Intelligent ultraviolet disinfection method and system

Abstract

This invention discloses an intelligent ultraviolet disinfection method and system. The system includes a multi-band ultraviolet light source component, an environmental and object sensing component, and an intelligent control component. The intelligent control component, through a built-in machine learning prediction model, predicts the disinfection effect under different bands based on real-time environmental parameters, the characteristics of the object to be disinfected, and the presence of personnel, and dynamically generates the optimal control strategy. This invention enables continuous and safe operation in a human-machine coexistence environment while ensuring that the disinfection effect meets the standards, reducing operating energy consumption and equipment wear, and is suitable for disinfection applications in multiple scenarios, including water, air, and object surfaces.

Description

Technical Field

[0001] This invention belongs to the field of disinfection and environmental hygiene technology, and particularly relates to an intelligent ultraviolet disinfection method and system. Background Technology

[0002] Ultraviolet (UV) disinfection has been widely used in drinking water treatment, wastewater reuse, air purification, and surface disinfection due to its advantages such as no need for chemical agents, rapid reaction, and no disinfection byproducts. Existing UV disinfection equipment typically inactivates microorganisms by irradiating the target medium with ultraviolet light of a specific wavelength, damaging the nucleic acids or key biomolecules of the microorganisms.

[0003] In existing water treatment systems, ultraviolet (UV) disinfection equipment mostly uses low-pressure or medium-pressure mercury lamps, with output wavelengths primarily concentrated around 254 nm. The irradiation dose is set based on empirical models or industry standards to ensure effective microbial inactivation. In the fields of air and surface disinfection, UV disinfection devices also mostly operate with fixed wavelengths, fixed power, or fixed irradiation times, and their disinfection strategies are usually preset at the factory.

[0004] However, studies have shown that different ultraviolet wavelengths differ significantly in terms of disinfection mechanisms, safety, and applicable scenarios. For example, the 254 nm band has a strong destructive ability against microbial nucleic acids and high inactivation efficiency, but it poses potential hazards to human skin and eyes, and therefore is usually only suitable for unoccupied or enclosed environments. On the other hand, far-ultraviolet bands such as 222 nm are considered relatively safe for humans under certain dosage conditions and are more suitable for environments where people may be present, but their inactivation efficiency and dosage requirements for different microorganisms still vary considerably.

[0005] Existing ultraviolet disinfection systems typically use only a single wavelength light source, or statically select a certain wavelength band as a compromise during the system design phase. This makes it difficult to flexibly switch between different wavelengths based on changes in actual usage scenarios (such as whether people are present), the type of object to be disinfected, and environmental conditions. This single or static configuration often leads to the following problems in practical applications: in unmanned environments, efficient wavelengths cannot be fully utilized to reduce operating time and energy consumption; in manned environments, additional power reduction or extended disinfection cycles are required, thus affecting disinfection efficiency.

[0006] Furthermore, the effectiveness of ultraviolet disinfection is not only related to wavelength but also influenced by a combination of factors, including irradiation dose, initial microbial concentration, ambient temperature, pH value, and media characteristics. Existing systems typically address these uncertainties by setting a large safety margin, employing a high irradiation dose under various conditions to ensure disinfection meets standards. While this approach improves the reliability of disinfection, it also significantly increases energy consumption, accelerates lamp aging, and reduces the overall system operating efficiency.

[0007] In the field of automated equipment or ultraviolet disinfection robots, improvements in existing technologies mainly focus on the movement control, path planning, and obstacle avoidance algorithms of the equipment, while the core disinfection module usually still operates as an independent execution unit with fixed parameters. These devices lack the ability to adjust the disinfection process based on real-time feedback from environmental bioburden or human activity. For example, when personnel are detected, they typically take passive protective measures by completely cutting off the light source, leading to disinfection interruptions. This "hard" operating mode, lacking adaptive band switching and dynamic dosage adjustment, makes it difficult to achieve a dynamic balance between disinfection effectiveness, operational continuity, and energy safety in complex human-machine coexistence environments.

[0008] Therefore, how to achieve reasonable configuration and switching of different wavelength light sources in ultraviolet disinfection systems, and optimize the ultraviolet irradiation dose in combination with the characteristics of the disinfection object and operating conditions, so as to ensure the disinfection effect while taking into account personnel safety and energy efficiency, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0009] Purpose of the invention: This invention provides an intelligent ultraviolet disinfection method and system, aiming to solve the technical problems of existing ultraviolet disinfection technology in complex application scenarios, especially in scenarios with dynamic changes in manned / unmanned environments, which are difficult to simultaneously achieve high-intensity sterilization and personnel safety protection, and have high operating energy consumption.

[0010] Technical solution: This invention provides an intelligent ultraviolet disinfection method, comprising:

[0011] Collect historical datasets, including historical disinfection data and historical inactivation rates. The historical disinfection data includes ultraviolet band, radiation dose, disinfection environmental parameters, and contact time. Preprocess the historical datasets to obtain a set of historical feature vectors.

[0012] A disinfection effect prediction model is built. The optimization objective is to minimize the mean square error between the predicted inactivation rate of the disinfection effect prediction model and the historical inactivation rate in the historical dataset on the historical feature vector set. The disinfection effect prediction model is trained iteratively to obtain a well-trained disinfection effect prediction model.

[0013] The current disinfection environment parameters of the disinfection target are collected. Based on the permutation and combination of the allowable range of ultraviolet band and the allowable range of radiation dose in the current disinfection environment parameters, the current parameter combination range of the disinfection target is formed. Each current parameter combination is concatenated with the current disinfection environment parameters to construct the current dataset of the disinfection target. The current feature vector set is obtained by preprocessing the current dataset.

[0014] Input the current feature vector set into the trained disinfection effect prediction model, predict the inactivation rate and action time based on each current feature vector, statistically analyze the ultraviolet band, radiation dose and corresponding action time that meet the preset target inactivation rate, and output the range of ultraviolet band, radiation dose and action time that the current feature vector set can meet the target inactivation rate, forming the optimal parameter range.

[0015] Based on the optimal parameter range, and combined with the preset safety threshold and the constraints of ultraviolet disinfection operation, the optimal parameter range is further defined; according to the preset ultraviolet disinfection rules, an executable parameter combination is selected from the defined optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and action time.

