Multi-parameter cooperative control method and device for double-prediction-mechanism dense medium coal separation

Abstract

The invention discloses a multi-parameter cooperative control method and device for double-prediction-mechanism dense medium coal separation. The method comprises the steps that multiple pieces of to-be-input data associated with the dense medium coal separation system at the current moment are obtained; inputting the plurality of to-be-input data into a pre-trained model prediction control mechanism model, and determining first prediction suspension density at a plurality of first moments in the future; and inputting a plurality of pieces of data to be input into a pre-trained long-short term memory network, and outputting second predicted suspension density at a plurality of second moments in the future, determining a target predicted suspension density at a third moment in the future through the two predicted suspension densities obtained through prediction; and adjusting at least one control parameter associated with the dense medium coal separation equipment based on the target predicted suspension density, wherein the control parameters comprise one or more of the opening degree of a water replenishing valve, the state of an underflow discharge valve, the frequency of a frequency converter, the opening degree of a pressure valve and the running state of a circulating pump. And the accuracy of multi-parameter cooperative control is improved.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, and in particular to a multi-parameter collaborative control method and device for heavy media coal preparation with a dual prediction mechanism. Background Technology

[0002] In recent years, the coal preparation industry has entered a period of rapid development towards intelligent manufacturing, with heavy media coal preparation being the most widely used separation method. Achieving intelligent control of the heavy media coal preparation process helps to increase the yield of clean coal, reduce the amount of coal entrained in gangue, thereby significantly enhancing production efficiency and greatly improving the operational stability and control precision of the heavy media coal preparation system. The core of this intelligent control lies in the precise regulation of the suspension density.

[0003] Currently, traditional PID control strategies are commonly used in industrial settings, but their control effectiveness is limited. This is mainly due to the strong coupling between the density and level of the suspension: adjusting the opening of the diversion valve, water supply valve, or medium supply valve simultaneously affects both density and level, making independent control difficult. Furthermore, the controlled object exhibits typical large inertia and large hysteresis characteristics, which traditional PID algorithms struggle to handle effectively. In developing this invention, it was discovered that existing technologies suffer from high computational complexity, long solution times, and difficulty in meeting the requirements of multi-parameter coordinated control, leading to poor suspension density control efficiency in heavy media coal preparation systems. Summary of the Invention

[0004] This invention provides a multi-parameter collaborative control method and device for heavy medium coal preparation with a dual prediction mechanism, in order to solve the problems of poor multi-parameter collaborative control effect and poor suspension density control efficiency in heavy medium coal preparation systems.

[0005] According to one aspect of the present invention, a multi-parameter collaborative control method for heavy media coal preparation with dual prediction mechanisms is provided, comprising:

[0006] During the heavy medium coal preparation process, multiple input data related to the heavy medium coal preparation system at the current moment are acquired. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0007] Multiple input data are input into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first time points; and multiple input data are input into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second time points.

[0008] Based on the first and second predicted suspension densities at the same future time, the target predicted suspension density at the third future time is determined; wherein, the third time is the time when multiple first times and multiple second times overlap.

[0009] Based on the target predicted suspension density, at least one control parameter associated with the heavy media coal preparation equipment is adjusted, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0010] Optionally, the method further includes: constructing a target simulation system on a simulation platform based on the hardware devices associated with the heavy media coal preparation system, to generate at least one working condition data in the target simulation system, wherein the working condition data includes simulated input data and corresponding simulated suspension concentration; performing simulation training on the at least one working condition data based on a pre-constructed simulation model to obtain model parameters under the simulation scenario; performing transfer learning based on the model parameters, and training the long short-term memory network to be trained based on the collected sample data to obtain a usable long short-term memory network.

[0011] Optionally, the target predicted suspension density at a third future time is determined based on the first and second predicted suspension densities at the same future time, including: determining first data based on the first predicted suspension density and the corresponding first weight at the same future time, and determining the target predicted suspension density at the third predicted time based on the second predicted suspension density and the corresponding second weight.

[0012] Optionally, after obtaining the target predicted suspension density, the method further includes: acquiring the first predicted suspension density, the second predicted suspension density, and the actual suspension density at the target prediction time; and acquiring the first historical predicted suspension density, the second historical predicted suspension density, and the actual historical suspension density corresponding to at least one historical time within the sliding window; determining a first prediction error based on the first predicted suspension density, the first historical predicted suspension density, the actual suspension density, and the actual historical suspension density; determining a second prediction error based on the second predicted suspension density, the second historical predicted suspension density, the actual suspension density, and the actual historical suspension density; and adjusting the weights of the model predictive control mechanism model and the long short-term memory network based on the first prediction error and the second prediction error, so as to determine the target predicted suspension density at the target prediction time based on the adjusted weights; wherein the prediction error is negatively correlated with the weights.

[0013] Optionally, the method further includes: adjusting the model parameters of the long short-term memory network based on the target predicted suspension density and the actual suspension density at the target prediction time, provided that the first prediction error and / or the second prediction error meet preset conditions.

[0014] Optionally, adjusting at least one control parameter associated with the heavy media coal preparation equipment based on the target predicted suspension density includes: using a parameter adjustment module in the model predictive control mechanism model to determine adjustment information based on the target predicted suspension density and the current data information of at least one control parameter, so as to adjust the corresponding control parameter based on the adjustment information.

[0015] Optionally, after obtaining the adjustment information, the method further includes: sending the adjustment information to a programmable logic controller (PLC) to coordinately control at least one control parameter based on the adjustment information.

[0016] According to another aspect of the present invention, a multi-parameter collaborative control device for heavy media coal preparation with dual prediction mechanisms is provided, comprising:

[0017] The input data acquisition module is used to acquire multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0018] The suspension density prediction module is used to input multiple input data into a pre-trained model prediction control mechanism model to determine the first predicted suspension density at multiple future first moments; and to input multiple input data into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments.

[0019] The target predicted suspension density determination module is used to determine the target predicted suspension density at a third future time based on the first and second predicted suspension densities at the same future time; wherein the third time is the time when multiple first times and multiple second times overlap.

[0020] The control parameter adjustment module is used to adjust at least one control parameter associated with the heavy medium coal preparation equipment based on the target predicted suspension density. The at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0021] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0022] At least one processor; and

[0023] A memory that is communicatively connected to at least one processor; wherein,

[0024] The memory stores a computer program that can be executed by at least one processor, such that the at least one processor is able to execute the multi-parameter collaborative control method for heavy media coal preparation with dual prediction mechanism according to any embodiment of the present invention.

[0025] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement the multi-parameter collaborative control method for heavy media coal preparation with a dual-prediction mechanism according to any embodiment of the present invention.

