Double-mode practical training platform for chemical experiment equipment based on machine learning

Abstract

The invention discloses a machine learning-based chemical experiment equipment bimodal practical training platform, and relates to the technical field of chemical experiments, and the platform comprises a physical layer which comprises chemical experiment equipment, network equipment and power supply equipment; the data acquisition layer is used for acquiring various data in the experiment process in real time; the cloud service layer is used for receiving and outputting the data information of the data acquisition layer; the analogue simulation layer is used for generating corresponding simulation experiment data; the display and interaction layer is used for displaying the simulation experiment data generated by the simulation software to a user in a visual form; and the management and control layer is used for receiving and outputting a coordination instruction. According to the method, the semi-physical mode and the physical mode are set, by combining physical operation and virtual operation, the intuition and authenticity of physical operation are provided, the limitation of pure virtual simulation software is avoided, the simulation experiment is closer to the actual experiment condition by constructing and optimizing the simulation model, and therefore the experiment period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical engineering experiments, and specifically to a dual-mode training platform for chemical engineering experimental equipment based on machine learning. Background Art

[0002] In the field of chemical engineering experiments, traditional training platforms mainly rely on physical material feeding operation, that is, the experimenter directly operates on the experimental equipment and uses real chemical raw materials for experiments. However, the training platform with physical material feeding operation requires continuous consumption of materials, increasing the training cost, and the single training cycle is relatively long, which limits the number of practice times of users. In addition, there are also some training platforms based on pure virtual simulation software. The above platforms simulate the experimental process through a computer, but lack the experience of physical operation. Although the training platform with pure virtual simulation software can simulate the experimental process, it lacks the intuitiveness and authenticity of physical operation, making it difficult for users to fully master the operation of experimental equipment and the experimental process. Moreover, for beginners or users who are not familiar with the experimental device and process, directly conducting physical material feeding experiments may increase the risk of misoperation, resulting in experimental failure or safety accidents, with a relatively high degree of danger.

[0003] Based on this, a dual-mode training platform for chemical engineering experimental equipment based on machine learning is now provided, which can eliminate the drawbacks existing in the prior art solutions. Summary of the Invention

[0004] The purpose of the present invention is to provide a dual-mode training platform for chemical engineering experimental equipment based on machine learning to solve the problems of high material consumption cost, limited number of training times, lack of intuitiveness of physical operation, and high risk of misoperation in the chemical engineering experimental training platform in the background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A dual-mode training platform for chemical engineering experimental equipment based on machine learning, including a physical layer for chemical engineering unit operation training, the physical layer includes chemical engineering experimental equipment, network equipment, and power supply equipment, and the chemical engineering experimental equipment, network equipment, and power supply equipment are all electrically connected by cables;

[0007] A data acquisition layer, which is arranged on one side of the chemical engineering experimental equipment and is used to collect various data in the experimental process in real time;

[0008] A cloud service layer, which is used to receive and output the data information of the data acquisition layer, perform data processing, feature extraction, and model training operations using machine learning algorithms, and establish an accurate simulation model;

[0009] The simulation layer is used to generate corresponding simulated experimental data in real time using the above simulation model, and display the reaction phenomena, material changes, and product quality information under different parameter settings through simulation software;

[0010] The display and interaction layer is used to display the simulated experimental data generated by the above simulation software to users in a visual form;

[0011] The management and control layer is used to receive and output the coordination instructions of the data acquisition layer, cloud service layer, simulation layer, and display and interaction layer, and perform information management operations according to the instructions.

[0012] Preferably, the data acquisition layer includes:

[0013] A flow meter, which is fixedly installed on the chemical experimental equipment and is used to monitor the change of the material flow rate inside the chemical experimental equipment in real time during the experiment;

[0014] A temperature sensor, which is used to monitor the temperature change inside the chemical experimental equipment during the actual operation in real time;

[0015] A pressure sensor, which is used to monitor the pressure condition inside the chemical experimental equipment during the actual operation in real time;

[0016] A liquid level gauge, which is used to monitor the liquid level height inside the chemical experimental equipment during the actual operation in real time;

[0017] A valve sensor, which is installed on one side of the valve in the chemical experimental equipment and is used to read the valve opening and closing state.