[0016] During the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed, the actual disinfection environment parameters and the actual inactivation rate are obtained; if the actual inactivation rate reaches the target inactivation rate, the disinfection is stopped; if the actual disinfection environment parameters change or the actual inactivation rate is lower than the target inactivation rate, the ultraviolet disinfection control strategy is regenerated.

[0017] Furthermore, the historical dataset includes at least one of the following datasets:

[0018] (1) Historical disinfection data and historical inactivation rates recorded in ultraviolet disinfection experiments conducted under controlled experimental conditions;

[0019] (2) Historical disinfection data and historical inactivation rate of ultraviolet disinfection experiments recorded in publicly available literature or industry databases.

[0020] Furthermore, the disinfection environment parameters include disinfection environment type, disinfection environment attribute parameters, and disinfection environment status parameters; the disinfection environment attribute parameters include the allowable range of ultraviolet band, the allowable range of radiation dose, the size parameters, and the bacterial concentration of the disinfection environment; the disinfection environment status parameters include whether there are people in the disinfection environment, the temperature, humidity, and pH of the disinfection environment.

[0021] Furthermore, the preprocessing includes at least one of data cleaning, outlier removal, normalization, and feature encoding.

[0022] Furthermore, the iterative training of the disinfection effect prediction model includes:

[0023] The historical feature vector set is divided into a training set and a validation set according to a preset ratio;

[0024] Set the maximum number of iterations, learning rate, and initial parameters for the disinfection effect prediction model;

[0025] In each iteration, the training set is input into the disinfection effect prediction model to calculate the predicted inactivation rate;

[0026] Based on the predicted inactivation rate and the historical inactivation rate, the mean square error of the disinfection effect prediction model is calculated, and the mean square error is used as the loss function.

[0027] Calculate the gradient of the loss function with respect to the disinfection effect prediction model, and update the parameters of the disinfection effect prediction model based on the gradient and the learning rate.

[0028] After each or several rounds of parameter updates, the predictive performance of the disinfection effect prediction model is calculated based on the validation set. If the predictive performance reaches the preset accuracy or the number of iterations reaches the maximum number of iterations, training stops; otherwise, the next round of iteration training continues. The predictive performance metrics include R0. 2 RMSE.

[0029] Furthermore, the iterative training of the disinfection effect prediction model, and the updating of the parameters of the disinfection effect prediction model, include:

[0030] Calculate the loss function with respect to the parameters The derivative of the loss function is used as the gradient, and the formula for the loss function is:

[0031] ;

[0032] in, For loss function, For parameters, Let i be the actual inactivation rate of the i-th data set. Let i be the predicted inactivation rate of the i-th data set. The amount of data in the training set;

[0033] The parameters of the disinfection effect prediction model are updated based on the gradient and learning rate, using the following formula:

[0034] ;

[0035] in, These are the parameters for the current disinfection effectiveness prediction model. For the updated model parameters, For learning rate, For the loss function with respect to the parameters The derivative of .

[0036] Furthermore, the preset rules for ultraviolet disinfection include energy consumption priority rules, safety priority rules, efficiency priority rules, and equipment lifespan protection rules. The energy consumption priority rule requires selecting the parameter combination with the lowest total power consumption within the optimal parameter range. The safety priority rule requires that when people are present in the disinfection environment, the ultraviolet wavelength should be forced to 222nm, and parameter combinations containing the 254nm ultraviolet wavelength should be excluded. The efficiency priority rule requires that when no one is present in the disinfection environment and time is limited, the combination with the shortest action time should be selected. The equipment lifespan protection rule requires that parameter combinations that can avoid frequent switching of ultraviolet lamps or long-term full-load disinfection should be selected.

[0037] The present invention also provides an intelligent ultraviolet disinfection system, comprising: an environment and object sensing component and an intelligent control component;

[0038] The environment and object perception component is used to collect historical datasets, including historical disinfection data and historical inactivation rates. The historical disinfection data includes ultraviolet band, radiation dose, disinfection environment parameters, and actual action time. It is also used to collect the current disinfection environment parameters of the disinfection target. Furthermore, it is used to obtain the actual disinfection environment parameters and actual inactivation rate during the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed.

[0039] The intelligent control component includes a data storage and analysis module and a control decision module; the data storage and analysis module includes a data preprocessing module, a machine learning prediction module, and an adaptive adjustment module.

[0040] The data preprocessing module is used to preprocess the historical dataset to obtain a historical feature vector set. It is also used to form the current parameter combination range of the disinfection target according to the permutation and combination of the allowable range of ultraviolet band and the allowable range of radiation dose in the current disinfection environment parameters. Each current parameter combination is concatenated with the current disinfection environment parameters to construct the current dataset. The current dataset is then preprocessed to obtain the current feature vector set.

[0041] The machine learning prediction module is used to build a disinfection effect prediction model. On the historical feature vector set, the optimization objective is to minimize the mean square error between the predicted inactivation rate of the disinfection effect prediction model and the historical inactivation rate in the historical dataset. The disinfection effect prediction model is iteratively trained to obtain a well-trained disinfection effect prediction model. It is also used to predict the inactivation rate and action time for each current feature vector in the current feature vector set based on the well-trained disinfection effect prediction model.

[0042] The control decision module is used to statistically analyze the ultraviolet bands, radiation doses, and corresponding exposure times that meet the preset target inactivation rate, obtain the range of ultraviolet bands, radiation doses, and exposure times that the current feature vector set can satisfy the target inactivation rate, and form the optimal parameter range; it is also used to further limit the optimal parameter range based on the optimal parameter range, combined with preset safety thresholds and constraints of ultraviolet disinfection operation; according to preset ultraviolet disinfection rules, it selects an executable parameter combination from the limited optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and exposure time.

[0043] The adaptive adjustment module is used to obtain the actual disinfection environment parameters and the actual inactivation rate during the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed; if the actual inactivation rate reaches the target inactivation rate, the disinfection is stopped; if the actual disinfection environment parameters change or the actual inactivation rate is lower than the target inactivation rate, the ultraviolet disinfection control strategy is regenerated.

[0044] Furthermore, it also includes functional control components, ultraviolet irradiation components, communication components, and power supply components;

[0045] The functional control component is used to set the working mode, start / stop status and ultraviolet disinfection control parameters of the ultraviolet disinfection system; the ultraviolet disinfection control parameters include ultraviolet band, irradiation dose and action time.

[0046] The ultraviolet irradiation component is used to adjust the ultraviolet band, irradiation dose and action time of the ultraviolet emission source according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters, so as to perform the ultraviolet disinfection task.

