A whole-process supervision method for medical waste based on a smart medical waste monitoring platform

Abstract

The present application provides a method for the whole-process supervision of medical waste based on an intelligent medical waste monitoring platform, including: matching the collected odor data with the pre-established medical waste odor component data to determine the type of medical waste to which the odor source belongs and determining whether the odor concentration exceeds a preset threshold; if the odor concentration exceeds the standard, obtaining microbial information based on the correlation between the odor component information and microorganisms, and combining with the temperature and humidity control parameters of the air conditioner to evaluate the risk of microbial growth; simulating the air flow distribution of the air conditioning system, combining with the layout of the air outlet positions of the air conditioner to predict the diffusion path and influence range of the odor in different areas of the hospital, and generating an odor diffusion heat map; according to the optimized disinfection plan, controlling the air conditioning system to enter the disinfection mode, adjusting the fresh air introduction ratio and the filtration efficiency level, and spraying disinfectant in key areas, and monitoring the disinfection effect in real time.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a full-process medical waste supervision method based on a smart medical waste monitoring platform. Background Art

[0002] During the intra-hospital transportation of medical waste, there is a complex relationship between the generation of odor and the spread of microorganisms. The odor components released by different types of medical waste on the transportation route are different and fluctuate with changes in temperature and humidity. These odors not only affect the environment, but may also indicate potential biological hazards. However, the existing regulatory system makes it difficult to track the source of odor in real time and predict its diffusion trend. At the same time, although the hospital air-conditioning system can be used to control odor, its airflow distribution often does not match the waste transportation route, making it difficult to accurately control the odor diffusion range. In addition, the preliminary disinfection treatment for different microorganisms needs to ensure the effect while avoiding chemical reactions with odor components to produce secondary pollution. However, there is currently a lack of intelligent regulatory systems that can comprehensively consider odor diffusion, microbial activity, air conditioning airflow, and disinfection reactions, making it difficult to achieve accurate control and risk warning of the entire process of medical waste transportation in a complex and changeable intra-hospital environment. Summary of the invention

[0003] The present invention provides a method for monitoring the entire process of medical waste based on a smart medical waste monitoring platform, which mainly includes:

[0004] According to the classification of medical waste, a distributed sensor network is used to sample the odor of waste transfer routes in different areas to obtain odor data, including odor concentration data and odor component information;

[0005] Match the collected odor data with the pre-established medical waste odor component data to determine the type of medical waste to which the odor source belongs and whether the odor concentration exceeds a preset threshold;

[0006] If the odor concentration exceeds the standard, the microbial information is obtained based on the correlation between the odor component information and the microorganisms, and the risk of microbial growth is evaluated in combination with the temperature and humidity control parameters of the air conditioner;

[0007] Simulate the air flow distribution of the air conditioning system, combine the location layout of the air conditioning outlets, predict the diffusion path and impact range of odors in different areas of the hospital, and generate an odor diffusion heat map;

[0008] Based on the odor diffusion heat map, combined with the frequency of air conditioning filter replacement and pipeline cleaning cycle factors, determine the areas that need to be disinfected. At the same time, based on the risk of microbial growth, select a disinfection plan to deal with such odors and microorganisms;

[0009] According to the selected disinfection plan, extract the chemical components of the disinfectant used in the disinfection plan. Through chemical reaction simulation, evaluate the potential reaction of the chemical components with the known odor components in the air. According to the evaluation results, if toxic and harmful substances will be produced, optimize the disinfection plan to avoid secondary pollution;

[0010] According to the optimized disinfection plan, control the air conditioning system to enter the disinfection mode, adjust the fresh air introduction ratio and the filtration efficiency level, and spray the disinfectant in key areas, and monitor the disinfection effect in real time.

[0011] The technical solution provided by the embodiment of the present invention may include the following beneficial effects:

[0012] The present invention discloses a whole-process supervision method for medical waste based on an intelligent medical waste monitoring platform. Collect the odor data of the medical waste transfer path through a distributed sensor network, and match it with the pre-built database to judge the odor source and concentration. When the odor concentration exceeds the standard, query the microorganism database based on the odor component information, and evaluate the microorganism breeding risk in combination with the air conditioning system parameters. At the same time, simulate the odor diffusion path, generate a heat map to determine the key disinfection area. Select a suitable plan from the disinfectant database, evaluate the risk of generating potential harmful substances through chemical reaction simulation and optimize and adjust to avoid secondary pollution. Finally, control the air conditioning system to enter the disinfection mode and monitor the effect in real time. The present invention realizes the intelligent monitoring, risk assessment and precise disinfection of medical waste odors, effectively reduces the risk of nosocomial infection, and improves the safety of the medical environment. Description of the Drawings

[0013] Figure 1 It is a flowchart of a whole-process supervision method for medical waste based on an intelligent medical waste monitoring platform of the present invention.

[0014] Figure 2 It is a schematic diagram of a whole-process supervision method for medical waste based on an intelligent medical waste monitoring platform of the present invention.

[0015] Figure 3 It is another schematic diagram of a whole-process supervision method for medical waste based on an intelligent medical waste monitoring platform of the present invention. Detailed Embodiments

[0016] To further understand the content of the present invention, the present invention will be described in detail in combination with the drawings and embodiments. The following will further elaborate on the present application with reference to the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention and are not intended to limit the invention. In addition, it should be noted that for the sake of description, only the parts related to the invention are shown in the drawings.

[0017] Such as Figures 1-3, a method for the whole-process supervision of medical waste based on an intelligent medical waste monitoring platform in this embodiment may specifically include:

[0018] Step S101, according to the classification of medical waste, use a distributed sensor network to sample the odor of the waste transfer paths in different regions to obtain odor data, including odor concentration data and odor component information.