[0018] Preferably, the cloud service layer includes:

[0019] A data cleaning unit, which is used to receive the original data information in the data acquisition layer and perform data screening, noise removal, and non-standard data information removal operations based on the above information;

[0020] A model training unit, which is used to construct and train a simulation model using machine learning algorithms;

[0021] A model optimization unit, which is used to find the weak links of the model and adjust the model structure based on the above information;

[0022] A model deployment unit, which is used to monitor the running status of the model in real time and output prediction results.

[0023] Preferably, the simulation layer includes:

[0024] A state simulation unit, which is used to dynamically simulate the physical state and chemical reaction state inside the chemical experimental equipment through simulation software according to the simulation model provided by the cloud service layer and real-time data feedback;

[0025] A process simulation unit, which is used to adjust reaction conditions, material ratios, and operation steps in chemical process requirements according to different chemical process requirements through simulation software, and calculate and display corresponding experimental results through a simulation model;

[0026] A process simulation unit, which is used to simulate the series connection of several chemical experimental devices according to the actual production process, and display the flow path, processing sequence, and mutual relationship between each link of materials between different chemical experimental devices.

[0027] Preferably, the display and interaction layer includes an optoelectronic sound effect device and an operation panel. The optoelectronic sound effect device is used to display the simulation experimental data generated by the above simulation software in a visual form to the user. The operation panel is used to provide an interface for the user to fully interact with chemical experimental devices, facilitating real-time viewing of the operation status, experimental parameters, and experimental results of the experiment. The optoelectronic sound effect device includes:

[0028] An optoelectronic liquid level gauge, which is installed on one side of the liquid level gauge and is used to display the simulated liquid level height;

[0029] A display module, which is installed on one side of the flowmeter, temperature sensor, pressure sensor, and valve sensor, and is used to display the simulated flow rate, temperature, pressure, and valve opening;

[0030] A sound module, which is used to simulate the operation sound effects during the experiment of chemical experimental devices. The operation sound effects include but are not limited to alarm sounds, prompt sounds, and start-up sound effects of chemical experimental devices;

[0031] A start-stop indicator, which is used to display the start and stop states of different components in the simulated chemical experimental device;

[0032] A virtual valve, which is used to simulate the operation function of a real valve in the operation panel interface;

[0033] A virtual switch, which is used to simulate various switch operations on chemical experimental devices.

[0034] Preferably, the management and control layer includes:

[0035] A user information management unit, which is used to perform user registration and authentication, permission allocation and management, and user information maintenance and update operations;

[0036] An experimental data management unit, which is used to perform data storage and backup, data retrieval and query, and data security and privacy protection operations;

[0037] An assessment result management unit, which is used to formulate an assessment index system according to experimental courses and teaching objectives, and perform automatic scoring and assessment feedback operations on the user's experimental performance based on the above assessment indexes;

[0038] A system configuration unit for monitoring and managing chemical experiment equipment and network equipment.

[0039] Preferably, the simulation software is installed on the network equipment.

[0040] Preferably, the simulation model adopts a prediction model of the neural network type.

[0041] Preferably, the data acquisition layer further includes a diagnosis component for judging whether the sensors at the corresponding monitoring nodes are faulty. The diagnosis component includes a fault judgment module and an indicator light connected electrically. The fault judgment module is used to identify the fault conditions of different sensors and circuits, and the indicator light is used to prompt whether the sensors and circuits meet the fault conditions.