[0026] The technical solution of this invention involves acquiring multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. These multiple input data include at least one or more of the following: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity in the heavy medium coal preparation equipment, and coal slurry-to-water ratio data of the cavity in the heavy medium coal preparation equipment. The multiple input data are then input into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments. Finally, the multiple input data are input into a pre-trained long-term predictive control mechanism model. In the short-time memory network, the second predicted suspension density at multiple future second moments is output; based on the first and second predicted suspension densities at the same future moment, the target predicted suspension density at a third future moment is determined; wherein, the third moment is the overlapping moment of multiple future first moments and multiple future second moments; based on the target predicted suspension density, at least one control parameter associated with the heavy medium coal preparation equipment is adjusted, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status. This solution acquires real-time and accurate core operating data of the system, and makes parallel predictions using both mechanistic and neural network models. It fully integrates the process fit of mechanistic deduction with the time-series fitting advantages of network models, effectively improving the accuracy and reliability of suspension density prediction. By anchoring the coincidence time to determine the target prediction value, the density prediction results are more in line with the actual time requirements of control. At the same time, based on the prediction results, it conducts forward-looking and targeted coordinated adjustments to multiple control parameters, solving the problems of poor multi-parameter coordinated control effect and poor suspension density control efficiency in heavy media coal preparation systems. It can avoid large fluctuations in suspension density in advance, ensuring that the density of the separation medium is stable within the process standard range, significantly improving the automation control level and separation accuracy of the heavy media coal preparation system, reducing human intervention errors, balancing equipment operation stability and production separation efficiency, and flexibly adapting to the parameter acquisition and control needs of different production conditions, further reducing production energy consumption and medium loss.

[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart of a multi-parameter collaborative control method for heavy media coal preparation with a dual prediction mechanism, provided in Embodiment 1 of the present invention.

[0030] Figure 2 This is a flowchart of a multi-parameter collaborative control method for heavy media coal preparation with a dual prediction mechanism, provided in Embodiment 2 of the present invention.

[0031] Figure 3 This is a schematic diagram of the structure of a multi-parameter collaborative control device for heavy media coal preparation with a dual prediction mechanism, provided in Embodiment 3 of the present invention.

[0032] Figure 4 This is a schematic diagram of the structure of an electronic device for implementing the multi-parameter collaborative control method of heavy media coal preparation with dual prediction mechanism according to embodiments of the present invention. Detailed Implementation

[0033] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] Example 1

[0036] Figure 1 This is a flowchart of a multi-parameter collaborative control method for heavy medium coal preparation using a dual-prediction mechanism, provided in Embodiment 1 of the present invention. This embodiment is applicable to heavy medium coal preparation. The method can be executed by a multi-parameter collaborative control device for heavy medium coal preparation using a dual-prediction mechanism. This multi-parameter collaborative control device can be implemented in hardware and / or software, and can be configured in electronic devices such as computers, servers, and controllers. Figure 1 As shown, the method includes:

[0037] S110. During the heavy medium coal preparation process, acquire multiple input data associated with the heavy medium coal preparation system at the current moment. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0038] The heavy media coal preparation system can be specifically understood as an integrated industrial production system that uses a heavy media suspension as the separation medium and integrates multiple process units and supporting equipment, including separation, media circulation, density control, and coal slurry water treatment. It enables precise density-based separation of raw coal. During the heavy media coal preparation process, the precise control of the density of the heavy media suspension directly affects the quality of the clean coal and the rational utilization of resources. The input data can be understood as various parameter data that need to be entered into the system control terminal during the process control stage of heavy media coal preparation. These parameters reflect the current production conditions and equipment operating status, including but not limited to one or more of the core data such as the suspension density, suspension level, equipment cavity pressure, and the proportion of coal slurry water in the cavity. Multiple input data related to the heavy media coal preparation system can be acquired in real time through the system's sensors and data acquisition modules. Heavy media coal preparation equipment can be understood as the core component of a heavy media coal preparation system. It is the main device that directly carries out the heavy media separation of raw coal, carries the suspension circulation and separation process, and is also the direct carrier for generating various core process data.

[0039] Specifically, density sensors, level sensors, pressure transmitters, and coal slurry concentration detectors are deployed in key areas of the heavy media coal preparation equipment, such as the suspension circulation pipeline and the separation chamber. During the heavy media coal preparation process, various sensors collect specified input data in real time, including suspension density, suspension level, equipment chamber pressure, and the proportion of coal slurry to water in the chamber. Optionally, the single or multiple types of data collected from each point are transmitted to the system data acquisition terminal via an industrial communication network. After the terminal completes real-time reception, preliminary verification, and integration of the data, a complete input dataset corresponding to the heavy media coal preparation system at the current moment is formed, which is finally synchronized to the coal preparation control system for subsequent process control.

[0040] For example, for input data collected by various dedicated sensors deployed in the heavy media coal preparation system, missing values ​​can be filled in based on historical data matching methods under similar working conditions. Millisecond-level multi-source data synchronization can be achieved by relying on the unified timing mechanism of the programmable logic controller, avoiding causal misjudgment caused by sampling delay. In addition, real-time normalization or standardization of various data can be performed to eliminate dimensional differences, which helps to improve the accuracy of subsequent prediction of suspension density.

[0041] In this embodiment, by capturing the key process status of heavy media coal preparation in real time and accurately, it provides real and effective data support for subsequent system control. At the same time, it can flexibly adapt to the collection needs under different production conditions, taking into account the comprehensiveness and relevance of data acquisition. It can also effectively avoid the problems of low efficiency and large error caused by manual collection, greatly improving the efficiency and accuracy of data acquisition, helping to timely control the real-time changes in equipment operation and process conditions, and ensuring the stable and efficient operation of the heavy media coal preparation system.

[0042] S120. Input multiple input data into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments; and input multiple input data into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments.

