include:

[0113] Collect vibration data of the target centrifuge and extract the vibration spectrum.

[0114] An operation stability analysis is performed based on the vibration spectrum, and risk management is performed based on the stability analysis result.

[0115] Optionally, vibration sensors are installed at key parts of the centrifuge (such as bearing seats, transmission systems, etc.), and vibration signals are collected in real time based on preset collection parameters. Preferably, the vibration sensor has a high sampling rate to capture the high-frequency components of the vibration signal. Then, the collected vibration data is preprocessed, including denoising, filtering and other steps to remove unnecessary noise and interference, and then the preprocessed vibration signal is Fourier transformed to convert the time domain signal into a frequency domain signal, extract the vibration spectrum, analyze the main frequency components in the spectrum, and identify the vibration amplitude corresponding to each frequency component.

[0116] Specifically, the operation stability analysis is performed based on the vibration spectrum. First, key features such as main frequency, harmonics, sidebands, resonance frequency, etc. are extracted from the vibration spectrum. These key features can reflect the operation status and health of the centrifuge. Then, based on the historical vibration spectrum features and historical operation data, an operation stability analysis model is constructed using algorithms such as support vector machine (SVM) and random forest, and the key features of the vibration spectrum are input for classification prediction, the operation stability of the centrifuge is evaluated, and the stability analysis results are obtained.

[0117] Furthermore, based on the results of the stability analysis, the risk level of the centrifuge is evaluated to determine whether its operating status is within the safe range. The risk assessment indicators include excessive vibration amplitude, abnormal frequency, etc. Finally, based on the risk assessment results, a preventive maintenance plan is formulated, including equipment inspection, component replacement, lubrication and other measures. In addition, for equipment in a high-risk state, emergency shutdown measures are taken in a timely manner to avoid serious failures.

[0118] Furthermore, the execution steps of the system also include:

[0119] The historical centrifugal control records are analyzed to obtain the filter cake crushing thickness and the filter cake thickness of the shaking car, and based on the pushing interval, the filter cake accumulation limit is calculated and determined, wherein the filter cake accumulation limit includes the minimum filtering accumulation rate and the maximum filter cake accumulation rate.

[0120] Monitor the filter cake thickness growth rate in real time and obtain the real-time filter cake accumulation rate.

[0121] The filter load is judged and the filter load is controlled in real time based on the filter cake accumulation limit.

[0122] Specifically, first, analyze the historical centrifuge control records to obtain the filter cake crushing thickness and the shaking car filter cake thickness, among which, the filter cake crushing thickness refers to the minimum critical thickness at which the filter cake is broken during the pushing process due to being too thin and then accumulated on the edge of the drum, and the shaking car filter cake thickness refers to the maximum critical thickness at which the target centrifuge or drum produces abnormal vibration after the filter cake passes.

[0123] Specifically, based on the pushing interval, the filter cake accumulation limit is calculated and determined. The filter cake accumulation limit includes the minimum filter cake accumulation rate based on the filter cake crushing thickness and the maximum filter cake accumulation rate based on the filter cake thickness of the shaking car, which are used to judge the accumulation of filter cake.

[0124] Furthermore, the thickness of the filter cake is monitored in real time by installing a sensor or other measuring device inside the target centrifuge. The thickness growth rate of the filter cake, that is, the rate of change of the filter cake thickness, is analyzed and calculated, and stored as the real-time filter cake accumulation rate. Thus, the formation speed of the filter cake can be understood. Next, based on the filter cake accumulation limit and the real-time filter cake accumulation rate, the load level of the current target centrifuge is determined. If the real-time filter cake accumulation rate is less than or equal to the minimum accumulation rate, the feed amount is appropriately increased to increase the load level of the centrifuge. Conversely, if the real-time filter cake accumulation rate is greater than or equal to the maximum filter cake accumulation rate, the amount of crystal slurry fed should be reduced to reduce the load of the centrifuge.