[0047] The communication component is used for data interaction between the ultraviolet disinfection system and external terminals or networks.

[0048] The power supply component is used to provide stable power to all components of the system.

[0049] Furthermore, the intelligent control component also includes an information transceiver module and a power management module; the information transceiver module is used for data input and output; the power management module is used to provide stable power to the intelligent control component.

[0050] The ultraviolet irradiation assembly includes an ultraviolet light source unit, a band selection / switching module, and a dose adjustment module. The ultraviolet light source unit generates ultraviolet radiation and includes at least one ultraviolet emission source that emits ultraviolet radiation at a predetermined peak wavelength when powered on. The band selection / switching module physically switches the ultraviolet band by switching the power supply circuit of a specific wavelength emission source on and off, according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters. The dose adjustment module adjusts the ultraviolet radiation intensity by adjusting the driving voltage or duty cycle, according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters. The ultraviolet irradiation assembly includes at least two sets of ultraviolet light source units with different peak wavelength emission sources.

[0051] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0052] 1. Achieving intelligent multi-band switching based on spectral characteristics, breaking the limitations of traditional single-wavelength methods and improving energy efficiency. This invention overcomes the limitations of traditional equipment that relies solely on a single wavelength. Through the configuration of multi-band ultraviolet light source components, under intelligent control, it dynamically selects the optimal band, such as 254nm or 222nm, based on the differences in spectral sensitivity of the target microorganisms and environmental conditions. While ensuring that the inactivation efficiency meets the standards, this invention can automatically match the operating mode with the lowest energy consumption, avoiding energy waste and lamp life loss caused by excessive irradiation in traditional equipment.

[0053] 2. This invention addresses the challenge of disinfection under excessive radiation conditions, ensuring operational continuity while prioritizing human safety. Unlike existing technologies that passively shut down upon detecting personnel, this invention utilizes the collaborative work of human body sensing and environmental perception components. Upon detecting personnel, it can switch to far-ultraviolet light, such as the 222nm band, within milliseconds, ensuring safety for human tissue. This mechanism not only eliminates the health risks posed by ultraviolet leakage but also ensures the continuity and integrity of disinfection operations in scenarios with frequent personnel movement, such as hospital wards and train carriages, eliminating the need for frequent interruptions due to personnel movement.

[0054] 3. Introducing machine learning prediction models to achieve a leap from experience-based control to precise predictive control. This invention integrates a machine learning prediction module, capable of deep feature extraction of water quality parameters (such as turbidity and transmittance) or environmental characteristics, and accurately predicting the UV dose-response relationship under current conditions based on a pre-trained model. Furthermore, it performs feedforward precise adjustments based on irradiance intensity and time, solving the problem of incomplete disinfection or dose redundancy caused by the inability of existing technologies to cope with complex environmental fluctuations, thus enhancing adaptability and stability under different water qualities and complex spatial environments.

[0055] 4. Supports full lifecycle digital management and remote operation and maintenance, reducing operation and maintenance costs. This invention utilizes an IoT architecture built through a communication module, enabling the system to upload operating bands, cumulative dose, and fault status to a cloud server in real time. Users can not only remotely monitor and adjust parameters but also analyze lamp aging trends and disinfection effect distribution based on historical data, providing data support for preventative maintenance and process optimization, significantly improving system management efficiency.

[0056] 5. Possessing a versatile architecture applicable to multiple scenarios involving water, air, and materials, this invention has broad application prospects. The hardware and software architecture design of this invention fully considers the disinfection characteristics of different media, making it suitable for both drinking water and reclaimed water treatment processes with extremely high dosage accuracy requirements, as well as indoor air and object surface disinfection scenarios with stringent safety requirements. This high degree of versatility facilitates the rapid promotion and industrialization of this technical solution in multiple fields such as healthcare, public transportation, and water engineering. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the system structure of the present invention.

[0058] Figure 2 This is a schematic diagram of the intelligent control component in this invention.

[0059] Figure 3 This is a schematic diagram of the structure of the ultraviolet irradiation component in this invention.

[0060] Figure 4 This is a schematic diagram illustrating the prediction effect of the machine learning prediction module in this invention.

[0061] Figure 5 This is a schematic diagram of the adaptive intelligent monitoring visualization interface of the present invention. Detailed Implementation

[0062] Example 1

[0063] The intelligent ultraviolet disinfection method of the present invention includes:

[0064] Collect historical datasets, including historical disinfection data and historical inactivation rates; the historical datasets include at least one of the following datasets:

[0065] (1) Historical disinfection data and historical inactivation rates recorded in ultraviolet disinfection experiments conducted under controlled experimental conditions;

[0066] (2) Historical disinfection data and historical inactivation rate of ultraviolet disinfection experiments recorded in publicly available literature or industry databases.

[0067] The historical disinfection data includes historical data on the operation status of the disinfection system and environmental parameters during the disinfection period.

[0068] The disinfection system's operating status data includes ultraviolet band, radiation dose, contact time, light intensity feedback, energy consumption, etc.

[0069] The disinfection environment parameters include the type of disinfection environment (such as air, surface, or water quality), the attribute parameters of the disinfection environment, and the state parameters of the disinfection environment; the attribute parameters of the disinfection environment include the allowable range of ultraviolet light band, the allowable range of radiation dose, and the size parameters of the disinfection environment (such as material, size, and bioburden); the state parameters of the disinfection environment include whether there are people in the disinfection environment, the temperature, humidity, pH, air flow, bacterial concentration, and water quality parameters (such as turbidity and light transmittance).

[0070] The historical dataset is preprocessed to obtain a set of historical feature vectors; the preprocessing includes data cleaning, outlier removal, normalization, and feature encoding.

[0071] The historical feature vector set is divided into a training set and a validation set according to a preset ratio; a disinfection effect prediction model is built based on the CatBoost model, and the maximum number of iterations, learning rate, and initial parameters of the disinfection effect prediction model are set; in each iteration, the training set is input into the disinfection effect prediction model to calculate the predicted inactivation rate;

[0072] Based on the predicted inactivation rate and the historical inactivation rate, the mean square error of the disinfection effect prediction model is calculated. This mean square error is then used as the loss function, as shown in the formula:

[0073] ;

[0074] in, For loss function, For parameters, Let i be the actual inactivation rate of the i-th data set. Let i be the predicted inactivation rate of the i-th data set. The amount of data in the training set;

[0075] Calculate the gradient of the loss function relative to the disinfection effect prediction model, and update the parameters of the disinfection effect prediction model based on the gradient and the learning rate. The formula is as follows:

[0076] ;

[0077] in, These are the parameters for the current disinfection effectiveness prediction model. For the updated model parameters, For learning rate, For the loss function with respect to the parameters The derivative of indicates the direction in which the error decreases the most.