[0042] A usage method of a dual-modal training platform for chemical experiment equipment based on machine learning, and the specific usage steps are as follows:

[0043] S1. The user conducts multiple simulation practice operations through the semi-physical simulation mode, manipulates the optoelectronic sound effect device through the operation panel for virtual experiment operations. The virtual operations include but are not limited to liquid level height, flow rate change, temperature change, and pressure change. The data records and experimental results of the semi-physical simulation mode are transmitted to the cloud service layer through the management and control layer to optimize the simulation model;

[0044] S2. When the user is familiar with the experimental process and the assessment indicators meet the qualified requirements, the user can conduct physical feeding operations through the chemical experiment equipment. The data records and experimental results during the physical operation are also transmitted to the cloud service layer to optimize the simulation model again.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] The present invention provides a dual-modal training platform for chemical experiment equipment based on machine learning, which has two modes: semi-physical mode and physical mode. It can effectively reduce material consumption and environmental pollution. By combining physical operations and virtual operations, the total number of physical feeding experiments is reduced. It not only provides the intuitiveness and authenticity of physical operations but also avoids the limitations of pure virtual simulation software. Through multiple semi-physical simulation exercises, users can be more familiar with the experimental devices and procedures, reduce the risk of misoperation, and reduce the occurrence of experimental failures and safety accidents. By constructing and optimizing the simulation model, the simulation experiment is closer to the actual experimental situation, thereby shortening the experimental cycle, improving teaching efficiency, and helping users repeatedly practice and consolidate experimental skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the overall structure of the present invention.

[0048] Figure 2 It is a schematic structural diagram of the physical layer of the present invention.

[0049] Figure 3 It is a schematic structural diagram of the data acquisition layer of the present invention.

[0050] Figure 4 It is a schematic structural diagram of the cloud service layer of the present invention.

[0051] Figure 5 It is a schematic structural diagram of the simulation layer of the present invention.

[0052] Figure 6 It is a schematic structural diagram of the display and interaction layer of the present invention.

[0053] Figure 7 It is a schematic structural diagram of the management and control layer of the present invention.

[0054] Figure 8 It is a schematic process diagram of the present invention.

[0055] Annotation of reference numerals: Physical layer 100, Chemical engineering experimental equipment 110, Network equipment 120, Power supply equipment 130, Data acquisition layer 200, Flowmeter 210, Temperature sensor 220, Pressure sensor 230, Liquid level gauge 240, Valve sensor 250, Fault judgment module 260, Indicator light 270, Cloud service layer 300, Data cleaning unit 310, Model training unit 320, Model optimization unit 330, Model deployment unit 340, Simulation layer 400, Status simulation unit 410, Process simulation unit 420, Process simulation unit 430, Display and interaction layer 500, Photoelectric sound effect device 510, Photoelectric liquid level gauge 511, Display module 512, Sound module 513, Start / stop indication 514, Virtual valve 515, Virtual switch 516, Operation panel 520, Management and control layer 600, User information management unit 610, Experimental data management unit 620, Assessment result management unit 630, System configuration unit 640. Detailed implementation manners

[0056] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0057] In this embodiment, as Figures 1 - 8 shown, a dual-modal training platform for chemical engineering experimental equipment based on machine learning includes a physical layer 100 for performing chemical engineering unit operation training. The physical layer 100 includes chemical engineering experimental equipment 110, network equipment 120 and power supply equipment 130. The chemical engineering experimental equipment 110, network equipment 120 and power supply equipment 130 are all electrically connected by cables;

[0058] The data acquisition layer 200 is arranged on one side of the chemical engineering experimental equipment 110 and is used for collecting various data in the experiment process in real time;

[0059] The cloud service layer 300 is used for receiving and outputting the data information of the data acquisition layer 200, performing data processing, feature extraction and model training operations by using machine learning algorithms, and establishing an accurate simulation model;

[0060] The simulation layer 400 is used for generating corresponding simulated experimental data in real time by using the above simulation model, and displaying reaction phenomena, material changes and product quality information under different parameter settings through simulation software;

[0061] The display and interaction layer 500 is used for displaying the simulated experimental data generated by the above simulation software to users in a visual form;

[0062] The management and control layer 600 is used for receiving and outputting coordination instructions of the data acquisition layer 200, the cloud service layer 300, the simulation layer 400 and the display and interaction layer 500, and performing information management operations according to the instructions;

[0063] Specifically, before the dual-mode training platform of chemical engineering experimental equipment based on machine learning is used, first check whether the platform can be used normally, and then adjust the experimental data to make the platform enter the state to be used.

