A behavioral analysis and control method for safe operation of fume hoods

Abstract

This invention discloses a behavior analysis and control method for safe operation of fume hoods, belonging to the field of behavior analysis and control technology. The method includes the following steps: defining a safety operation benchmark and real-time monitoring system for the fume hood; performing behavior pattern analysis based on benchmark parameters and real-time monitoring data; issuing safety warnings based on the behavior analysis results; implementing automatic control measures based on the warning signals; optimizing energy efficiency based on the control results; providing long-term operational data feedback based on the optimization scheme; further optimizing the behavior patterns and energy efficiency based on the operational data; and automatically upgrading and self-adjusting the system based on the optimization results. This invention effectively improves the safety, energy efficiency, and operational efficiency of fume hood operation, ensuring the stability and high efficiency of the system in complex environments, and enhancing the system's flexibility and adaptability to different environments. It enables the fume hood to maintain efficient and safe operation under frequently changing operating conditions in laboratory or industrial environments.

Description

Technical Field

[0001] This invention relates to the field of behavior analysis and control technology, specifically to a behavior analysis and control method for safe operation of a fume hood. Background Technology

[0002] With increasingly stringent laboratory safety management requirements, fume hoods, as crucial equipment for ensuring laboratory environmental safety and operator health, are receiving growing attention for their operational safety and energy efficiency. Fume hood control systems typically rely on rule-based settings, such as sensor data on airflow velocity, vibration range, and temperature and humidity. However, these traditional control methods often exhibit limitations when facing complex and dynamic experimental environments. With the rapid development of intelligent technologies, reinforcement learning, deep learning, and transfer learning are gradually being introduced into fume hood control systems to achieve adaptive optimization. Nevertheless, existing technologies still face two main challenges when applying these advanced methods:

[0003] 1. Existing technologies have shortcomings in multi-factor comprehensive optimization: Current reinforcement learning methods typically focus on optimizing a single objective, neglecting the complex relationships between safety, energy efficiency, and equipment stability during fume hood control. Traditional methods rely more on rule setting, and when faced with complex airflow fluctuations, vibration changes, and abnormal temperature and humidity, the system struggles to make real-time and effective adjustments.

[0004] 2. Existing technologies suffer from insufficient environmental adaptability: Current deep learning and transfer learning technologies have weak adaptability to new environments. When equipment is moved to different experimental environments or faces unknown operating states, existing learning models often require a large amount of new data for training, resulting in a slow adaptation process and an inability to quickly adjust control strategies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a behavior analysis and control method for safe operation of fume hoods, thereby solving the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a behavior analysis and control method for safe operation of a fume hood, comprising the following steps:

[0008] S1. Define the safe operating standards and real-time monitoring system for fume hoods to obtain fume hood operation data;

[0009] S2. Based on the fume hood operation data, use machine learning algorithms to perform behavioral pattern analysis and obtain behavioral analysis results;

[0010] S3. Based on the behavioral analysis results, a safety warning is issued, and a warning signal is obtained;

[0011] S4. Implement automatic control measures based on the early warning signal to obtain the control result;

[0012] S5. Based on the control results, optimize the energy efficiency by using an optimization algorithm to optimize the control parameters and obtain the optimization scheme;

[0013] S6. Based on the optimization plan, long-term operational data feedback is obtained to get the optimized operational data;

[0014] S7. Based on the operational data, further optimize the behavior patterns and energy efficiency to obtain the optimization results;

[0015] S8. Based on the optimization results, automatically upgrade and self-adjust the system.

[0016] To further optimize this technical solution, the steps in S1 for defining the safe operation benchmark and real-time monitoring system for the fume hood include:

[0017] Define the safe operating standards for the fume hood and obtain the standard parameters;

[0018] The status of the fume hood is monitored in real time through a sensor network to obtain fume hood operation data;

[0019] Using Internet of Things (IoT) technology, real-time monitoring data is transmitted to a data processing center;

[0020] Determine whether the fume hood is within the safe reference range;

[0021] If the safety baseline is exceeded, the abnormal event will be recorded and an alarm signal will be generated.

[0022] To further optimize this technical solution, the step of providing a security warning in S3 includes:

[0023] Based on the behavioral analysis results, the system determines whether there are abnormal behavioral patterns;

[0024] An abnormal behavior pattern is detected, and a warning signal is generated based on a pattern recognition algorithm.