[0078] After each or several rounds of parameter updates, the prediction performance of the disinfection effect prediction model is calculated based on the validation set. If the prediction performance reaches a preset accuracy (e.g., RMSE < 0.5) or the number of iterations reaches the maximum number of iterations, training stops, and a trained disinfection effect prediction model is obtained; otherwise, the next round of iteration training continues. The prediction performance index includes R... 2 RMSE.

[0079] This invention also includes an external validation set, which is a dataset that was not used in the training process. The trained disinfection effect prediction model is validated on this external validation set. Figure 4 As shown, in the external validation set, R 2 =0.778, RMSE=0.535.

[0080] Collect the current disinfection environment parameters of the disinfection target. Based on the permutations (such as all permutations) of the allowable range of ultraviolet band and allowable range of radiation dose in the current disinfection environment parameters, form the current parameter combination range of the disinfection target. Concatenate each current parameter combination with the current disinfection environment parameters to construct the current dataset of the disinfection target. By preprocessing the current dataset, obtain the current feature vector set.

[0081] Input the current feature vector set into the trained disinfection effect prediction model, predict the inactivation rate and action time based on each current feature vector, statistically analyze the ultraviolet band, radiation dose and corresponding action time that meet the preset target inactivation rate (e.g. 99.99%), and output the ultraviolet band range, radiation dose range and action time range that the current feature vector set can meet the target inactivation rate, forming the optimal parameter range.

[0082] Based on the optimal parameter range, combined with preset safety thresholds (such as human exposure limits) and constraints of ultraviolet disinfection operation, the optimal parameter range is further defined; according to preset ultraviolet disinfection rules, an executable parameter combination is selected from the defined optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and exposure time.

[0083] The preset rules for ultraviolet disinfection include energy consumption priority rules, safety priority rules, efficiency priority rules, and equipment life protection rules.

[0084] The energy consumption priority rule requires that the parameter combination with the lowest total power consumption (such as low power or short time) be selected first within the optimal parameter range;

[0085] The safety priority rule requires that when people are present in the disinfection environment, the ultraviolet band must be 222nm, and parameter combinations that include the 254nm ultraviolet band must be excluded.

[0086] The efficiency priority rule requires that, in the case of an unoccupied disinfection environment and when time is of the essence, the combination with the shortest action time (i.e., the highest dose rate) should be selected first.

[0087] The equipment life protection rules require prioritizing parameter combinations that avoid frequent switching of ultraviolet lamps or long-term full-load disinfection.

[0088] During the implementation of the ultraviolet disinfection control strategy or after the disinfection task is completed, the actual disinfection environment parameters are obtained, and the actual inactivation rate is calculated based on the initial bacterial concentration and the bacterial concentration after ultraviolet lamp inactivation.

[0089] If the actual inactivation rate reaches the target inactivation rate, then disinfection should be stopped;

[0090] If the actual inactivation rate is lower than the target inactivation rate or the actual disinfection environment parameters change, the ultraviolet disinfection control strategy will be regenerated. If the temperature, pH, humidity of the disinfection environment change, the turbidity of the water suddenly increases, the presence of personnel causes the safety status of the disinfection area to switch from "unmanned" to "manned", the range of allowed ultraviolet bands changes, or the maximum allowable irradiation dose or power limit changes, the original ultraviolet control strategy will be modified, such as increasing the power or switching the band, to achieve closed-loop control.

[0091] Example 2

[0092] like Figures 1 to 3 As shown in Embodiment 1, the intelligent ultraviolet disinfection system 10 of the present invention includes a functional control component 101, an environment and object sensing component 102, an intelligent control component 103, an ultraviolet irradiation component 104, a communication component 105, and a power supply component 106; the above components work together under the unified coordination of the intelligent control component 103 to achieve intelligent control of the disinfection process.

[0093] The functional control component 101 is used to set the working mode, start / stop status, and ultraviolet disinfection control parameters of the intelligent ultraviolet disinfection system; the working mode includes automatic / intelligent operation mode, manual / human intervention mode, powerful disinfection mode, safe coexistence mode, standby / energy saving mode, and maintenance / calibration mode; the ultraviolet disinfection control parameters include ultraviolet band, irradiation dose, and action time.

[0094] The environment and object perception component 102 is used to collect historical datasets, including historical disinfection data and actual inactivation rates. The historical disinfection data includes ultraviolet band, radiation dose, disinfection environment parameters, and actual action time. It is also used to collect the disinfection environment parameters of the current disinfection target. Furthermore, it is used to obtain the actual disinfection environment parameters and actual inactivation rate during the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed.

[0095] The historical dataset includes at least one of the following datasets:

[0096] (1) Historical disinfection data and historical inactivation rates recorded in ultraviolet disinfection experiments conducted under controlled experimental conditions;

[0097] (2) Historical disinfection data and historical inactivation rate of ultraviolet disinfection experiments recorded in publicly available literature or industry databases.

[0098] The historical disinfection data includes historical data on the operation status of the disinfection system and environmental parameters during the disinfection period.

[0099] The disinfection system's operating status data includes ultraviolet band, radiation dose, contact time, light intensity feedback, energy consumption, etc.

[0100] The disinfection environment parameters include the type of disinfection environment (such as air, surface, or water quality), the attribute parameters of the disinfection environment, and the state parameters of the disinfection environment; the attribute parameters of the disinfection environment include the allowable range of ultraviolet light band, the allowable range of radiation dose, and the size parameters of the disinfection environment (such as material, size, and bioburden); the state parameters of the disinfection environment include whether there are people in the disinfection environment, the temperature, humidity, pH, air flow, bacterial concentration, and water quality parameters (such as turbidity and light transmittance).

[0101] The intelligent control component 103 includes an information transceiver module 1031, a data storage and analysis module 1032, a control decision module 1033, and a power management module 1034; the data storage and analysis module 1032 includes a data preprocessing module 10321, a machine learning prediction module 10322, and an adaptive adjustment module 10323.