[0064] Among them, as Figure 1 and Figure 3 shown, the data acquisition layer 200 includes:

[0065] The flowmeter 210 is fixedly installed on the chemical engineering experimental equipment 110 and is used for monitoring the change of the material flow rate inside the chemical engineering experimental equipment 110 in real time during the experiment process;

[0066] The temperature sensor 220 is used for monitoring the temperature change inside the chemical engineering experimental equipment 110 during the physical operation process in real time;

[0067] The pressure sensor 230 is used for monitoring the pressure condition inside the chemical engineering experimental equipment 110 during the physical operation process in real time;

[0068] The liquid level gauge 240 is used for monitoring the liquid level height inside the chemical engineering experimental equipment 110 during the physical operation process in real time;

[0069] The valve sensor 250 is installed on one side of the valve in the chemical engineering experimental equipment 110 and is used for reading the valve opening and closing state;

[0070] Specifically, during the physical operation process, experimental data is collected in real-time by the data acquisition layer 200. The data should include key system state parameters such as temperature, pressure, flow rate, liquid level, etc. The MQTT data transmission protocol is adopted to ensure that the experimental data can be transmitted to the cloud service layer 300 in real-time, avoiding data loss or delay.

[0071] Among them, as Figure 1 and Figure 4 shown, the cloud service layer 300 includes:

[0072] A data cleaning unit 310, which is used to receive the original data information in the data acquisition layer 200 and perform data screening, noise removal, and non-standard data information removal operations based on the above information. For example, the sensors of the chemical experiment equipment 110 may occasionally show abnormal readings. The data cleaning unit 310 will identify, correct, or eliminate these abnormal data according to the set reasonable threshold range and data change rules, and at the same time perform a normalization operation to map the data uniformly to a specific interval, such as [0, 1] or [-1, 1], so that the data reaches the high-quality and standardized state required for model input;

[0073] A model training unit 320, which is used to construct and train a simulation model using machine learning algorithms. Taking data such as reaction temperature, pressure, and raw material concentration as input features, and the output of the reaction product's yield and quality as output labels, continuously adjust the model's parameters through a large number of data samples, so that the model can accurately capture the complex relationship between input and output, avoid overfitting, and thus be able to make accurate predictions for new experimental data;

[0074] A model optimization unit 330, which is used to find the weak links of the model and adjust the model structure based on the above information. By analyzing the performance of the model on the validation set, such as indicators such as accuracy, recall rate, and mean square error, find the weak links of the model. For example, if the false alarm rate of the model is relatively high when predicting the failure of chemical equipment, the model optimization unit 330 will try to adjust the model structure, such as increasing or decreasing the number of layers and nodes of the neural network, or changing the hyperparameter settings of the algorithm, such as the learning rate and regularization coefficient, to make it more in line with the requirements of practical applications;

[0075] A model deployment unit 340, which is used to monitor the running status of the model and output prediction results in real-time;

[0076] Specifically, the above simulation model can adopt the kernel ridge regression algorithm, which is convenient for effectively predicting and analyzing the key parameters in the simulation of chemical experiment equipment. The parameters include temperature, pressure, flow rate, liquid level:

[0077] For example, when predicting pressure, assume that the input feature x includes factors such as feed flow rate F, the proportion of gas components G in the equipment, and the reaction progress index R. Through M experimental samples (x 1 , P 1 ), (x 2 , P 2 )... (x M , P M ), where P i is the measured pressure value corresponding to the i-th sample. The objective function for training the model to solve is Determine the coefficient α i and the bias b. When there is a new input feature x l , the predicted pressure P l is Thus, the predicted value is obtained;

[0078] When predicting flow rate, assume that the input feature x includes factors such as pipe diameter D, pump power, and material viscosity. Through S experimental samples (x 1 , Q 1 ), (x 2 , Q 2 )... (x S , Q S ), where Q i is the measured pressure value corresponding to the i-th sample. The objective function for training the model to solve is Determine the coefficient α i and the bias b. When there is a new input feature x l , the predicted pressure Q l is Thus, the predicted value is obtained;

[0079] The training and fine-tuning operations of the simulation model are implemented through the following steps:

[0080] Through the data cleaning unit 310 in the cloud service layer 300, clean and format the collected data, remove noise and outliers, make it meet the input requirements of the simulation model, use the collected experimental data as training samples to train the simulation model, adopt the supervised learning method, adjust the parameters and structure of the model by comparing the difference between the output of the simulation model and the actual experimental data. On the basis of model training, fine-tune the model according to the new experimental data. The optimization algorithm adopts the Bayesian optimization algorithm to facilitate improving the accuracy and generalization ability of the simulation model. Establish a feedback mechanism to compare and analyze the output of the simulation model with the actual experimental data. As the experimental data continues to increase and accumulate, continuously learn and update the simulation model, and use the new experimental data to train and optimize the model so that it can adapt to different experimental conditions and scenarios, thereby increasing the accuracy, stability, and reliability of the model.

[0081] Among them, as Figure 1 and Figure 5 shown, the simulation layer 400 includes:

[0082] A state simulation unit 410, which is used to dynamically simulate the physical state and chemical reaction state inside the chemical experimental equipment 110 through simulation software according to the simulation model and real-time data feedback provided by the cloud service layer 300. The simulation software is installed on the network device 120 and is used to accurately reproduce the real-time operating state of the chemical experimental equipment 110 under various working conditions and dynamically simulate the physical state and chemical reaction state inside the equipment;

[0083] A process simulation unit 420, which is used to adjust the reaction conditions, material ratios, and operation steps in the chemical process requirements through simulation software according to different chemical process requirements, and calculate and display the corresponding experimental results through the simulation model;

[0084] A process simulation unit 430, which is used to serially simulate several chemical experimental equipment 110 according to the actual production process, and display the flow path, processing sequence, and mutual relationship between each link of the material between different chemical experimental equipment 110;

[0085] Specifically, when simulating the operation of the reaction kettle, the state simulation unit 410 can display in real time the changes in the temperature and pressure of the material in the reaction kettle over time, as well as the phase change of the material (such as the transition from liquid to gas). For example, when simulating the process of synthesizing a certain chemical substance, parameters such as reaction temperature, catalyst dosage, and reaction time can be changed, and then the changes in product yield, purity, and by-product generation amount can be intuitively seen.

[0086] Among them, as Figure 1 and Figure 6 shown, the display and interaction layer 500 includes an optoelectronic sound effect device 510 and an operation panel 520. The optoelectronic sound effect device 510 is used to display the simulated experimental data generated by the above simulation software in a visual form to the user. The operation panel 520 is used to provide an interface for the user to fully interact with the chemical experimental equipment 110, facilitating real-time viewing of the operation state, experimental parameters, and experimental results of the experiment. The optoelectronic sound effect device 510 includes:

[0087] An optoelectronic liquid level gauge 511, which is installed on one side of the liquid level gauge 240 and is used to display the simulated liquid level height;

[0088] A display module 512, which is installed on one side of the flow meter 210, temperature sensor 220, pressure sensor 230, and valve sensor 250 and is used to display the simulated flow rate, temperature, pressure, and valve opening;

[0089] The sound module 513 is used to simulate the operating sound effects during the experiment process of the chemical experiment equipment 110. The operating sound effects include, but are not limited to, alarm sounds, prompt sounds, and startup sound effects of the chemical experiment equipment 110;

[0090] The start-stop indicator 514 is used to display the start and stop states of different components in the simulated chemical experiment equipment 110. For example, it lights green when starting and red when stopping;

[0091] The virtual valve 515 is used to simulate the operation function of a real valve in the operation panel 520 interface. The virtual valve 515 can be opened, closed, and adjusted in terms of opening degree by means of clicking, dragging, etc. Its operation effect is the same as operating a valve on a real device;

[0092] The virtual switch 516 is used to simulate various switch operations on the chemical experiment equipment 110. By clicking the virtual switch 516, operations such as starting, stopping, and switching the working mode of the equipment can be achieved.