[0043] Specifically, the Model Predictive Control (MPC) mechanism can be understood as a predictive control model (MPC) established based on the process mechanism of heavy media separation, the hydrodynamic laws of heavy media suspension, and the principle of media density control, combined with the material balance and equipment operating characteristics of raw coal separation. This model relies on input data such as suspension density, liquid level, chamber pressure, and coal slurry-water ratio in the heavy media coal preparation system. Following the inherent technical laws of heavy media separation, it quantitatively describes the impact of changes in each input parameter on the raw coal separation accuracy and system operational stability. It can predict the suspension density in advance and then generate optimal control instructions through rolling optimization control logic, providing precise control basis for the programmable logic controller (PLC control system), achieving adaptive and refined control of the process parameters of the heavy media coal preparation system. The "first moment" can be understood as a set of several continuous or discrete time nodes in the future, defined by the MPC mechanism based on a predetermined time step, used to predict changes in suspension density. The time step can be set according to actual needs and is not limited here. The first predicted suspension density can be understood as the suspension density prediction values ​​for each future time node calculated by the model based on the input data at the first time point dimension. The pre-trained Long Short-Term Memory (LSTM) network can be understood as a deep learning network trained on heavy media coal preparation time-series data, possessing the ability to capture the temporal correlation characteristics and long-term variation patterns of the data. It serves as a dynamic compensation unit for predicting suspension density, effectively capturing the deviation between the actual process and the ideal mechanism model, thereby providing real-time and dynamic compensation for the overall prediction results and significantly improving the model's ability to fit the real system behavior. The second time point can be understood as multiple future time nodes, distinct from the first time point, defined by the LSM network according to a preset time granularity. The second predicted suspension density can be understood as the suspension density prediction results for each future time node output by the LSM network based on the input data at the second time point dimension.

[0044] Specifically, in the process control stage of heavy media coal preparation, the multi-dimensional input data collected from the heavy media coal preparation system is standardized and preprocessed. Then, two data input operations are completed simultaneously. On the one hand, the preprocessed input data is imported into a pre-constructed model predictive control mechanism model after sample training and iterative optimization. Based on the mechanism deduction and calculation logic of the model, the first predicted suspension density corresponding to multiple first moments in the future is accurately solved and determined. On the other hand, the same batch of input data is input into a long short-term memory network that has been trained. With the help of the network's deep learning and feature mining capabilities for time series data, the second predicted suspension density at multiple second moments in the future is calculated and output.

[0045] In this embodiment, the dual advantages of the model predictive control mechanism model and the neural network model are used to complete the multi-time period prediction of suspension density. The mechanism model can ensure that the prediction results are consistent with the actual operation of the separation process, while the long short-term memory network can accurately capture the temporal change characteristics of the data, improving the accuracy and adaptability of density prediction at different times. At the same time, the parallel prediction of the two models can obtain multiple sets of density prediction results from different dimensions, providing more comprehensive and diverse data references for subsequent suspension density control. This effectively avoids the limitations of single model prediction, greatly improves the reliability and scientific nature of suspension density prediction, helps to predict density change trends in advance, accurately formulate control strategies, and ensure the stability of heavy media coal preparation separation effect.

[0046] S130. Based on the first and second predicted suspension densities at the same future time, determine the target predicted suspension density at the third future time; wherein, the third time is the time when multiple first times and multiple second times overlap.

[0047] Specifically, the third moment can be understood as the time point where the multiple future first moments corresponding to the first predicted suspension density output by the model prediction control mechanism in the heavy media coal preparation density prediction stage coincide with the multiple future second moments corresponding to the second predicted suspension density output by the pre-trained long short-term memory network. The target predicted suspension density can be understood as the final accurate predicted suspension density value that can be used for subsequent process control, determined after fusing and calibrating the first and second predicted suspension densities at the same time point at this overlapping third moment.

[0048] Specifically, in the density prediction stage of heavy media coal preparation, the time nodes where multiple first moments and multiple second moments overlap are first screened out and defined as the third moment. Then, the first predicted suspension density corresponding to the overlapping moment is extracted from the model predictive control mechanism model, and the second predicted suspension density corresponding to the overlapping moment is extracted from the long short-term memory network. Data fusion processing is carried out on the two sets of predicted density data at the same moment. Combined with the actual process requirements of heavy media coal preparation and the prediction characteristics of the two types of models, data verification, weight allocation and result integration are completed. Finally, the target predicted suspension density at the third moment is accurately determined.

[0049] In this embodiment, the target predicted suspension density is obtained by fusing the first and second predicted suspension densities. This integrates the dual predictive advantages of the mechanistic model and the neural network model, avoiding the biases and limitations of a single model in density prediction. It effectively improves the accuracy and reliability of the suspension density prediction results at future overlap times. At the same time, relying on the precise anchoring of the overlap time, the target predicted value can accurately match the actual production time control requirements, providing accurate and reliable core data for the advance control of the suspension density in subsequent heavy media coal preparation. This helps to more efficiently control the rhythm of suspension density changes and ensure the stability and separation accuracy of the heavy media coal preparation process.

[0050] S140. Adjust at least one control parameter associated with the heavy media coal preparation equipment based on the target predicted suspension density, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0051] In this context, control parameters can be understood as various operational parameters that can directly affect the heavy medium coal preparation equipment and thus change the equipment's operating conditions and the state of the heavy medium suspension. These parameters include the opening degree of the water supply valve, the state of the underflow discharge valve, the frequency of the frequency converter, the opening degree of the pressure valve, and the operating state of the circulating pump. By making targeted adjustments to one or more of these parameters, precise intervention can be achieved in the operating rhythm of the heavy medium coal preparation equipment, the medium circulation efficiency, and the state of the suspension system. This will ultimately achieve the core objective of regulating the suspension density and ensuring the stable progress of the heavy medium coal preparation process.

[0052] Specifically, during the heavy media coal preparation process, the predicted suspension density is used as the core control benchmark. Deviation analysis and control logic calculations are performed by combining real-time data collected from the heavy media coal preparation equipment, such as the actual suspension density, liquid level, and chamber pressure. This allows for targeted dynamic adjustments to the equipment's associated control parameters. Based on actual operating conditions, any single parameter can be adjusted individually, including the water supply valve opening, underflow discharge valve status, frequency converter frequency, pressure valve opening, and circulating pump operating status. Alternatively, multiple parameters can be combined and optimized simultaneously. For example, when the actual suspension density is higher than the target value, the water supply valve opening can be increased; the circulating pump operating status can be switched and the frequency converter frequency adjusted according to the medium circulation requirements; or the pressure valve opening and underflow discharge valve start / stop can be precisely controlled based on the chamber pressure deviation. This achieves precise intervention in the heavy media coal preparation equipment's operating conditions, realizing the control target of bringing the suspension density closer to the standard value.

[0053] In this embodiment, proactive control based on accurate target prediction of suspension density can avoid abnormal fluctuations in suspension density in advance, ensuring that the density of the separating medium remains stable within the process requirement range, effectively improving the separation accuracy and efficiency of heavy media coal preparation. Simultaneously, it supports flexible selection and combination of multiple control parameters, adapting to control needs under different density deviation scenarios, significantly improving the targeting and adaptability of equipment control. Furthermore, automated parameter adjustment is achieved throughout the process based on predicted data, reducing the involvement of manual control. This avoids human error, improves the automation level of the heavy media coal preparation system, and reduces production energy consumption and media loss through refined parameter adjustment, further ensuring the stable and efficient operation of the heavy media coal preparation equipment.