[0102] The information transceiver module 1031 is used for data input and output;

[0103] The data preprocessing module 10321 is used to preprocess the historical dataset to obtain a historical feature vector set; it is also used to form the current parameter combination range of the disinfection target according to the permutation and combination (such as full permutation) of the allowable range of ultraviolet band and allowable range of radiation dose in the current disinfection environment parameters, and to concatenate each current parameter combination with the current disinfection environment parameters to construct the current dataset; and to preprocess the current dataset to obtain the current feature vector set; the preprocessing includes data cleaning, outlier removal, normalization processing and feature encoding.

[0104] The machine learning prediction module 10322 is used to build a disinfection effect prediction model based on the CatBoost model. On the historical feature vector set, the optimization objective is to minimize the mean squared error between the predicted inactivation rate of the disinfection effect prediction model and the historical inactivation rate in the historical dataset. The model is iteratively trained to obtain a trained disinfection effect prediction model. Specifically, this includes: dividing the historical feature vector set into a training set and a validation set according to a preset ratio; building the disinfection effect prediction model based on the CatBoost model, setting the maximum number of iterations, the learning rate, and the initial parameters of the disinfection effect prediction model; and in each iteration, inputting the training set into the disinfection effect prediction model to calculate the predicted inactivation rate.

[0105] Based on the predicted inactivation rate and the historical inactivation rate, the mean square error of the disinfection effect prediction model is calculated. This mean square error is then used as the loss function, as shown in the formula:

[0106] ;

[0107] in, For loss function, For parameters, Let i be the actual inactivation rate of the i-th data set. Let i be the predicted inactivation rate of the i-th data set. The amount of data in the training set;

[0108] Calculate the gradient of the loss function relative to the disinfection effect prediction model, and update the parameters of the disinfection effect prediction model based on the gradient and the learning rate. The formula is as follows:

[0109] ;

[0110] in, These are the parameters for the current disinfection effectiveness prediction model. For the updated model parameters, For learning rate, For the loss function with respect to the parameters The derivative of indicates the direction in which the error decreases the most.

[0111] After each or several rounds of parameter updates, the prediction performance of the disinfection effect prediction model is calculated based on the validation set. If the prediction performance reaches a preset accuracy (e.g., RMSE < 0.5) or the number of iterations reaches the maximum number of iterations, training stops, and a trained disinfection effect prediction model is obtained; otherwise, the next round of iteration training continues. The prediction performance index includes R... 2 RMSE.

[0112] This invention also includes an external validation set, which is a dataset that was not used in the training process. The trained disinfection effect prediction model is validated on this external validation set. Figure 4As shown, in the external validation set, R 2 =0.778, RMSE=0.535.

[0113] The machine learning prediction module 10322 is also used to predict the inactivation rate and action time for each current feature vector in the current feature vector set based on the trained disinfection effect prediction model.

[0114] The control decision module 1033 is used to statistically analyze the ultraviolet bands, radiation doses, and corresponding action times that meet the preset target inactivation rate, obtain the range of ultraviolet bands, radiation doses, and action times that the current feature vector set can meet the target inactivation rate, and form the optimal parameter range; it is also used to further limit the optimal parameter range based on the optimal parameter range, combined with preset safety thresholds and constraints of ultraviolet disinfection operation; and select an executable parameter combination from the limited optimal parameter range as the ultraviolet disinfection control strategy according to preset ultraviolet disinfection rules; the parameter combination includes ultraviolet band, radiation dose, and action time.

[0115] The preset rules for ultraviolet disinfection include energy consumption priority rules, safety priority rules, efficiency priority rules, and equipment life protection rules.

[0116] The energy consumption priority rule requires that the parameter combination with the lowest total power consumption (such as low power or short time) be selected first within the optimal parameter range;

[0117] The safety priority rule requires that when people are present in the disinfection environment, the ultraviolet band must be 222nm, and parameter combinations that include the 254nm ultraviolet band must be excluded.

[0118] The efficiency priority rule requires that, in the case of an unoccupied disinfection environment and when time is of the essence, the combination with the shortest action time (i.e., the highest dose rate) should be selected first.

[0119] The equipment life protection rules require prioritizing parameter combinations that avoid frequent switching of ultraviolet lamps or long-term full-load disinfection.

[0120] The adaptive adjustment module 10323 is used to obtain the actual disinfection environment parameters and the actual inactivation rate during the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed; if the actual inactivation rate reaches the target inactivation rate, the disinfection is stopped; if the actual disinfection environment parameters change or the actual inactivation rate is lower than the target inactivation rate, the ultraviolet disinfection control strategy is regenerated.

[0121] The power management module 1034 is used to perform low-level drive control and provide stable power to the intelligent control component 103.

[0122] The ultraviolet irradiation component 104 is used to adjust the ultraviolet band, irradiation dose, and action time of the ultraviolet emission source according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters to perform the ultraviolet disinfection task; the ultraviolet irradiation component 104 includes an ultraviolet light source unit 1041, a band selection / switching module 1042, and a dose adjustment module 1043; the ultraviolet irradiation component 104 includes at least two sets of ultraviolet light source units with different peak wavelength emission sources (such as a 222nm excimer lamp group and a 254nm mercury lamp or LED group).

[0123] The ultraviolet light source unit 1041 is used to generate ultraviolet radiation, and includes at least one ultraviolet emission source that emits ultraviolet light with a predetermined peak wavelength when powered on.

[0124] The band selection / switching module 1042 is used to physically switch the ultraviolet band by switching on and off the power supply circuit of a specific wavelength emission source according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters.

[0125] The dose adjustment module 1043 is used to adjust the ultraviolet radiation intensity by adjusting the driving voltage or duty cycle PWM according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters.

[0126] The communication component 105 is used for data interaction between the intelligent ultraviolet disinfection system and external terminals or networks.

[0127] The power supply component 106 is used to provide stable power to the various components of the system.

[0128] To visually represent the adaptive closed-loop adjustment process described above, this invention also provides an adaptive intelligent monitoring visualization interface. Please refer to [link / reference]. Figure 5 This interface, generated by intelligent control components, displays the complete data flow of the system's perception, decision-making, and execution in real time, and specifically includes the following three functional areas:

[0129] (1) Multi-source sensing status area, located on the left side of the interface: The multi-source sensing status area displays the parameters collected by the environmental and object sensing components in real time. The interface can synchronously monitor and display the current environmental status parameters (such as turbidity / PM value, flow rate, temperature, and humidity). When a personnel intrusion signal is detected, a high-brightness warning will immediately pop up in this area to intuitively indicate the current source of the interruption.