[0093] Specifically, the combined operation of physical operation and virtual operation is achieved through the following steps:

[0094] Install the optoelectronic sound effect device 510 at the corresponding positions of pipelines, valves, and power components in the chemical experiment equipment 110. Install the display module 512 at the pressure gauge, temperature measurement point, and electric valve accessory to display the simulated pressure, temperature, and valve opening degree. Install the optoelectronic liquid level gauge 511 on the liquid level gauge 240 to display the simulated liquid level. Install the sound module 513 on power equipment such as centrifugal pumps, centrifuges, and compressors to simulate the operating sound effects. Then assign physical addresses to several optoelectronic sound effect devices 510, connect them to the local area network, bind the corresponding modules on the control software, start the listening service, so as to read the status values transmitted by each module. Distribute the initial status values, including initial pressure, initial temperature, initial liquid level, and initial valve opening degree, to several optoelectronic sound effect devices 510 on the operation panel 520. When it is monitored that the user operates the valve, read the status value transmitted by the data acquisition layer 200 as the input value, calculate the corresponding output value through the machine learning algorithm in the simulation model, and then distribute the calculation result to the optoelectronic sound effect device 510, and synchronously display the corresponding changes on the operation panel 520. When operations such as starting and stopping the pump, starting and stopping the heater, opening and closing the electric valve, and adding and discharging materials are required, virtual operations are performed through the start-stop indicator 514, the virtual valve 515, and the virtual switch 516. Clicking the corresponding button can perform virtual operations.

[0095] Among them, as Figure 1 and Figure 7 shown, the management and control layer 600 includes:

[0096] The user information management unit 610 is used to perform user registration and authentication, permission allocation and management, and user information maintenance and update operations. It is responsible for handling the registration process of users on the training platform, collecting users' basic information, and allocating corresponding operation permissions according to the identity and role of users. It regularly organizes and maintains user information, and timely updates information such as users' contact information and role changes.

[0097] The experimental data management unit 620 is used to perform data storage and backup, data retrieval and query, and data security and privacy protection operations. It uses a reliable database management system to classify and store the massive data generated during the experiment. The data includes the operation data of experimental equipment, experimental operation records, experimental result data, etc. Users can quickly locate the required experimental data according to keywords such as experiment name, experiment time, and experiment personnel.

[0098] The assessment result management unit 630 is used to formulate an assessment index system according to experimental courses and teaching objectives, and perform automatic scoring and assessment feedback operations on users' experimental performance based on the above assessment indicators. When users repeatedly practice and become familiar with experimental operations in the practice mode, they can enter the assessment mode for assessment tests. In the assessment mode, users need to independently determine the operation points and parameter settings, and the system will also make real-time responses to users' operations. The assessment mode can score users' operations according to the step-marked point status values, and the step scores can be set. After users submit their scores, the scores are uploaded to the system administrator side for teachers to evaluate.

[0099] The system configuration unit 640 is used to monitor and manage the chemical experimental equipment 110 and network equipment 120, monitor and manage the hardware resources relied on by the training platform, such as servers, storage devices, network equipment 120, etc., real-time monitor the operation status of hardware devices, and can regularly update and maintain the software system of the platform to improve users' usage experience.

[0100] Among them, as Figure 1 shown, the simulation model adopts a prediction model of the neural network type and uses the neural network machine learning algorithm, which is convenient for constructing and training the simulation model. On the basis of model training, the model is fine-tuned according to new experimental data. The fine-tuning process includes adjusting the model parameters, adding new feature variables, improving the model structure, etc. Through iterative training and fine-tuning, the output of the simulation model is made closer to the actual experimental data. The real-time simulation algorithm adopts algorithms such as kernel ridge regression to ensure that the simulation model can accurately respond to the changes in experimental data. During the model training process, the Bayesian optimization algorithm is used to find the optimal model parameters, combined with the optimization algorithm to improve the accuracy and generalization ability of the simulation model.

[0101] Among them, as Figure 1 and Figure 3As shown, the data acquisition layer 200 further includes a diagnosis component for determining whether the sensors at the corresponding monitoring nodes are faulty. The diagnosis component includes a fault judgment module 260 and an indicator light 270 that are electrically connected. The fault judgment module 260 is used to identify the fault conditions of different sensors and circuits, and the indicator light 270 is used to indicate whether the sensors and circuits meet the fault conditions. Through the fault judgment module 260, the circuits of the sensors and circuits can be repaired to avoid short-circuit situations. When the sensors and circuits are faulty, the indicator light 270 emits red light, which is convenient for the subsequent operation of the training platform.