[0054] Optionally, adjusting at least one control parameter associated with the heavy media coal preparation equipment based on the target predicted suspension density includes: using a parameter adjustment module in the model predictive control mechanism model to determine adjustment information based on the target predicted suspension density and the current data information of at least one control parameter, so as to adjust the corresponding control parameter based on the adjustment information.

[0055] The parameter adjustment module within the model predictive control mechanism can be understood as a core functional unit specifically designed for the control of the heavy media coal preparation process. It incorporates control algorithms and logic rules tailored to the actual production patterns of heavy media coal preparation. It receives target predicted suspension density and related control parameter data, performs calculations and analysis, and ultimately outputs a suitable parameter control scheme. The current data information specifically refers to the real-time actual values ​​and operating status information of control parameters associated with the heavy media coal preparation equipment, such as the opening degree of the water supply valve, the status of the underflow discharge valve, and the frequency converter frequency, at the current moment when parameter adjustments are performed. The adjustment information specifically refers to the specific instructions generated by the parameter adjustment module after calculation and analysis, based on the target predicted suspension density and the current control parameter data. These instructions guide parameter control, including the adjustment direction, adjustment range, and start / stop switching methods for each control parameter, and can be directly used as the basis for precise adjustment of the corresponding control parameters.

[0056] Specifically, based on the target predicted suspension density at the third future moment, the system simultaneously retrieves current data information of at least one control parameter from the heavy media coal preparation equipment, including the opening of the water supply valve, the status of the underflow discharge valve, the frequency converter frequency, the opening of the pressure valve, and the operating status of the circulating pump. Both types of data are input into the parameter adjustment module built into the model predictive control mechanism. This module relies on the preset process control logic and algorithm model to complete data calculation and analysis, accurately determine the adjustment direction, adjustment range, and adjustment method of each control parameter, generate corresponding parameter adjustment information, and finally perform automated adaptation adjustment of the corresponding control parameters associated with the heavy media coal preparation equipment according to the adjustment information, thereby achieving precise pre-control of the suspension density.

[0057] For example, the core mechanism of the parameter adjustment module in the model predictive control mechanism is based on rolling time-domain optimization and feedback correction strategies. Specifically, it can be summarized as follows: First, at each sampling time, based on the current state of the system and the preset internal dynamic model, the system behavior within a finite future time domain is predicted to obtain the target predicted suspension density at the third future time. Second, a constrained optimization problem is constructed and solved to generate a control input sequence, ensuring that the predicted output within this interval closely approximates the desired trajectory of the suspension density while satisfying all set operational constraints. Third, only the first control variable in this control sequence is applied to the actual system. At the next sampling time, the system state information is re-acquired, and the above prediction and optimization process is repeated to achieve closed-loop control. That is, if the predicted target suspension density at the third future time approaches the desired trajectory of the suspension density, the target predicted suspension density at the first time of the third future time can be used in the actual system. The parameter adjustment module determines the adjustment information corresponding to the control parameters based on this target predicted suspension density. If the predicted target suspension density at the third future time does not approach the desired trajectory of the suspension density, the prediction process needs to be repeated.

[0058] In this embodiment, by using the parameter adjustment module of the model predictive control mechanism model to carry out integrated parameter calculation and control command generation, the control parameter adjustment scheme can be made more in line with the heavy media coal preparation process mechanism and equipment operation law, which greatly improves the scientificity and accuracy of parameter adjustment. At the same time, by combining the target predicted suspension density and the current parameter data to determine the adjustment information in two dimensions, the targeted and reasonable parameter adjustment can be effectively guaranteed, avoiding problems such as equipment condition disorder and suspension density fluctuation caused by blind control. Furthermore, the entire parameter adjustment process can be completed automatically by the module, reducing the error and lag of manual intervention, significantly improving the automation control efficiency of the heavy media coal preparation system, ensuring that the suspension density is stable within the process requirements range, and thus continuously optimizing the separation accuracy and production stability of heavy media coal preparation.

[0059] Optionally, after obtaining the adjustment information, the method further includes: sending the adjustment information to a programmable logic controller (PLC) to coordinately control at least one control parameter based on the adjustment information.

[0060] Specifically, after generating the adjustment information of the control parameters of the heavy medium coal preparation equipment, the adjustment information is transmitted completely and in real time to the programmable logic controller supporting the heavy medium coal preparation system through the industrial communication link. The programmable logic controller receives and analyzes the core instructions such as the adjustment direction, amplitude, and start-stop state switching requirements of the control parameters contained in the adjustment information, and according to the preset equipment collaborative control logic, conducts centralized and precise collaborative regulation on at least one control parameter such as the opening degree of the water replenishing valve, the state of the underflow discharge valve, the frequency of the frequency converter, the opening degree of the pressure valve, and the operating state of the circulating pump, and synchronously completes the issuance of the action instructions for the corresponding execution components and the switching of the operating states, and finally implements the adjustment operations of various control parameters.

[0061] In this embodiment, by leveraging the hardware control advantages of the programmable logic controller, the efficient connection and precise implementation of the adjustment information from the algorithm layer to the equipment execution layer are achieved, which can ensure that the adjustment instructions of various control parameters are quickly responded to and strictly executed. At the same time, the programmable logic controller can conduct collaborative overall regulation on multiple control parameters, avoiding problems such as equipment condition imbalance and ineffective suspension liquid density regulation caused by independent adjustment of single parameters, greatly improving the synchronization and coordination of parameter adjustment, and can also utilize the stable industrial control performance of the programmable logic controller to ensure the safety, reliability, and deviation-free adjustment process of various control parameters of the heavy medium coal preparation equipment, further enhancing the suspension liquid density regulation effect, and ensuring the stability of the overall operation of the heavy medium coal preparation system and the accuracy of the separation process.

[0062] Based on the above embodiments, the method further includes: constructing a target simulation system on the simulation platform based on the hardware equipment associated with the heavy medium coal preparation system to generate at least one working condition data in the target simulation system, where the working condition data includes simulated input data and the corresponding simulated suspension liquid concentration; performing simulation training on at least one working condition data based on the pre-constructed simulation model to obtain the model parameters in the simulation scenario; based on the transfer learning of the model parameters, and performing model training on the long short-term memory network to be trained for transfer learning based on the collected sample data to obtain a usable long short-term memory network.