[0025] Compare behavioral patterns with historical data to determine whether they belong to known dangerous patterns;

[0026] If the condition is classified as a known hazard, an emergency alert will be issued and recommendations for appropriate safety measures will be provided.

[0027] To further optimize this technical solution, the step of comparing behavioral patterns with historical data includes:

[0028] Data collection and preprocessing: Real-time monitoring data is obtained through sensor networks, and the data is preprocessed to obtain preprocessed data;

[0029] Behavioral pattern recognition: The system analyzes preprocessed data and identifies behavioral patterns using machine learning algorithms;

[0030] Storage and management of historical data: Store historical operation data and labeled behavior patterns in the system's historical database, and label historical data as a reference standard;

[0031] Pattern comparison and matching: After identifying the behavior pattern at the current moment, the system compares it with historical data, calculates the similarity between the current pattern and historical patterns, and matches it with historical patterns based on the similarity.

[0032] Decision-making and feedback: If a new behavior pattern is highly similar to an abnormal pattern, a safety warning or automatic adjustment measure will be triggered; if it is highly similar to a normal pattern, the current operating status will be maintained and data changes will continue to be monitored.

[0033] To further optimize this technical solution, the step of implementing automatic control measures in S4 includes:

[0034] Real-time detection and early warning signals;

[0035] Upon detecting a warning signal, the system automatically activates preset safety control measures;

[0036] By utilizing adaptive control technology, strategies are adjusted based on feedback from safety control measures to enhance the system's ability to respond to abnormal situations.

[0037] To further optimize this technical solution, the step of providing long-term operational data feedback in S6 includes:

[0038] The system continuously monitors the optimized operational data;

[0039] The system continuously evaluates the persistence and reliability of the optimization effect based on changes in real-time data;

[0040] The system adjusts based on the evaluation results to ensure that the equipment operates in a safe and efficient state.

[0041] To further optimize this technical solution, the step in S7 of further optimizing the behavior pattern and energy efficiency includes:

[0042] Based on historical data and real-time feedback, the system continuously learns and improves the control algorithm by modifying the algorithm weight coefficients, using deep reinforcement learning technology with multi-factor dynamic adjustment, combined with behavioral analysis results and energy efficiency optimization, thereby further optimizing system performance.

[0043] To further optimize this technical solution, the deep reinforcement learning technique based on multi-factor dynamic adjustment includes:

[0044] Model building:

[0045] Construct the following reward function for training and optimization of reinforcement learning:

[0046] ;

[0047] ;

[0048] in:

[0049] Airflow velocity : Difference in vibration frequency Temperature and humidity changes are acquired through a sensor network, among which... Calculated using a weighted average method: For temperature, For humidity, Temperature weighting coefficient, Humidity weighting coefficient ;

[0050] , , : Reference parameters, defined in step S1;

[0051] The smaller the deviation in airflow speed, the greater the reward.

[0052] The smaller the deviation in vibration frequency, the greater the reward.

[0053] The smaller the deviation in temperature and humidity, the greater the reward.

[0054] Energy efficiency score: the higher the score, the more energy-efficient the system.

[0055] Ideal power consumption;

[0056] Actual power consumption;

[0057] , , , Weighting coefficients, adjusted based on actual results;

[0058] Model usage:

[0059] Input data; calculate energy efficiency score; through the calculation of reward function, the system obtains control strategy and adjusts it according to actual execution results to obtain the optimal control strategy; as the environment changes or the equipment operating time increases, the system continues to adjust the control strategy to adaptively respond to new operating states.

[0060] To further optimize this technical solution, the steps for automated system upgrades and self-adjustments in S8 include:

[0061] Based on the optimization results, the system continuously updates its control strategy through deep learning and transfer learning techniques, thereby adjusting and upgrading itself.

[0062] To further optimize this technical solution, the deep learning and transfer learning techniques include:

[0063] Model building:

[0064] By combining deep learning and transfer learning, the following model is constructed:

[0065] ;

[0066] in:

[0067] In time Time-based control output;

[0068] : These are the weighting coefficients between deep learning and transfer learning, controlling the relative influence of the two in the final output, and are adjusted according to the actual results;

[0069] Environmental differences are derived by comparing data acquired by sensor networks in the new and old environments.