[0102] During use, in the physical feeding mode, the experimenter directly operates on the chemical experiment equipment 110 and conducts experiments using real chemical raw materials. The data acquisition layer 200 internally collects experimental data in real time through a flow meter 210, a temperature sensor 220, a pressure sensor 230, a liquid level gauge 240, and a valve sensor 250, including temperature, pressure, flow rate, liquid level, and valve opening and closing states. After cleaning the above data through the data cleaning unit 310, a simulation model is constructed and trained through the model training unit 320 to make the simulation experiment closer to the actual experiment situation; in the semi-physical simulation operation mode, the user conducts virtual operations (such as feeding, heating, refrigeration, etc.) through the operation panel 520. The above achieves the simulation effect through simulation software, simulates experimental data according to the trained model, and simulates the temperature, pressure, sound effect, valve opening and closing states, equipment startup states, etc. through the optoelectronic sound effect device 510 to provide a near-real experimental experience. At the same time, the user can also perform physical operations, which not only provides the intuitiveness and authenticity of physical operations but also avoids the limitations of pure virtual simulation software.

[0103] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A dual-modal training platform for chemical experimental equipment based on machine learning, characterized in that: The invention comprises a physical layer (100) for performing chemical unit operation training, wherein the physical layer (100) comprises chemical experiment equipment (110), network equipment (120) and power supply equipment (130), and the chemical experiment equipment (110), network equipment (120) and power supply equipment (130) are all electrically connected through cables; A data collection layer (200), the data collection layer (200) being arranged on one side of the chemical experiment equipment (110) and used for collecting various data in the experiment process in real time; The cloud service layer (300) is used to receive and output data information from the data collection layer (200), perform data processing, feature extraction and model training operations using a machine learning algorithm, and establish an accurate simulation model; The simulation layer (400) is used to generate corresponding simulation experimental data in real time using the above simulation model, and to display reaction phenomena, material changes and product quality information under different parameter settings through simulation software; The display and interaction layer (500) is used to display the simulation experiment data generated by the simulation software to the user in a visual form; The management and control layer (600) is used to receive and output coordination instructions from the data collection layer (200), the cloud service layer (300), the simulation layer (400), and the display and interaction layer (500), and perform information management operations according to the instructions.

2. According to claim 1, a dual-modal training platform for chemical experimental equipment based on machine learning is characterized in that: The data collection layer (200) comprises: A flow meter (210), wherein the flow meter (210) is fixedly mounted on the chemical experiment equipment (110) and is used to monitor in real time the change in the flow rate of materials inside the chemical experiment equipment (110) during the experiment; A temperature sensor (220) is used to monitor the temperature change inside the chemical experiment equipment (110) in real time during the actual operation process; A pressure sensor (230) is used to monitor the pressure condition inside the chemical experiment equipment (110) in real time during the actual operation process; A liquid level meter (240) for real-time monitoring of the liquid level inside the chemical experimental equipment (110) during actual operation; A valve sensor (250) is installed on one side of a valve in the chemical experimental equipment (110) and is used to read the valve switch status.

3. According to claim 1, a dual-modal training platform for chemical experimental equipment based on machine learning is characterized in that: The cloud service layer (300) includes: A data cleaning unit (310), used for receiving raw data information in the data acquisition layer (200) and performing data screening, noise removal and irregular data information operations based on the raw data information; A model training unit (320), used to perform simulation model construction and training operations using a machine learning algorithm; A model optimization unit (330), used to find out the weak links of the model and adjust the model structure based on the above information; The model deployment unit (340) is used to monitor the operation of the model in real time and output the prediction results.