[0063] Specifically, the target simulation system can be understood as a digital simulation system built on a high-fidelity simulation platform, highly compatible with the actual heavy media coal preparation system, based on the structure, operating characteristics, and process logic of the actual hardware equipment. It can accurately replicate various operating scenarios in on-site production. For example, the target simulation system simulates the dynamic operating behavior and multi-parameter coupling mechanisms of key equipment including heavy media separators, flotation machines, and dewatering screens, enabling accurate reproduction of the physical and logical constraints in the actual process. This generates a large number of training samples covering a wide range of operating conditions, accurately labeled, and without privacy risks, providing a sufficient pre-training foundation for long short-term memory networks. The operating condition data can be specifically understood as a set of various data generated in the target simulation system, adapted to heavy media coal preparation production. Its core includes simulated input data for model training and corresponding simulated suspension concentration data. Model parameter transfer learning can be understood as transferring the simulation model parameters trained in a simulation scenario to adapt to the heavy media coal preparation process to the Long Short-Term Memory (LSTM) network to be trained as initial parameters. This allows the network to directly inherit the basic feature extraction capabilities of the simulation model. Then, combined with real-world sample data, subsequent targeted training is completed, achieving simulation model parameter reuse and rapid model optimization. This makes it possible to quickly adapt to on-site operating conditions when deployed to the actual production system by using only a small amount of online collected real-world operating data for fine-tuning. Model parameter transfer learning significantly reduces the dependence on large-scale real-world data while effectively protecting the company's data security and process secrets.

[0064] Specifically, the actual structure, operating characteristics, and process logic of various hardware devices in the heavy media coal preparation system are replicated. A target simulation system that closely matches the on-site system is built in the simulation platform. The target simulation system simulates and recreates various production conditions of heavy media coal preparation, generating at least one set of operating condition data containing simulated input data and corresponding simulated suspension concentrations. The generated operating condition data is then imported into the pre-built simulation model for simulation training. Through multiple rounds of iterative calculations and parameter optimization, model parameters adapted to the simulation scenario are extracted. Subsequently, transfer learning is carried out based on these model parameters, transferring the simulation parameters to the Long Short-Term Memory (LSTM) network to be trained. This gives the network basic parameter characteristics that fit the heavy media coal preparation process. At the same time, combined with real sample data collected on-site, targeted training and optimization iterations are carried out on the LSM network to be trained after parameter transfer, continuously fitting the time-series data characteristics of actual production and correcting model biases. Finally, the training is completed, and a LSM network that can be directly applied to the density prediction of heavy media coal preparation is obtained.

[0065] In this embodiment, a simulation system is built through a simulation platform, which can generate sufficient working condition data at low cost and high efficiency. This effectively solves the problems of difficulty and insufficient quantity of on-site sample data collection. By using transfer learning to reuse the high-quality parameters of the simulation model, the training cycle of the long short-term memory network is significantly shortened. At the same time, by combining training with real sample data, the model can not only fit the process mechanism but also adapt to the actual on-site working conditions, significantly improving the model's generalization ability and prediction accuracy. It can also reduce the manpower and time costs of model training and ensure that the trained network can stably and reliably serve the work of predicting the suspension density of heavy media coal preparation.

[0066] Optionally, to further enhance the robustness and generalization ability of the model predictive control mechanism model and long short-term memory network under complex, abnormal, or even extreme conditions, generative adversarial network (GAN) technology can be introduced to construct a high-fidelity virtual PLC (programmable logic controller) simulation environment. This simulation environment can automatically generate adversarial examples with physical consistency and reasonable control logic to simulate edge or fault scenarios such as sensor drift, actuator failure, and drastic feed fluctuations. By incorporating such adversarial examples into the model training process, the model can be reinforced, enabling it to exhibit stronger stability and reliability when facing uncertainties and disturbances in real-world systems.

[0067] The technical solution of this embodiment involves acquiring multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. These multiple input data include at least one or more of the following: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity in the heavy medium coal preparation equipment, and coal slurry-to-water ratio data of the cavity in the heavy medium coal preparation equipment. The multiple input data are then input into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments. Finally, the multiple input data are input into a pre-trained longitude... In the time-memory network, the second predicted suspension density at multiple future second moments is output; based on the first and second predicted suspension densities at the same future moment, the target predicted suspension density at a third future moment is determined; wherein, the third moment is the overlapping moment of multiple future first moments and multiple future second moments; based on the target predicted suspension density, at least one control parameter associated with the heavy medium coal preparation equipment is adjusted, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status. This solution acquires real-time and accurate core operating data of the system, relies on parallel prediction using both mechanistic and neural network models, and fully integrates the process fit of mechanistic deduction with the time-series fitting advantages of network models. This effectively improves the accuracy and reliability of suspension density prediction. By anchoring the coincidence time to determine the target prediction value, the density prediction results are more in line with the actual time requirements of control. At the same time, based on the prediction results, multiple control parameters are adjusted in a forward-looking and targeted manner, which can avoid the problem of large fluctuations in suspension density in advance, ensure that the density of the separation medium is stable within the process standard range, significantly improve the automation control level and separation accuracy of the heavy media coal preparation system, reduce human intervention errors, balance equipment operation stability and production separation efficiency, and can flexibly adapt to the parameter acquisition and control needs of different production conditions, further reducing production energy consumption and media loss.

[0068] Example 2

[0069] Figure 2 This is a flowchart of a multi-parameter collaborative control method for heavy media coal preparation with a dual-prediction mechanism, provided in Embodiment 2 of the present invention. The method in this embodiment is a further optimization of the method in the above embodiments. Optionally, based on the first predicted suspension density and the corresponding first weight at the same future time, first data is determined; and based on the second predicted suspension density and the corresponding second weight, the target predicted suspension density at the third prediction time is determined. Figure 2 As shown, the method includes:

[0070] S210. During the heavy medium coal preparation process, acquire multiple input data associated with the heavy medium coal preparation system at the current moment. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0071] S220. Input multiple input data into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments; and input multiple input data into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments.

[0072] S230. Based on the first predicted suspension density and the corresponding first weight at the same future time, determine the first data, and the second predicted suspension density and the corresponding second weight, determine the target predicted suspension density at the third prediction time.

[0073] Specifically, the first weight can be understood as a proportion coefficient assigned to the first predicted suspension density output by the model, based on the model's predictive control mechanism's process adaptability, prediction accuracy, and practical application effect in heavy media coal preparation density prediction. Its magnitude reflects the influence of the first predicted suspension density on the final target prediction result. The second weight can be understood as a proportion coefficient set for the second predicted suspension density output by the network, combining the long short-term memory network's fitting ability and prediction stability to the time-series operating data of heavy media coal preparation. It reflects the contribution proportion of the second predicted suspension density to the target prediction result. The first data can be understood as the intermediate data used in the final calculation of the target predicted suspension density, obtained by weighting the first predicted suspension density at the same future third time point with the corresponding first weight.