[0125] Through the above steps, the feed amount can be dynamically adjusted based on real-time data, which can make the centrifugal filtration process more stable and efficient, while avoiding damage to the machine or loss of efficiency due to abnormal cake thickness.

[0126] In summary, the multi-stage centrifugal control system for heavy alkali filtration provided by the present invention has the following technical effects:

[0127] Based on the historical centrifugal control records, the control record analysis module obtains multiple groups of typical control parameters and corresponding typical filtration performance indicators; the simulation construction module establishes an initial centrifugal filtration simulation model, and sets the simulation control series based on multiple groups of typical control parameters, calls the initial centrifugal filtration simulation model for simulation analysis, and obtains the initial simulation analysis results; the simulation correction module compares the typical filtration performance indicators with the initial simulation analysis results, obtains the simulation difference information, and takes the minimization of the simulation difference as the optimization goal to tune and correct the initial centrifugal filtration simulation model to obtain the characteristic centrifugal filtration simulation model; the slurry characteristic acquisition module obtains the historical slurry sampling characteristics of the target scene and determines the incoming material characteristic space, where the slurry sampling characteristics include slurry particle size characteristics, slurry solid content, and slurry temperature; the parameter adjustment and calibration module performs random scatter point distribution according to the incoming material characteristic space to obtain a calibrated incoming material characteristic set, activates the characteristic centrifugal filtration simulation model to simulate and optimize the calibrated incoming material characteristic set, outputs multiple groups of calibrated control parameters, and obtains a calibrated control parameter set; the control execution module performs multi-stage centrifugal control of the heavy alkali slurry based on the calibrated control parameter set. This achieves the technical effects of improving product stability, reducing parameter control costs, and improving monitoring and feedback capabilities.

[0128] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A multi-stage centrifugal control system for heavy alkali filtration, characterized in that: The system comprises: A control record analysis module, which is used to obtain multiple groups of typical control parameters and corresponding typical filtration performance indicators based on historical centrifugal control records; A simulation construction module, wherein the simulation construction module is used to establish an initial centrifugal filtration simulation model, and to set a simulation control series based on multiple groups of typical control parameters, to call the initial centrifugal filtration simulation model for simulation analysis, and to obtain initial simulation analysis results; A simulation correction module, wherein the simulation correction module is used to compare the typical filtration performance index with the initial simulation analysis result, obtain simulation difference information, and perform tuning and correction of the initial centrifugal filtration simulation model with minimization of the simulation difference as the optimization goal to obtain a characteristic centrifugal filtration simulation model; A slurry characteristic acquisition module, which is used to obtain historical slurry sampling characteristics of the target scene and determine the incoming material characteristic space, wherein the slurry sampling characteristics include slurry particle size characteristics, slurry solid content, and slurry temperature; A parameter adjustment and calibration module, wherein the parameter adjustment and calibration module is used to perform random scatter point distribution according to the incoming material characteristic space, obtain a calibration incoming material characteristic set, activate the characteristic centrifugal filtration simulation model to perform simulation and parameter tuning on the calibration incoming material characteristic set, output multiple groups of calibration control parameters, and obtain a calibration control parameter set; A control execution module is used to perform multi-stage centrifugal control of heavy alkali slurry based on the calibration control parameter set.

2. The multi-stage centrifugal control system for heavy alkali filtration according to claim 1, characterized in that: Establishing an initial centrifugal filtration simulation model, and setting a simulation control series based on multiple groups of typical control parameters, calling the initial centrifugal filtration simulation model for simulation analysis, and obtaining initial simulation analysis results, the execution steps include: Establishing a centrifugal filtration numerical model, wherein the numerical model includes a centrifugal force field model, a sedimentation dynamics model, and a control conversion model; Obtaining structural parameters and performance parameters of a target centrifuge, and constructing a simplified centrifugal filtration model based on simulation software, wherein the simplified centrifugal filtration model at least includes a multi-stage drum, a material flow path, and a material discharge path; Fitting the centrifugal filtration numerical model and the centrifugal filtration simplified model to construct the initial centrifugal filtration simulation model, and using multiple groups of typical control parameters as simulation control parameters and slurry characteristics corresponding to the multiple groups of typical control parameters as slurry simulation characteristics to set the simulation control series; The initial centrifugal filtration simulation model is configured based on the simulation control series to perform simulation analysis, and simulation analysis data is recorded and stored as the initial simulation analysis result.