4. According to the dual-modal training platform for chemical experimental equipment based on machine learning in claim 1, it is characterized in that: The simulation layer (400) comprises: A state simulation unit (410) is used to dynamically simulate the physical state and chemical reaction state inside the chemical experimental equipment (110) through simulation software based on the simulation model and real-time data feedback provided by the cloud service layer (300); A process simulation unit (420) is used to adjust the reaction conditions, material ratios, and operation steps in the chemical process requirements through simulation software according to different chemical process requirements, and to calculate and display corresponding experimental results through simulation models; The process simulation unit (430) is used to simulate the serial connection of a plurality of chemical experimental equipment (110) according to the actual production process, and to display the flow path of materials between different chemical experimental equipment (110), the processing sequence and the mutual relationship between each link.

5. The dual-modal training platform for chemical experimental equipment based on machine learning according to claim 1 is characterized in that: The display and interaction layer (500) includes a photoelectric sound effect device (510) and an operation panel (520). The photoelectric sound effect device (510) is used to display the simulation experiment data generated by the simulation software to the user in a visual form. The operation panel (520) is used to provide the user with an interface for comprehensive interaction with the chemical experiment equipment (110), so as to facilitate real-time viewing of the operating status, experimental parameters, and experimental results of the experiment. The photoelectric sound effect device (510) includes: A photoelectric liquid level meter (511), which is installed on one side of the liquid level meter (240) and is used to display the height of the simulated liquid level; A display module (512), the display module (512) being installed on one side of the flow meter (210), the temperature sensor (220), the pressure sensor (230), and the valve sensor (250), and being used to display simulated flow, temperature, pressure, and valve opening; A sound module (513) is used to simulate the operating sound effects of the chemical experiment equipment (110) during the experiment process, wherein the operating sound effects include but are not limited to alarm sounds, prompt sounds, and startup sound effects of the chemical experiment equipment (110); A start and stop indication (514) for displaying the start and stop status of different components in the simulated chemical experimental equipment (110); A virtual valve (515) is used to simulate the operation function of a real valve in the operation panel (520) interface; The virtual switch (516) is used to simulate various switch operations on the chemical experimental equipment (110).

6. The dual-modal training platform for chemical experimental equipment based on machine learning according to claim 1 is characterized in that: The management and control layer (600) comprises: A user information management unit (610) is used to perform user registration and authentication, authority allocation and management, user information maintenance and update operations; An experimental data management unit (620) is used to perform data storage and backup, data retrieval and query, and data security and privacy protection operations; The assessment result management unit (630) is used to formulate an assessment indicator system according to the experimental course and teaching objectives, and automatically score and provide assessment feedback to the user's experimental performance based on the above assessment indicators; The system configuration unit (640) is used to monitor and manage the chemical experimental equipment (110) and the network equipment (120).

7. The dual-modal training platform for chemical experimental equipment based on machine learning according to claim 1 is characterized in that: The simulation software is installed on the network device (120).

8. The dual-modal training platform for chemical experimental equipment based on machine learning according to claim 3 is characterized in that: The simulation model adopts a prediction model of a neural network type.

9. The dual-modal training platform for chemical experimental equipment based on machine learning according to claim 2 is characterized in that: The data collection layer (200) also includes a diagnostic component for determining whether a sensor at a corresponding monitoring node has a fault, the diagnostic component including an electrically connected fault determination module (260) and an indicator light (270), the fault determination module (260) being used to identify fault conditions of different sensors and circuits, and indicating whether the sensor and circuit meet the fault condition through the indicator light (270).

10. A method for using the dual-modal training platform for chemical experimental equipment based on machine learning according to any one of claims 1 to 9, wherein the specific steps of use are as follows: S1. The user performs multiple simulation exercises in a semi-physical simulation mode, and operates the photoelectric sound effect device (510) through an operation panel (520) to perform virtual experimental operations. The virtual operations include but are not limited to liquid level height, flow rate change, temperature change and pressure change. The data records and experimental results of the semi-physical simulation mode are transmitted to the cloud service layer (300) through the management and control layer (600), and the simulation model is optimized. S2. When the user is familiar with the experimental process and the assessment indicators meet the qualified requirements, the user can perform physical material feeding operations through the chemical experimental equipment (110). The data records and experimental results during the physical operation are also transmitted to the cloud service layer (300) to optimize the simulation model again.

Images