[0070] Deep learning functions output a control strategy based on the input data and continuously optimize it;

[0071] : Transfer learning function, adjust control strategy to adapt it to new environment;

[0072] Model usage:

[0073] Input data into a deep learning model to obtain an initial control policy; optimize the initial control policy through the deep learning model to obtain an optimized control policy; adjust the optimized control policy through transfer learning to obtain an adjusted control policy; continuously adjust the weight coefficients to adapt to the environment and provide feedback, and finally output the optimal control policy.

[0074] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, they implement the steps of a behavior analysis and control method for safe operation of a fume hood as described in the first aspect of the present invention.

[0075] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a behavior analysis and control method for safe operation of a fume hood as described in the first aspect of the present invention.

[0076] Compared with the prior art, the present invention provides a behavior analysis and control method for safe operation of fume hoods, which has the following beneficial effects:

[0077] This behavioral analysis and control method for safe operation of fume hoods successfully addresses the shortcomings of existing technologies by combining multi-factor comprehensive optimization with environmental adaptability. Traditional fume hood control methods typically rely on single factors or static rules for operational decisions, failing to comprehensively consider the complex relationships between multiple dimensions of data, such as airflow velocity, vibration range, and temperature and humidity changes. By introducing a multi-factor comprehensive optimization method combining reinforcement learning and deep learning, the system achieves reinforcement learning objectives. This allows the system to dynamically adjust the equipment's operating status and comprehensively evaluate real-time changes in multiple factors, including airflow, vibration, and temperature and humidity, thereby maximizing energy efficiency and operational efficiency while ensuring equipment safety. Reinforcement learning optimizes the control strategy through multiple explorations and feedback, while deep learning improves the accuracy of multi-factor decision-making and the system's adaptability by automatically extracting and fusing key features. This comprehensive optimization method effectively enhances the operational safety of fume hoods while avoiding the risk of neglecting certain key factors in traditional methods, ensuring the system's stability and efficiency in complex environments.

[0078] Furthermore, this system addresses the limitations of traditional equipment in terms of environmental adaptability. Traditional control methods often require readjustment or retraining when the environment changes, lacking the ability to quickly adapt to new environments. However, by introducing transfer learning, this system can rapidly transfer existing knowledge between different experimental environments, achieving rapid adaptation to new environments. Transfer learning, through fine-tuning of deep learning models and utilizing the similarity between historical data and the current environment, quickly updates control strategies, reducing the time and cost of retraining. This rapid adaptability greatly improves the system's flexibility, enabling it to cope with frequently changing operating conditions in laboratory or industrial environments, ensuring the equipment is always in optimal working condition. In this way, the system not only improves the adaptability of the fume hood to different environments but also maintains efficient and safe operation under constantly changing experimental needs and environmental conditions. Attached Figure Description

[0079] To more clearly illustrate the technical solutions of 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.

[0080] Figure 1 This is a flowchart illustrating a behavior analysis and control method for safe operation of a fume hood proposed in this invention.

[0081] Figure 2 This is a flowchart illustrating the comparison of behavioral patterns and historical data in a behavioral analysis and control method for safe operation of a fume hood proposed in this invention. Detailed Implementation

[0082] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0083] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0084] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0085] Example 1:

[0086] Reference Figures 1-2 This is the first embodiment of the present invention, which provides a behavior analysis and control method for safe operation of a fume hood, including the following steps:

[0087] S1. Define the safety operation benchmarks and real-time monitoring system for fume hoods to obtain fume hood operation data.

[0088] In this embodiment, the steps for defining the safe operating standards and real-time monitoring system for fume hoods include:

[0089] Define safe operating standards for fume hoods:

[0090] Based on industry standards, historical maintenance records, and equipment manufacturer recommendations, establish safe operating benchmarks for fume hoods, including key parameters such as airflow velocity, vibration range, and temperature and humidity changes.

[0091] Real-time monitoring of fume hood status via sensor network:

[0092] Use sensors such as airflow sensors, vibration sensors, and temperature and humidity sensors to monitor the status of the fume hood in real time and compare it with the baseline value.

[0093] Using IoT technology, the operating data of the fume hood collected by sensors is transmitted to the data processing center in real time;

[0094] Determine if the fume hood is within a safe operating range;

[0095] If the abnormal event exceeds the safety baseline range, an alarm signal will be generated.

[0096] S2. Conduct behavioral pattern analysis based on fume hood operation data, use machine learning algorithms to analyze real-time monitoring data, identify safety hazards in advance, and obtain behavioral analysis results.