3. The multi-stage centrifugal control system for heavy alkali filtration according to claim 2, characterized in that: Comparing the typical filtration performance index with the initial simulation analysis result, obtaining simulation difference information, and taking minimization of the simulation difference as the optimization goal, performing tuning and correction on the initial centrifugal filtration simulation model to obtain a characteristic centrifugal filtration simulation model, the execution steps include: Based on the filtering performance requirements of the target scenario, performance indicator weights are assigned and a difference information evaluation function is constructed; Taking the control coefficient of the numerical model in the initial centrifugal filtration simulation model as an independent variable, constructing a control coefficient space, and randomly generating a control particle group; The simulation control series is used as multiple target dimensions, and the difference information evaluation function is combined to perform iterative optimization of the regulated particle group to obtain the characteristic centrifugal filtration simulation model.

4. The multi-stage centrifugal control system for heavy alkali filtration according to claim 3, characterized in that: The multi-stage centrifugal control of heavy alkali slurry is performed based on the calibration control parameter set, and the execution steps include: Obtain the incoming sample characteristics of heavy alkali slurry in real time, and obtain the real-time slurry characteristics; Mapping the real-time slurry characteristics to the incoming material characteristic set for similarity evaluation to obtain a slurry characteristic similarity sequence; Based on the slurry characteristic similarity sequence, parameter call is performed in the calibration control parameter set to obtain multi-stage centrifugal control parameters for multi-stage centrifugal control of heavy alkali slurry.

5. The multi-stage centrifugal control system for heavy alkali filtration according to claim 4, characterized in that: The multi-stage centrifugal control of the heavy alkali slurry is performed based on the calibration control parameter set, and the execution step also includes: Using the calibration incoming material characteristic set and the calibration control parameter set as sample data, constructing a decision training set and a decision verification set; Constructing a centrifugal filtration baseline model based on a neural network, performing supervised training through the decision training set, and verifying the centrifugal filtration baseline model with the decision verification set; The real-time slurry characteristics are input into the centrifugal filtration baseline model to make a centrifugal filtration decision, and multi-stage centrifugal control of the heavy alkali slurry is performed according to the decision result.

6. The multi-stage centrifugal control system for heavy alkali filtration according to claim 1, characterized in that: The execution steps of the system also include: Collect vibration data of the target centrifuge and extract the vibration spectrum; An operation stability analysis is performed based on the vibration spectrum, and risk management is performed based on the stability analysis result.

7. The multi-stage centrifugal control system for heavy alkali filtration according to claim 1, characterized in that: The execution steps of the system also include: Analyze the historical centrifugal control records to obtain the filter cake crushing thickness and the filter cake thickness of the shaking car, and calculate and determine the filter cake accumulation limit based on the pushing interval, wherein the filter cake accumulation limit includes the minimum filter cake accumulation rate and the maximum filter cake accumulation rate; Monitor the filter cake thickness growth rate in real time and obtain the real-time filter cake accumulation rate; The filter load is judged and the filter load is controlled in real time based on the filter cake accumulation limit.

8. The multi-stage centrifugal control system for heavy alkali filtration according to claim 1, characterized in that: According to the incoming material characteristic space, random scattered point distribution is performed to obtain a calibration incoming material characteristic set, and the execution step further includes: Based on the historical slurry sampling characteristics, analyzing and determining the probability distribution of the incoming material characteristic space; According to the probability distribution, random scattered point allocation is performed to obtain the calibrated incoming material characteristic set.

Images