[0130] (2) The prediction-actual dual-track monitoring area is located in the middle of the interface: The prediction-actual dual-track monitoring area plots a dynamic fitting curve based on a machine learning model. Among them, the dashed line (Predicted) represents the theoretically predicted inactivation rate calculated by the model based on the current turbidity, flow rate, and temperature and humidity; the solid line (Actual) represents the actual inactivation rate based on sensor feedback. The system calculates the deviation between the two in real time. When environmental interference (such as high humidity or sudden changes in turbidity) causes the actual effect to be lower than the predicted value, the interface will automatically trigger a compensation mechanism and display the compensation dose value (as shown in "+3.0mJ / cm" in the figure). 2 This allows for a transparent display of the algorithm's logic.

[0131] (3) Dual-band engine status area, located on the right side of the interface: The dual-band engine status area reflects the working mode of the actuator in real time through a logical topology diagram. In the "safe coexistence mode", the interface clearly shows that the 222 nm safety light source is in the ACTIVE and adaptive enhancement state, while the 254 nm main light source is in the HARD-CUTOFF and locked state. This visualized state mapping effectively verifies the system's safe switching logic in a human-machine coexistence environment.

[0132] Example 3

[0133] Based on the intelligent ultraviolet disinfection system 10 described in Embodiment 2 above, this embodiment further provides an exemplary application implementation method based on a specific disinfection scenario, which is used to illustrate the operation mode and effect of the intelligent ultraviolet disinfection system of the present invention in the actual water disinfection process.

[0134] The intelligent ultraviolet disinfection system 10 is deployed in a water disinfection scenario to perform ultraviolet disinfection on the water to be treated. The water is the effluent from a conventional drinking water or reclaimed water treatment unit, and its initial microbial load, turbidity, and water quality parameters are within the range of conventional engineering applications.

[0135] After the system is started, the initial working mode and initial start / stop status of the system are set through the function control component 101, and the ultraviolet disinfection control parameters are initialized. The environment and object perception component 102 collects the current disinfection environment parameters of the water body and the operating status data of the disinfection system, including but not limited to water body type, water temperature, light transmittance, flow rate, ultraviolet band, radiation dose, action time, light intensity feedback, energy consumption, etc., and sends the parameters to the intelligent control component 103.

[0136] The data preprocessing module 10321 takes the permutations and combinations of the allowable range of ultraviolet bands (e.g., 222, 254) and the allowable range of radiation doses (e.g., 7.7, 8.0) in the current disinfection environment parameters collected (e.g., the permutations are divided into 4 groups, namely ultraviolet band 222, radiation dose 7.7, ultraviolet band 222, radiation dose 8.0, ultraviolet band 254, radiation dose 7.7, ultraviolet band 254, radiation dose 8.0), and constructs the current parameter combination range of the disinfection target. Each current parameter combination is concatenated with the current disinfection environment parameters to construct the current dataset; the current dataset is preprocessed to obtain the current feature vector set.

[0137] The current feature vector set is input into the machine learning prediction module 10322. The machine learning prediction module predicts the inactivation rate under different combinations of ultraviolet bands and radiation doses based on the trained disinfection effect prediction model. Some prediction results are shown in Table 1 below (1 log=90% inactivation rate, 2 log=99% inactivation rate, 3 log=99.9% inactivation rate).

[0138] Table 1 Examples of predicted disinfection effects under different combinations of ultraviolet bands and dosages

[0139] Ultraviolet band (nm) Irradiation dose (mJ / cm²) Predicted inactivation rate (log) 222 7.7 2.0 222 8.0 3.0 222 9.8 4.0 254 4.9 2.0 254 7.7 3.0 254 10.9 4.0

[0140] The above values ​​are merely exemplary predictions output by the machine learning model under specific conditions, used to illustrate the working process of the system of the present invention, and do not constitute a limitation on the disinfection parameters.

[0141] The control decision module 1033 statistically analyzes the ultraviolet bands, radiation doses, and corresponding action times that meet the preset target inactivation rate, obtaining the range of ultraviolet bands, radiation doses, and action times that the current feature vector set can satisfy the target inactivation rate (e.g., ≥2 log or ≥4 log), forming the optimal parameter range. Based on the optimal parameter range, combined with preset safety thresholds and constraints of ultraviolet disinfection operation, the optimal parameter range is further defined. According to preset ultraviolet disinfection rules, an executable parameter combination is selected from the defined optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and action time.

[0142] For example, when the system sets a target inactivation rate of ≥4 log, the control decision module 1033 comprehensively evaluates different combinations of wavebands and doses, and an example of its strategy generation is shown in Table 2:

[0143] Table 2 Example of UV control strategy generation

[0144] Target inactivation rate Candidate ultraviolet bands (nm) Dosage range (mJ / cm²) Control Strategy Description ≥2 log 222 / 254 5–30 Prioritize low-dose bands ≥4 log 254 10–50 Optimal dose efficiency ≥4 log 280 ≥70 High energy consumption, not the preferred choice

[0145] During or after the disinfection process, the system obtains the actual disinfection environment parameters (such as the actual irradiation status and disinfection effect feedback information) and the actual inactivation rate through the environment and object perception component 102, and returns the information to the data storage and analysis module 1032.

[0146] The adaptive adjustment module 10323 compares and analyzes the actual inactivation rate with the predicted inactivation rate during the execution of the ultraviolet disinfection control strategy, and automatically corrects the control parameters when there is a deviation.

[0147] If the actual inactivation rate reaches the target inactivation rate, disinfection is stopped; if the actual disinfection environmental parameters change, or the actual inactivation rate is lower than the target inactivation rate, the control decision module is triggered to regenerate the ultraviolet disinfection control strategy. The effect comparison before and after adjustment is shown in Table 3.

[0148] Table 3. Examples of Comparison of Disinfection Effects Before and After Adaptive Adjustment

[0149] Control phase Ultraviolet band (nm) Irradiation dose (mJ / cm²) Actual inactivation effect (log) Initial prediction 222 10.0 2.0 After adjustment 254 7.7 3.0

[0150] Through the above adaptive adjustment process, the system can dynamically correct the control strategy in actual operation, so that the disinfection effect gradually approaches or reaches the preset target, thus forming a complete adaptive closed-loop intelligent control process.