[0097] In this embodiment, machine learning algorithms (such as Support Vector Machine (SVM) or Deep Neural Network (DNN)) are used to analyze this real-time data and identify any abnormal behavior patterns. By conducting long-term analysis of trends in data such as airflow, vibration, temperature, and humidity, the system identifies normal and abnormal behavior patterns, and proactively identifies potential safety hazards and malfunction risks.

[0098] S3. Based on the behavioral analysis results, a safety warning is issued, and a warning signal is obtained.

[0099] In this embodiment, the steps for issuing a security warning include:

[0100] Based on the behavioral analysis results, the system determines whether there are abnormal behavioral patterns;

[0101] An abnormal behavior pattern is detected, and a warning signal is generated based on a pattern recognition algorithm.

[0102] Compare behavioral patterns with historical data to determine whether they belong to known dangerous patterns;

[0103] If the condition is classified as a known hazard, an emergency alert will be issued and recommendations for appropriate safety measures will be provided.

[0104] Furthermore, the step of comparing behavioral patterns with historical data includes:

[0105] Data Collection and Preprocessing: The system collects real-time data from the fume hood, including key parameters such as airflow velocity, vibration range, temperature, and humidity. To ensure data accuracy and consistency, the system preprocesses this data, including noise reduction, standardization, and missing data imputation.

[0106] Behavioral Pattern Recognition: Through pattern recognition algorithms (such as K-means clustering or random forest algorithms), the system analyzes current sensor data and identifies possible behavioral patterns. For example, the system can detect airflow fluctuations, abnormal vibration amplitudes, and abnormal changes in temperature and humidity. Based on these characteristics, the system can determine whether the equipment is in normal working condition or whether there are potential safety hazards.

[0107] Historical data storage and management: Historical operational data and labeled behavioral patterns are stored in the system's historical database, and the historical data is labeled. This data includes behavioral patterns under normal operation, fault modes, and corresponding equipment status and environmental conditions. Through long-term accumulation and analysis, it has been clearly labeled and serves as a reference standard for the system to judge whether anomalies have occurred.

[0108] Pattern comparison and matching: After identifying the behavior pattern at the current moment, the system compares it with historical data, including matching based on feature similarity, time series analysis, or classification matching through machine learning methods. It calculates the similarity between the current pattern and historical patterns, and matches them with historical patterns to determine whether it is a known abnormal behavior.

[0109] Decision-making and feedback: Based on the comparison results, the system makes corresponding decisions. If the new behavior pattern highly matches anomaly patterns in historical data, the system will trigger a security alert or automatic adjustment measures. If the behavior pattern is similar to normal patterns in historical data, the system will maintain the current operating state and continue to monitor data changes.

[0110] Comparing behavioral patterns with historical data can help the system quickly identify and respond to anomalies, improve the security and stability of the device, and provide strong support for subsequent intelligent optimization decisions.

[0111] S4. Implement automatic control measures based on the early warning signal to obtain the control result.

[0112] In this embodiment, the steps for implementing automatic control measures include:

[0113] Real-time detection and early warning signals;

[0114] Upon detecting a warning signal, the system automatically activates preset safety control measures to prevent the fume hood from being in an unsafe operating state. These control measures include, but are not limited to, automatically adjusting the airflow rate, temperature, humidity, or equipment status of the fume hood, or temporarily shutting down the fume hood for inspection as needed.

[0115] Adaptive control technology (i.e., a control method that automatically adjusts its control strategy or parameters based on dynamic changes in the controlled object or environment to achieve better control results) is used to automatically adjust the control parameters of the fume hood based on feedback from safety control measures, ensuring that the equipment always operates within a safe range. This automatic control method can quickly respond to safety warnings, reducing the risks associated with human error and response delays.

[0116] S5. Based on the control results, optimize energy efficiency by using optimization algorithms to optimize control parameters, reduce energy consumption costs, improve production efficiency, and obtain an optimized solution.

[0117] In this embodiment, by using optimization algorithms (such as Particle Swarm Optimization (PSO) or Genetic Algorithm (GA), the system finely adjusts control parameters such as airflow, temperature, and humidity to ensure that energy consumption is minimized while ensuring equipment safety. The goal of optimization is to find a balance point that allows the equipment to maintain safe operation while maximizing energy efficiency. This process is applicable to fume hoods that operate for extended periods; energy efficiency optimization helps reduce energy costs and improve the overall economic efficiency of the equipment. Through intelligent energy efficiency control, the system can dynamically adjust the operation of the fume hood according to the needs of the laboratory or industrial environment, avoiding energy waste and improving production efficiency.