[0074] Specifically, in the density prediction stage of heavy media coal preparation, the first predicted suspension density output by the model prediction control mechanism is matched with the corresponding first weight, and the second predicted suspension density output by the long short-term memory network is matched with the corresponding second weight. Then, for the two sets of predicted data at the same third time point in the future, the first predicted suspension density and the first weight are weighted to obtain the first data, and the second predicted suspension density and the second weight are weighted. Combined with the obtained first data, a weighted fusion calculation is carried out to accurately determine the target predicted suspension density at the third time point in the future.

[0075] For example, an adaptive fusion strategy based on gated weights is adopted to dynamically integrate the prediction outputs of the two types of models, and the corresponding expression is shown below:

[0076] ;

[0077] in, This indicates the target predicted suspension density. This represents the predicted output of the mechanistic model. This represents the prediction results of the Long Short-Term Memory network. Indicates the fusion weight. , Instead of setting a fixed constant, the system performs online evaluation and adaptive adjustment based on the prediction errors (such as mean square error and absolute deviation) of the two models at the current time or within a recent window. For example, when the mechanistic model fails to predict accurately due to sudden changes in operating conditions, the system will automatically reduce its settings. This assigns higher weights to Long Short-Term Memory (LSTM) networks; conversely, when the system operates stably and the mechanistic model is highly reliable, it increases... This enhances the physical consistency and robustness of control, thereby achieving a dynamic balance between prediction accuracy and model reliability, ensuring that the system can maintain high-precision predictions under various operating conditions.

[0078] In this embodiment, by assigning appropriate weights to the two types of prediction results to achieve weighted fusion, the advantages of model prediction control mechanism model fitting process mechanism and long short time memory network capturing time series features can be fully taken into account. This effectively avoids the bias and limitations of single model prediction, and significantly improves the accuracy and reliability of the target predicted suspension density. At the same time, the flexible configuration of weights can adapt to the prediction needs under different production conditions, so that the fused prediction results are more in line with the actual on-site separation process control requirements. This provides more scientific and reliable data support for the subsequent precise pre-control of suspension density, and helps to stabilize the separation accuracy and equipment operation status of heavy media coal preparation.

[0079] Based on the above embodiments, after obtaining the target predicted suspension density, the method further includes: acquiring the first predicted suspension density, the second predicted suspension density, and the actual suspension density at the target prediction time; and acquiring the first historical predicted suspension density, the second historical predicted suspension density, and the actual historical suspension density corresponding to at least one historical time within a sliding window; determining a first prediction error based on the first predicted suspension density, the first historical predicted suspension density, the actual suspension density, and the actual historical suspension density; determining a second prediction error based on the second predicted suspension density, the second historical predicted suspension density, the actual suspension density, and the actual historical suspension density; and adjusting the weights of the model predictive control mechanism model and the long short-term memory network based on the first prediction error and the second prediction error, so as to determine the target predicted suspension density at the target prediction time based on the adjusted weights; wherein the prediction error is negatively correlated with the weights.

[0080] Specifically, after obtaining the target predicted suspension density, the first predicted suspension density, the second predicted suspension density, and the actual suspension density measured on-site at the target prediction time are extracted. Simultaneously, the first historical predicted suspension density, the second historical predicted suspension density, and the corresponding actual historical suspension density at at least one historical time within the sliding window are retrieved. Then, error calculations are performed using the first predicted suspension density, the first historical predicted suspension density, and the corresponding actual density data to determine the first prediction error of the model predictive control mechanism. Error calculations are then performed using the second predicted suspension density, the second historical predicted suspension density, and the corresponding actual density data to determine the second prediction error of the Long Short-Term Memory (LSTM) network. Following the rule that prediction error and weight are negatively correlated, the first and second weights corresponding to the two models are dynamically adjusted based on the obtained first and second prediction errors. A larger error corresponds to a lower weight percentage in the model, and a smaller error corresponds to a higher weight percentage. Finally, the target predicted suspension density at the target prediction time is recalculated based on the adjusted weights.

[0081] In this embodiment, the prediction error of the dual models is accurately calculated by combining real-time and historical predicted values ​​and actual values. Based on the negative correlation between error and weight, the weights are dynamically optimized, making the weight allocation more consistent with the current prediction accuracy of the model. This effectively corrects the model prediction deviation and significantly improves the accuracy and reliability of the predicted suspension density. At the same time, the application of the sliding window can make full use of historical data to reflect the stability of the model prediction, making the weight adjustment more reasonable and targeted. This continuously optimizes the effect of dual-model fusion prediction, avoids the control errors caused by long-term deviations of a single model, and provides a more accurate basis for the subsequent adjustment of control parameters of the heavy media coal preparation equipment. This further ensures the stable control of the suspension density of the heavy media coal preparation system and the efficient advancement of the separation process.

[0082] Optionally, the method further includes: adjusting the model parameters of the long short-term memory network based on the target predicted suspension density and the actual suspension density at the target prediction time, provided that the first prediction error and / or the second prediction error meet preset conditions.

[0083] Specifically, after calculating the first and second prediction errors, the two types of errors are compared with the preset error threshold conditions. If the first and / or second prediction errors are determined to meet the preset conditions, the target predicted suspension density and the actual suspension density measured on-site at the target prediction time are retrieved. Using these two sets of core data as the basis for correction, the network weights, bias terms, and other model parameters of the Long Short-Term Memory Network are optimized and dynamically adjusted in reverse iteration. By continuously fitting the deviation between the predicted and actual values ​​and correcting the internal parameter configuration of the network, the model of the Long Short-Term Memory Network is iteratively updated.

[0084] In this embodiment, by accurately determining the error threshold, network parameter adjustments are triggered in a timely manner when the model's prediction accuracy fails to meet process requirements. This enables the long short-term memory network to self-optimize and self-iterate, effectively compensating for prediction deviations caused by changes in operating conditions during long-term model operation. This continuously improves the network's prediction accuracy and adaptability. Furthermore, parameter adjustments are only performed when the error meets the conditions, which not only accurately corrects model shortcomings but also avoids the waste of computing power caused by meaningless parameter iterations. This ensures that the long short-term memory network always aligns with the actual production conditions of heavy media coal preparation, guaranteeing the high reliability of its output prediction data. This further enhances the overall effect of dual-model fusion prediction, provides more accurate data support for suspension density control, and steadily improves the separation stability and process control efficiency of the heavy media coal preparation system.