[0151] Example 4

[0152] Based on the intelligent ultraviolet disinfection system 10 described in Embodiment 2 above, this embodiment provides a specific implementation method for indoor air and object surface disinfection scenarios, illustrating how the present invention achieves safe and efficient ultraviolet disinfection control under the condition of dynamic human presence.

[0153] The intelligent ultraviolet disinfection system 10 is deployed in indoor environments, such as wards, laboratories, public transportation spaces, or clean rooms, to disinfect indoor air and exposed surfaces.

[0154] During system operation, the environmental and object perception component 102 collects information about the disinfection environment. The collected information includes, but is not limited to: attribute parameters of the disinfection environment (indoor space type and size parameters); status parameters of the disinfection environment (such as temperature, humidity, air flow status, and personnel presence information); and type data of the disinfection environment (air or surface).

[0155] Personnel presence status information can be obtained through infrared human body sensing, visual sensing, or other equivalent methods. The collected data is sent to the information transceiver module 1031 in the intelligent control component 103.

[0156] The information transceiver module 1031 transmits the collected data to the data storage and analysis module 1032. The data preprocessing module 10321 preprocesses the data and constructs feature vectors before inputting it into the machine learning prediction module 10322.

[0157] The machine learning prediction module, based on a trained disinfection effect prediction model, predicts and analyzes the relationship between ultraviolet bands, irradiation doses, and disinfection effects under different personnel conditions. The prediction results are shown in Table 4.

[0158] Table 4 Examples of predicted disinfection effects under different personnel conditions

[0159] Personnel status Ultraviolet band (nm) <![CDATA[Irradiation dose (mJ / cm 2 )]]> Predicted inactivation rate (log) unmanned 254 40 4.2 unmanned 222 40 3.1 Someone 222 30 2.0 Someone 254 30 Not recommended (due to security restrictions)

[0160] The data above indicate that in unmanned environments, the system can preferentially use the 254 nm band to obtain higher unit dose inactivation efficiency; while in manned environments, the system preferentially selects the 222 nm band to achieve continuous disinfection while meeting safety requirements.

[0161] The system statistically analyzes the ultraviolet bands, radiation doses, and corresponding treatment times that meet the preset target inactivation rate, and outputs the range of ultraviolet bands, radiation doses, and treatment times that the current feature vector set can meet the target inactivation rate, forming the optimal parameter range.

[0162] The control decision module 1033 further limits the optimal parameter range based on the optimal parameter range, combined with safety thresholds, regulatory constraints, and disinfection targets; according to the preset rules for ultraviolet disinfection, it selects an executable parameter combination from the limited optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and exposure time.

[0163] The control decision module 1033 sends the control strategy to the ultraviolet irradiation component 104. During the disinfection process, the system continuously acquires environmental change information, ultraviolet operation status information, and personnel status feedback, and returns the information to the data storage and analysis module 1032.

[0164] The adaptive adjustment module 10323 compares and analyzes the actual operating results with the predicted results. When a deviation in the disinfection effect or a change in the environmental state is detected, the control strategy is automatically corrected, such as by adjusting the irradiation dose or reselecting the ultraviolet band, thereby realizing adaptive closed-loop intelligent control under the condition of dynamic presence of personnel.

[0165] For example, if the system adopts a "safety first" interrupt control logic: when no one is in the room, the system performs broad-spectrum disinfection in the 254 nm band and high-power mode; once the environmental sensing component 102 detects a person entering the room, it corrects the ultraviolet disinfection control strategy: turns off the 254 nm ultraviolet light source; turns on the 222 nm ultraviolet light source; reduces the single irradiation dose and extends the action time, thereby achieving seamless switching while strictly protecting the human body from harmful radiation.

Claims

1. A smart ultraviolet disinfection method, characterized in that, include: Collect historical datasets, including historical disinfection data and historical inactivation rates. The historical disinfection data includes ultraviolet band, radiation dose, disinfection environmental parameters, and contact time. Preprocess the historical dataset to obtain a set of historical feature vectors; A disinfection effect prediction model is built based on the CatBoost model. The optimization objective is to minimize the mean squared error between the predicted inactivation rate of the disinfection effect prediction model and the historical inactivation rate in the historical dataset. The model is iteratively trained to obtain a well-trained disinfection effect prediction model. The iterative training includes: dividing the historical feature vector set into a training set and a validation set according to a preset ratio; setting the maximum number of iterations, the learning rate, and the initial parameters of the disinfection effect prediction model; in each iteration, inputting the training set into the disinfection effect prediction model and calculating the predicted inactivation rate; calculating the mean squared error of the disinfection effect prediction model based on the predicted inactivation rate and the historical inactivation rate, using the mean squared error as the loss function; calculating the gradient of the loss function relative to the disinfection effect prediction model, and updating the parameters of the disinfection effect prediction model based on the gradient and the learning rate; after each or several rounds of parameter updates, calculating the prediction performance of the disinfection effect prediction model based on the validation set. If the prediction performance reaches a preset accuracy or the number of iterations reaches the maximum number of iterations, training stops; otherwise, the next round of iteration training continues. The prediction performance index includes R... 2 The parameters of the updated disinfection effect prediction model include: calculating the loss function with respect to the parameters. The derivative of the loss function is used as the gradient, and the formula for the loss function is: ； in, For loss function, For parameters, Let i be the actual inactivation rate of the i-th data set. Let i be the predicted inactivation rate of the i-th data set. The amount of data in the training set; The parameters of the disinfection effect prediction model are updated based on the gradient and learning rate, using the following formula: ； in, These are the parameters of the current disinfection effectiveness prediction model. For the updated model parameters, For learning rate, For the loss function with respect to the parameters The derivative; The current disinfection environment parameters of the disinfection target are collected. Based on the permutation and combination of the allowable range of ultraviolet band and the allowable range of radiation dose in the current disinfection environment parameters, the current parameter combination range of the disinfection target is formed. Each current parameter combination is concatenated with the current disinfection environment parameters to construct the current dataset of the disinfection target. The current feature vector set is obtained by preprocessing the current dataset. Input the current feature vector set into the trained disinfection effect prediction model, predict the inactivation rate and action time based on each current feature vector, statistically analyze the ultraviolet band, radiation dose and corresponding action time that meet the preset target inactivation rate, and output the range of ultraviolet band, radiation dose and action time that the current feature vector set can meet the target inactivation rate, forming the optimal parameter range. Based on the optimal parameter range, and combined with the preset safety threshold and the constraints of ultraviolet disinfection operation, the optimal parameter range is further defined; according to the preset ultraviolet disinfection rules, an executable parameter combination is selected from the defined optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and action time. During the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed, the actual disinfection environment parameters and the actual inactivation rate are obtained; if the actual inactivation rate reaches the target inactivation rate, the disinfection is stopped; if the actual disinfection environment parameters change or the actual inactivation rate is lower than the target inactivation rate, the ultraviolet disinfection control strategy is regenerated.