[0118] S6. Based on the optimization plan, long-term operational data feedback is obtained to get the optimized operational data.

[0119] In this embodiment, the steps for providing long-term operational data feedback include:

[0120] The system continuously monitors the optimized operating data, including the energy consumption, temperature, and humidity of the fume hoods;

[0121] The system continuously evaluates the persistence and reliability of the optimization effect based on changes in real-time data;

[0122] The system adjusts based on the evaluation results to ensure that the equipment operates in a safe and efficient state;

[0123] The long-term feedback mechanism enables the system to continuously optimize the control scheme based on factors such as environmental changes and equipment aging, maintaining optimal equipment performance. This process helps the system adapt to the needs of different experimental environments and ensures the long-term stable and efficient operation of the fume hood.

[0124] S7. Based on the operating data, further optimize the behavior patterns and energy efficiency to obtain the optimization results.

[0125] In this embodiment, the steps for further optimizing the behavior pattern and energy efficiency include:

[0126] Based on historical data and real-time feedback, the system continuously learns and improves its control algorithm by modifying algorithm weight coefficients through deep reinforcement learning technology with multi-factor dynamic adjustment, combined with behavioral pattern recognition and energy efficiency optimization. In this continuous adjustment process, the system not only optimizes the safety of equipment operation but also achieves maximum energy efficiency while ensuring safety. Deep reinforcement learning enables the system to gradually learn the optimal operating mode in uncertain environments, thereby improving control accuracy and responsiveness, making the operation of the fume hood more intelligent and efficient.

[0127] Furthermore, the deep reinforcement learning technique based on multi-factor dynamic adjustment includes:

[0128] Model building:

[0129] Construct the following reward function for training and optimization of reinforcement learning:

[0130] ;

[0131] ;

[0132] in:

[0133] Airflow velocity : Difference in vibration frequency Temperature and humidity changes are acquired through a sensor network, among which... Calculated using a weighted average method: For temperature, For humidity, Temperature weighting coefficient, Humidity weighting coefficient ;

[0134] , , : Reference parameters, defined in step S1;

[0135] The smaller the deviation in airflow speed, the greater the reward.

[0136] The smaller the deviation in vibration frequency, the greater the reward.

[0137] The smaller the deviation in temperature and humidity, the greater the reward.

[0138] Energy efficiency score: the higher the score, the more energy-efficient the system.

[0139] Ideal power consumption;

[0140] Actual power consumption;

[0141] , , , Weighting coefficient;

[0142] Model usage:

[0143] Input data: Input data and calculate the deviation;

[0144] Calculate energy efficiency score: The system simultaneously calculates the energy efficiency score of the current device. It evaluates the energy efficiency status of the device by the ratio between the actual power consumption and the ideal power consumption. This score will play a positive role in the reward function, thereby prompting the system to adjust parameters such as airflow, temperature and humidity according to the actual situation to optimize energy efficiency.

[0145] Obtaining the optimal control strategy: Based on the current state and reward function, the system selects the best control action through reinforcement learning, such as adjusting airflow speed, vibration frequency, and temperature and humidity settings. Through long-term interaction and feedback (i.e., adjusting the strategy based on the reward of actual execution results), it gradually learns an optimal control strategy, enabling the equipment to maintain the best safety and energy efficiency under different operating conditions.

[0146] Adaptive learning: As the environment changes or the equipment's operating time increases, the system continues to adjust its control actions to enable the equipment to adaptively respond to new operating states (e.g., fluctuations in airflow speed, changes in vibration frequency, sudden changes in temperature and humidity).

[0147] Selecting the optimal control action through reinforcement learning:

[0148] ;

[0149] in, Indicates state, ;

[0150] This refers to actions, including adjusting airflow speed, adjusting vibration frequency, and adjusting temperature and humidity.

[0151] This represents the learning rate, which controls the speed at which the Q-value is updated.

[0152] This represents the discount factor, which is the weight of future rewards;

[0153] Indicates a new state Next, select all possible actions. The maximum Q value represents the optimal reward in the future, and the action at this point is the optimal control action.

[0154] This method, based on data on airflow velocity, vibration range, and temperature and humidity changes, combined with the deviation of the baseline value, designs a dynamically adjusted reward function. During long-term operation, the reinforcement learning system will further optimize and improve its control strategy through accumulated rewards and feedback, thereby enabling the equipment to maintain high efficiency while maintaining good safety and stability.