[0085] S240. Adjust at least one control parameter associated with the heavy media coal preparation equipment based on the target predicted suspension density, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0086] The technical solution of this embodiment, through the system's supporting sensors and data acquisition modules, acquires multiple input data related to the current moment of the heavy medium coal preparation system. It can select one or more of the following data combinations based on actual separation requirements: suspension density data, suspension level data, equipment cavity pressure data, and cavity coal slurry-water ratio data. These input data are then fed into a pre-trained model predictive control mechanism model and a long short-term memory network, respectively, outputting the first predicted suspension density for multiple future first-moments. Simultaneously, the long short-term memory network, based on its learning ability of data temporal characteristics, generates multiple future second-moments predictions. The second predicted suspension density at the second time point; then, according to the preset weight allocation rule, the first predicted suspension density corresponding to the same time point in the future is weighted and calculated with the corresponding first weight to obtain the first data, and the second predicted suspension density is weighted and fused with the corresponding second weight to finally determine the target predicted suspension density at the third predicted time point; finally, using the target predicted suspension density as the control benchmark, at least one control parameter associated with the heavy medium coal preparation equipment is dynamically adjusted, and one or more of the following can be flexibly selected for adaptation and adjustment: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status. This system achieves dual-model fusion prediction using a mechanistic model and an LSTM network, balancing the scientific validity of the process mechanism with the accurate capture of data time-series characteristics. This significantly improves the reliability and foresight of the predicted suspension density. Based on the prediction results, the adjustment of control parameters is more targeted, ensuring that the suspension density stably meets the separation requirements, improving the raw coal separation accuracy and system operation stability. Furthermore, it allows for flexible adjustment of multiple parameters to adapt to different coal qualities and operating conditions, reducing manual intervention, lowering energy consumption and media loss, and promoting the heavy media coal preparation system towards a higher level of intelligent and efficient operation.

[0087] Example 3

[0088] Figure 3 This is a schematic diagram of a multi-parameter collaborative control device for heavy media coal preparation with a dual-prediction mechanism, provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes:

[0089] The input data acquisition module 310 is used to acquire multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. Among the multiple input data, at least one or more of the following are included: density data of the suspension of the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0090] The suspension density prediction module 320 is used to input multiple input data into a pre-trained model prediction control mechanism model to determine the first predicted suspension density at multiple future first moments; and to input multiple input data into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments.

[0091] The target predicted suspension density determination module 330 is used to determine the target predicted suspension density at a third future time based on the first and second predicted suspension densities at the same future time; wherein the third time is the time when multiple first times and multiple second times overlap.

[0092] The control parameter adjustment module 340 is used to adjust at least one control parameter associated with the heavy medium coal preparation equipment based on the target predicted suspension density. The at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0093] The technical solution of this embodiment involves an input data acquisition module acquiring multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. These multiple input data include at least one or more of the following: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity in the heavy medium coal preparation equipment, and coal slurry-to-water ratio data of the cavity in the heavy medium coal preparation equipment. A suspension density prediction module inputs these multiple input data into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments. Furthermore, the module inputs these multiple input data into a pre-trained longitude... In the time memory network, the second predicted suspension density at multiple future second moments is output; the target predicted suspension density determination module determines the target predicted suspension density at a third future moment based on the first and second predicted suspension densities at the same future moment; wherein, the third moment is the overlapping moment of multiple future first moments and multiple future second moments; the control parameter adjustment module adjusts at least one control parameter associated with the heavy medium coal preparation equipment based on the target predicted suspension density, wherein at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status. This solution acquires real-time and accurate core operating data of the system, and makes parallel predictions using both mechanistic and neural network models. It fully integrates the process fit of mechanistic deduction with the time-series fitting advantages of network models, effectively improving the accuracy and reliability of suspension density prediction. By anchoring the coincidence time to determine the target prediction value, the density prediction results are more in line with the actual time requirements of control. At the same time, based on the prediction results, it conducts forward-looking and targeted coordinated adjustments to multiple control parameters, solving the problems of poor multi-parameter coordinated control effect and poor suspension density control efficiency in heavy media coal preparation systems. It can avoid large fluctuations in suspension density in advance, ensuring that the density of the separation medium is stable within the process standard range, significantly improving the automation control level and separation accuracy of the heavy media coal preparation system, reducing human intervention errors, balancing equipment operation stability and production separation efficiency, and flexibly adapting to the parameter acquisition and control needs of different production conditions, further reducing production energy consumption and medium loss.

[0094] Based on the above embodiments, optionally, the apparatus is further configured to: construct a target simulation system on a simulation platform based on the hardware devices associated with the heavy media coal preparation system, so as to generate at least one working condition data in the target simulation system, wherein the working condition data includes simulated input data and corresponding simulated suspension concentration; perform simulation training on the at least one working condition data based on a pre-constructed simulation model to obtain model parameters under the simulation scenario; perform transfer learning based on the model parameters, and perform model training on the long short-term memory network to be trained based on the collected sample data to obtain a usable long short-term memory network.

[0095] Optionally, the target predicted suspension density determination module 330 is specifically used to determine first data based on the first predicted suspension density and the corresponding first weight at the same future time, and to determine the target predicted suspension density at the third prediction time based on the second predicted suspension density and the corresponding second weight.

[0096] Optionally, after obtaining the target predicted suspension density, the device is further configured to acquire the first predicted suspension density, the second predicted suspension density, and the actual suspension density at the target prediction time; and to acquire the first historical predicted suspension density, the second historical predicted suspension density, and the actual historical suspension density corresponding to at least one historical time within a sliding window; determine a first prediction error based on the first predicted suspension density, the first historical predicted suspension density, the actual suspension density, and the actual historical suspension density; determine a second prediction error based on the second predicted suspension density, the second historical predicted suspension density, the actual suspension density, and the actual historical suspension density; and adjust the weights of the model predictive control mechanism model and the long short-term memory network based on the first prediction error and the second prediction error, so as to determine the target predicted suspension density at the target prediction time based on the adjusted weights; wherein the prediction error is negatively correlated with the weights.

[0097] Optionally, the device is also used to adjust the model parameters of the long short-term memory network based on the target predicted suspension density and the actual suspension density at the target prediction time, provided that the first prediction error and / or the second prediction error meet preset conditions.

[0098] Optionally, the control parameter adjustment module 340 is specifically used to utilize the parameter adjustment module in the model predictive control mechanism model to determine adjustment information based on the target predicted suspension density and the current data information of at least one control parameter, so as to adjust the corresponding control parameters based on the adjustment information.

[0099] Optionally, after obtaining the adjustment information, the control parameter adjustment module 340 is further configured to send the adjustment information to the programmable logic controller (PLC) so that the PLC can coordinately control at least one control parameter based on the adjustment information.