2. The intelligent ultraviolet disinfection method of claim 1, wherein, The historical dataset includes at least one of the following datasets: (1) Historical disinfection data and historical inactivation rates recorded in ultraviolet disinfection experiments conducted under controlled experimental conditions; (2) Historical disinfection data and historical inactivation rate of ultraviolet disinfection experiments recorded in publicly available literature or industry databases.

3. The intelligent ultraviolet disinfection method of claim 2, wherein, The disinfection environment parameters include disinfection environment type, disinfection environment attribute parameters, and disinfection environment status parameters; the disinfection environment attribute parameters include the allowable range of ultraviolet band, the allowable range of radiation dose, the size parameters, and bacterial concentration; the disinfection environment status parameters include whether there are people in the disinfection environment, the temperature, humidity, and pH.

4. The intelligent ultraviolet disinfection method of claim 3, wherein, The preprocessing includes at least one of data cleaning, outlier removal, normalization, and feature encoding.

5. The intelligent ultraviolet disinfection method according to claim 4, characterized in that, The preset rules for ultraviolet disinfection include energy consumption priority rules, safety priority rules, efficiency priority rules, and equipment lifespan protection rules. The energy consumption priority rule requires selecting the parameter combination with the lowest total power consumption within the optimal parameter range. The safety priority rule requires that when people are present in the disinfection environment, the ultraviolet wavelength should be forced to 222nm, and parameter combinations containing the 254nm ultraviolet wavelength should be excluded. The efficiency priority rule requires that when no one is present in the disinfection environment and time is limited, the combination with the shortest action time should be selected. The equipment lifespan protection rule requires that parameter combinations that can avoid frequent switching of ultraviolet lamps or long-term full-load disinfection should be selected.

6. A system for performing the intelligent ultraviolet disinfection method according to any one of claims 1-5, characterized in that, include: Environment and object perception components and intelligent control components; The environment and object perception component is used to collect historical datasets, including historical disinfection data and historical inactivation rates. The historical disinfection data includes ultraviolet bands, radiation doses, disinfection environmental parameters, and actual action time. It is also used to collect the current disinfection environmental parameters of the disinfection target. It is also used to obtain actual disinfection environment parameters and actual inactivation rate during the execution of ultraviolet disinfection control strategies or after the disinfection task is completed; The intelligent control component includes a data storage and analysis module and a control decision module; the data storage and analysis module includes a data preprocessing module, a machine learning prediction module, and an adaptive adjustment module. The data preprocessing module is used to preprocess the historical dataset to obtain a historical feature vector set. It is also used to form the current parameter combination range of the disinfection target according to the permutation and combination of the allowable range of ultraviolet band and the allowable range of radiation dose in the current disinfection environment parameters. Each current parameter combination is concatenated with the current disinfection environment parameters to construct the current dataset. The current dataset is then preprocessed to obtain the current feature vector set. The machine learning prediction module is used to build a disinfection effect prediction model. On the historical feature vector set, the optimization objective is to minimize the mean square error between the predicted inactivation rate of the disinfection effect prediction model and the historical inactivation rate in the historical dataset. The disinfection effect prediction model is iteratively trained to obtain a well-trained disinfection effect prediction model. It is also used to predict the inactivation rate and action time for each current feature vector in the current feature vector set based on the well-trained disinfection effect prediction model. The control decision module is used to statistically analyze the ultraviolet bands, radiation doses, and corresponding exposure times that meet the preset target inactivation rate, obtain the range of ultraviolet bands, radiation doses, and exposure times that the current feature vector set can satisfy the target inactivation rate, and form the optimal parameter range; it is also used to further limit the optimal parameter range based on the optimal parameter range, combined with preset safety thresholds and constraints of ultraviolet disinfection operation; according to preset ultraviolet disinfection rules, it selects an executable parameter combination from the limited optimal parameter range as the ultraviolet disinfection control strategy; the parameter combination includes ultraviolet band, radiation dose, and exposure time. The adaptive adjustment module is used to obtain the actual disinfection environment parameters and the actual inactivation rate during the execution of the ultraviolet disinfection control strategy or after the disinfection task is completed; if the actual inactivation rate reaches the target inactivation rate, the disinfection is stopped; if the actual disinfection environment parameters change or the actual inactivation rate is lower than the target inactivation rate, the ultraviolet disinfection control strategy is regenerated.

7. The intelligent ultraviolet disinfection system of claim 6, wherein, It also includes functional control components, ultraviolet irradiation components, communication components, and power supply components; The functional control component is used to set the working mode, start / stop status and ultraviolet disinfection control parameters of the ultraviolet disinfection system; the ultraviolet disinfection control parameters include ultraviolet band, irradiation dose and action time. The ultraviolet irradiation component is used to adjust the ultraviolet band, irradiation dose and action time of the ultraviolet emission source according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters, so as to perform the ultraviolet disinfection task. The communication component is used for data interaction between the ultraviolet disinfection system and external terminals or networks. The power supply component is used to provide stable power to all components of the system.

8. The intelligent ultraviolet disinfection system according to claim 7, characterized in that, The intelligent control component also includes an information transceiver module and a power management module; the information transceiver module is used for data input and output; the power management module is used to provide stable power to the intelligent control component. The ultraviolet irradiation assembly includes an ultraviolet light source unit, a band selection / switching module, and a dose adjustment module. The ultraviolet light source unit generates ultraviolet radiation and includes at least one ultraviolet emission source that emits ultraviolet radiation at a predetermined peak wavelength when powered on. The band selection / switching module physically switches the ultraviolet band by switching the power supply circuit of a specific wavelength emission source on and off, according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters. The dose adjustment module adjusts the ultraviolet radiation intensity by adjusting the driving voltage or duty cycle, according to the ultraviolet disinfection control strategy or the set ultraviolet disinfection control parameters. The ultraviolet irradiation assembly includes at least two sets of ultraviolet light source units with different peak wavelength emission sources.

Images