[0155] S8. Based on the optimization results, automatically upgrade and self-adjust the system.

[0156] In this embodiment, the steps for automated system upgrades and self-adjustments include:

[0157] Based on the optimization results, the system continuously updates its control strategy through deep learning and transfer learning technologies, thereby self-adjusting and upgrading. Over time, the system will gradually improve its recognition and control accuracy and can automatically adapt to changes in different environments. This self-adjustment and upgrading capability ensures that the system can flexibly respond to new situations or problems, improve the safety and energy efficiency of the equipment, and still provide the optimal control strategy when the operating environment changes significantly.

[0158] Furthermore, the deep learning and transfer learning techniques include:

[0159] Model building:

[0160] By combining deep learning and transfer learning, the following model is constructed:

[0161] ;

[0162] in:

[0163] Indicates time The control output or decision result at any given moment, for example, This could indicate an adjustment command, such as a control command like "the airflow speed should be increased by 10%, and the vibration amplitude reduced by 5%";

[0164] : This is the weighting coefficient between deep learning and transfer learning, which controls the relative influence of the two in the final output. By adjusting this parameter, the system can balance the effects of deep learning and transfer learning, ensuring that the model can be accurately adjusted in the current environment and quickly adapt to the new environment.

[0165] Environmental differences are derived by comparing data acquired by sensor networks in the new and old environments, representing the differences between the current and original environments.

[0166] A deep neural network model represents an input based on airflow speed, vibration range, temperature and humidity change data, and baseline deviation. The model outputs a control strategy, which can be understood as the deep learning model's real-time optimization of the device.

[0167] ;

[0168] in, It's weight. It is a bias term;

[0169] : Represents the transfer learning function, used to improve the system's adaptability to new environments. By fine-tuning the parameters of the deep neural network model, the control strategy can be adapted to the new environment;

[0170] ;

[0171] in, and These are the fine-tuning parameters in the transfer learning process;

[0172] Model usage:

[0173] Obtaining the initial control strategy: Calculate based on the collected data and the deviation from the benchmark value, input the data into the deep neural network model, and the model outputs an initial control strategy;

[0174] Deep learning model optimization: Deep neural networks learn how to adjust device parameters based on input data through weight matrices and biases to achieve optimized control. The model is continuously trained using historical data to gradually improve the accuracy of the control strategy.

[0175] Adjustments in transfer learning: When equipment is migrated to a new experimental environment, transfer learning fine-tunes the parameters of the deep neural network based on the differences between the original and new environments. This process helps the system quickly adapt to the new environment without requiring a large amount of new data.

[0176] Adaptation and feedback: The outputs of deep learning and transfer learning models will jointly determine the control decisions of the device. Whenever new data is input into the system, the model will continuously update the output and optimize the parameters to ensure that the device always operates in the best safety and energy efficiency state.

[0177] Example 2:

[0178] This embodiment also provides a computer device applicable to a behavior analysis and control method for safe operation of a fume hood, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the behavior analysis and control method for safe operation of a fume hood as proposed in the above embodiment.

[0179] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements a behavior analysis and control method for safe operation of a fume hood as proposed in the above embodiments.

[0180] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0181] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0182] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0183] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0184] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0185] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A behavior analysis and control method for safe operation of a fume hood, characterized in that, Includes the following steps: S1. Define the safe operating standards and real-time monitoring system for fume hoods to obtain fume hood operation data; S2. Based on the fume hood operation data, use machine learning algorithms to perform behavioral pattern analysis and obtain behavioral analysis results; S3. Based on the behavioral analysis results, a safety warning is issued, and a warning signal is obtained; S4. Implement automatic control measures based on the early warning signal to obtain the control result; S5. Based on the control results, optimize the energy efficiency by using an optimization algorithm to optimize the control parameters and obtain the optimization scheme; S6. Based on the optimization plan, long-term operational data feedback is obtained to get the optimized operational data; S7. Based on the operational data, further optimize the behavior patterns and energy efficiency to obtain the optimization results; Based on historical data and real-time feedback, the system continuously learns and improves the control algorithm by modifying the algorithm weight coefficients, using deep reinforcement learning technology based on multi-factor dynamic adjustment, combined with behavior analysis results and energy efficiency optimization, thereby further optimizing system performance. Among them, deep reinforcement learning techniques based on multi-factor dynamic adjustment include: Model building: Construct the following reward function for training and optimization of reinforcement learning: ； ; in: Airflow velocity : Difference in vibration frequency Temperature and humidity changes are acquired through a sensor network, among which... Calculated using a weighted average method: For temperature, For humidity, Temperature weighting coefficient, Humidity weighting coefficient ; , , Baseline parameters; The smaller the deviation in airflow speed, the greater the reward. The smaller the deviation in vibration frequency, the greater the reward. The smaller the deviation in temperature and humidity, the greater the reward. Energy efficiency score: the higher the score, the more energy-efficient the system. Ideal power consumption; Actual power consumption; , , , Weighting coefficient; S8. Based on the optimization results, automatically upgrade and self-adjust the system.