[0100] The multi-parameter collaborative control device for heavy medium coal preparation with dual prediction mechanism provided in the embodiments of the present invention can execute the multi-parameter collaborative control method for heavy medium coal preparation with dual prediction mechanism provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0101] Example 4

[0102] Figure 4This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0103] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0104] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0105] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the multi-parameter collaborative control method for heavy media coal preparation with dual prediction mechanisms.

[0106] In some embodiments, the multi-parameter coordinated control method for dual-prediction mechanism heavy medium coal preparation can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the multi-parameter coordinated control method for dual-prediction mechanism heavy medium coal preparation described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the multi-parameter coordinated control method for dual-prediction mechanism heavy medium coal preparation by any other suitable means (e.g., by means of firmware).

[0107] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0108] Computer programs for implementing the dual-prediction mechanism heavy media coal preparation multi-parameter collaborative control method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer programs can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0109] Example 5

[0110] Embodiment 5 of the present invention also provides a computer-readable storage medium storing computer instructions for causing a processor to execute a multi-parameter coordinated control method for heavy media coal preparation with a dual-prediction mechanism. The method includes:

[0111] During the heavy medium coal preparation process, multiple input data related to the heavy medium coal preparation system at the current moment are acquired. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment.

[0112] Multiple input data are input into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first time points; and multiple input data are input into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second time points.

[0113] Based on the first and second predicted suspension densities at the same future time, the target predicted suspension density at the third future time is determined; wherein, the third time is the time when multiple first times and multiple second times overlap.

[0114] Based on the target predicted suspension density, at least one control parameter associated with the heavy media coal preparation equipment is adjusted, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

[0115] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0116] To provide interaction with an object, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the object; and a keyboard and pointing device (e.g., a mouse or trackball) through which the object provides input to the electronic device. Other types of devices can also be used to provide interaction with the object; for example, feedback provided to the object can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the object can be received in any form (including sound input, voice input, or tactile input).

[0117] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., a computer with a graphical user interface or web browser through which an item can interact with the implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0118] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0119] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and no limitation is imposed herein.

[0120] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A multi-parameter collaborative control method for heavy media coal preparation with a dual prediction mechanism, characterized in that, include: During the heavy medium coal preparation process, multiple input data associated with the heavy medium coal preparation system at the current moment are acquired. Among these multiple input data, at least one or more of the following are included: density data of the suspension in the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment. Multiple input data are input into a pre-trained model predictive control mechanism model to determine the first predicted suspension density at multiple future first moments; and multiple input data are input into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments. Based on the first and second predicted suspension densities at the same future time, the target predicted suspension density at a third future time is determined; wherein, the third time is the time when multiple first times and multiple second times overlap. Based on the target predicted suspension density, at least one control parameter associated with the heavy media coal preparation equipment is adjusted, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

2. The method according to claim 1, characterized in that, The method further includes: Based on the hardware equipment associated with the heavy media coal preparation system, a target simulation system is constructed on a simulation platform to generate at least one set of operating condition data in the target simulation system, wherein the operating condition data includes simulation input data and corresponding simulation suspension concentration; The model parameters under the simulation scenario are obtained by simulating and training the at least one working condition data based on the pre-built simulation model. Based on the model parameters, transfer learning is performed, and the long short-term memory network to be trained is trained based on the collected sample data to obtain a usable long short-term memory network.

3. The method according to claim 1, characterized in that, The determination of the target predicted suspension density at a third future time based on the first and second predicted suspension densities at the same future time includes: Based on the first predicted suspension density and the corresponding first weight at the same future time, the first data is determined, and the second predicted suspension density and the corresponding second weight are determined to determine the target predicted suspension density at the third prediction time.

4. The method according to claim 1 or 3, characterized in that, After obtaining the target predicted suspension density, the method further includes: Obtain the first predicted suspension density, the second predicted suspension density, and the actual suspension density at the target prediction time; and, Obtain the first historical predicted suspension density, the second historical predicted suspension density, and the actual historical suspension density corresponding to at least one historical moment within the sliding window; The first prediction error is determined based on the first predicted suspension density, the first historical predicted suspension density, the actual suspension density, and the actual historical suspension density. The second prediction error is determined based on the second predicted suspension density, the second historical predicted suspension density, the actual suspension density, and the actual historical suspension density. Based on the first prediction error and the second prediction error, the weights of the model prediction control mechanism model and the long short-term memory network are adjusted to determine the target predicted suspension density at the target prediction time based on the adjusted weights. Among them, the prediction error is negatively correlated with the weight.

5. The method according to claim 4, characterized in that, The method further includes: If the first prediction error and / or the second prediction error meet the preset conditions, the model parameters of the long short-term memory network are adjusted based on the target predicted suspension density and the actual suspension density at the target prediction time.

6. The method according to claim 1, characterized in that, The adjustment of at least one control parameter associated with the heavy media coal preparation equipment based on the target predicted suspension density includes: Using the parameter adjustment module in the model predictive control mechanism model, adjustment information is determined based on the target predicted suspension density and the current data information of at least one control parameter, so as to adjust the corresponding control parameters based on the adjustment information.

7. The method according to claim 6, characterized in that, After obtaining the adjustment information, the method further includes: The adjustment information is sent to the programmable logic controller (PLC) so that the PLC can coordinately control at least one control parameter based on the adjustment information.

8. A multi-parameter collaborative control device for heavy media coal preparation with a dual prediction mechanism, characterized in that, include: The input data acquisition module is used to acquire multiple input data associated with the heavy medium coal preparation system at the current moment during the heavy medium coal preparation process. Among the multiple input data, at least one or more of the following are included: density data of the suspension of the heavy medium coal preparation equipment, liquid level data of the suspension, pressure data of the cavity of the heavy medium coal preparation equipment, and coal slurry water ratio data of the cavity of the heavy medium coal preparation equipment. The suspension density prediction module is used to input multiple sets of input data into a pre-trained model prediction control mechanism model to determine the first predicted suspension density at multiple future first moments; and to input multiple sets of input data into a pre-trained long short-term memory network to output the second predicted suspension density at multiple future second moments. The target predicted suspension density determination module is used to determine the target predicted suspension density at a third future time based on the first and second predicted suspension densities at the same future time; wherein the third time is the overlapping time of multiple future first times and multiple future second times; The control parameter adjustment module is used to adjust at least one control parameter associated with the heavy medium coal preparation equipment based on the target predicted suspension density, wherein the at least one control parameter includes one or more of the following: water supply valve opening degree, underflow discharge valve status, frequency converter frequency, pressure valve opening degree, and circulating pump operating status.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the multi-parameter collaborative control method for heavy media coal preparation with dual prediction mechanism as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the multi-parameter collaborative control method for heavy media coal preparation with a dual-prediction mechanism as described in any one of claims 1-7.

Images