2. The behavior analysis and control method for safe operation of a fume hood according to claim 1, characterized in that, The safe operation standards and real-time monitoring system for fume hoods in S1 include: Define the safe operating standards for the fume hood and obtain the standard parameters; The status of the fume hood is monitored in real time through a sensor network to obtain fume hood operation data; Using Internet of Things (IoT) technology, real-time monitoring data is transmitted to a data processing center; Determine whether the fume hood is within the safe reference range; If the safety baseline is exceeded, the abnormal event will be recorded and an alarm signal will be generated.

3. The behavior analysis and control method for safe operation of a fume hood according to claim 1, characterized in that, The steps for providing a security warning in S3 include: Based on the behavioral analysis results, the system determines whether there are abnormal behavioral patterns; An abnormal behavior pattern is detected, and a warning signal is generated based on a pattern recognition algorithm. Compare behavioral patterns with historical data to determine whether they belong to known dangerous patterns; If the condition is classified as a known hazard, an emergency alert will be issued and recommendations for appropriate safety measures will be provided.

4. The behavior analysis and control method for safe operation of a fume hood according to claim 3, characterized in that, The steps for comparing behavioral patterns with historical data include: Data collection and preprocessing: Real-time monitoring data is obtained through sensor networks, and the data is preprocessed to obtain preprocessed data; Behavioral pattern recognition: The system analyzes preprocessed data and identifies behavioral patterns using machine learning algorithms; Storage and management of historical data: Store historical operation data and labeled behavior patterns in the system's historical database, and label historical data as a reference standard; Pattern comparison and matching: After identifying the behavior pattern at the current moment, the system compares it with historical data, calculates the similarity between the current pattern and historical patterns, and matches it with historical patterns based on the similarity. Decision-making and feedback: If a new behavior pattern is highly similar to an abnormal pattern, a safety warning or automatic adjustment measure will be triggered; if it is highly similar to a normal pattern, the current operating status will be maintained and data changes will continue to be monitored.

5. The behavior analysis and control method for safe operation of a fume hood according to claim 1, characterized in that, The steps for implementing automatic control measures in S4 include: Real-time detection and early warning signals; Upon detecting a warning signal, the system automatically activates preset safety control measures; By utilizing adaptive control technology, adjustments are made based on feedback from safety control measures to enhance the system's ability to respond to abnormal situations.

6. The behavior analysis and control method for safe operation of a fume hood according to claim 1, characterized in that, The steps for long-term operational data feedback in S6 include: The system continuously monitors the optimized operational data; The system continuously evaluates the persistence and reliability of the optimization effect based on changes in real-time data; The system adjusts based on the evaluation results to ensure that the equipment operates in a safe and efficient state.

7. The behavior analysis and control method for safe operation of a fume hood according to claim 1, characterized in that, The steps for automated system upgrades and self-adjustments in S8 include: Based on the optimization results, the system continuously updates its control strategy through deep learning and transfer learning techniques, thereby adjusting and upgrading itself.

8. The behavior analysis and control method for safe operation of a fume hood according to claim 7, characterized in that, The deep learning and transfer learning techniques include: Model building: By combining deep learning and transfer learning, the following model is constructed: ； in: In time Time-based control output; : These are the weighting coefficients between deep learning and transfer learning, controlling the relative influence of the two in the final output, and are adjusted according to the actual results; Environmental differences are derived by comparing data acquired by sensor networks in the new and old environments. Deep learning functions output a control strategy based on the input data and continuously optimize it; : Transfer learning function, adjust the control strategy to adapt it to the new environment.

Images