[0019] Obtain the sampling area information divided according to the pre-established medical waste classification standard, set multiple odor sampling points in the sampling area to form a distributed sensor network covering the waste transfer path. Use a gas sensor array to detect the odor sampling points in real time to obtain the odor concentration and component data of each sampling point; fuse different types of sensor data according to the weighted average algorithm to determine the comprehensive odor data. Perform preprocessing on the comprehensive odor data, eliminate environmental noise interference through a filtering algorithm; reduce the dimension of the multi-dimensional odor feature data, and extract an odor fingerprint feature vector containing key parameters such as the concentration ratio of gas components, odor intensity, and duration.

[0020] Exemplarily, the sampling area is divided according to the pre-established medical waste classification standard. Multiple odor sampling points are set in each waste category area to form a distributed sensor network covering the waste transfer path. A high-sensitivity gas sensor array is used to detect the odor concentration and composition at each sampling point in real time. The sensor array includes metal oxide, electrochemical, and photoionization sensors. The weighted average algorithm is used to fuse the data of different types of sensors to obtain comprehensive odor data. The odor data collected at each sampling point is preprocessed. The Kalman filtering algorithm is used to eliminate environmental noise interference, and the singular value decomposition (SVD) algorithm is used to reduce the dimension of the multi-dimensional odor feature data to extract the odor fingerprint feature vector. The odor fingerprint feature vector includes key parameters such as the concentration ratio of gas components, odor intensity, and duration. An odor database is established, and the odor data is associated with the waste transfer path according to the geographical location information of the sampling points to form an odor distribution map. A convolutional neural network with the ResNet-50 architecture is used to perform pattern recognition and classification on the preprocessed odor data. The transfer learning method is adopted, using the pre-trained model as a feature extractor and then fine-tuning on the medical waste odor dataset. The recognition results are matched with the preset medical waste odor feature library to judge the waste category and concentration level at each sampling point. A waste distribution heat map is drawn to mark the main aggregation areas and transfer channels of different categories of waste. Based on the odor distribution heat map, the rationality of the current transfer path is evaluated, and the odor intensity and waste category distribution of each section are calculated. A medical waste transfer path optimization model is constructed, considering factors such as the hazard level of various wastes and the requirements for treatment timeliness, and combining information such as waste generation volume and the location of treatment facilities. The Max-Min ant colony algorithm is used to dynamically optimize and adjust the existing transfer path, where the odor intensity on the path is used as heuristic information and the hazard degree of waste categories is used as a constraint condition for path selection. Through multiple iterative calculations, an optimized classified transfer path plan is generated to achieve efficient classified transfer of medical waste in different regions. In the division of the medical waste sampling area, the area is divided into five categories: infectious, injurious, pathological, chemical, and pharmaceutical according to the pre-established classification standard. 20 odor sampling points are set in each category area to form a distributed sensor network with 100 nodes. Each sampling point is equipped with three sensors. The metal oxide sensor detects volatile organic compounds, the electrochemical sensor detects ammonia and hydrogen sulfide, and the photoionization sensor detects trace organic compounds. The weighted average algorithm is used to fuse the data, and the weight ratio is 4:3:3. In the Kalman filtering algorithm, the measurement noise covariance R is set to 0.1, and the process noise covariance Q is set to 0.01 to balance sensitivity and stability. The SVD algorithm reduces the dimension of the original 50-dimensional odor data to a 10-dimensional feature vector, retaining 95% of the information volume. The odor fingerprint feature vector includes the relative concentration ratios of 10 key gas components such as formaldehyde, ethanol, and acetone. The odor intensity is represented by a quantization value from 0 to 100, and the duration is recorded in seconds.The odor distribution map uses shades of color to represent concentration, and different patterns to represent waste categories. The ResNet-50 network uses pre-trained weights, and the last three layers are replaced with fully connected layers adapted to medical waste classification, with the number of layers being 1024-512-5. Transfer learning is performed by fine-tuning with 5000 pre-annotated medical waste odor maps, with a learning rate set to 0.001, a batch size of 64, and training for 50 epochs. The waste distribution heat map uses a three-color gradient of red, yellow, and green to represent the level of concentration, and different-shaped icons to mark five categories of waste. When evaluating the rationality of the transfer path, the average odor intensity and the distribution ratio of waste categories on each path are calculated. In the improved Max-Min ant colony algorithm, the reciprocal of the odor intensity on the path is used as the heuristic information, and the risk level of waste categories is used as the constraint condition for path selection. The pheromone evaporation coefficient is set to 0.5, the local update coefficient is 0.1, and the global update coefficient is 0.2. Through 100 iterations of calculation, an optimized classification and transfer path plan is generated, realizing the efficient classification and transfer of medical waste in different regions. The average path length is shortened by 15% compared to before optimization, and the exposure time of hazardous waste is reduced by 20%.

[0021] Step S102: Match the collected odor data with the pre-established medical waste odor component data, determine the type of medical waste to which the odor source belongs, and determine whether the odor concentration exceeds a preset threshold.

[0022] Obtain odor data collected by a multi-channel gas sensor array, where the odor data includes the outputs of metal oxide, electrochemical, and photoionization sensors; dynamically adjust the weights according to the accuracy of each sensor in historical data, and use the weighted average method to fuse the odor data; perform standardization processing on the fused data to obtain an odor feature vector in a unified format. Extract feature templates from the pre-established medical waste odor component database, where the feature templates contain typical odor components and concentration ranges of different types of medical waste; perform dimensionality reduction processing on the odor feature vector, and select the first N principal components with a cumulative contribution rate reaching a preset value as the key odor component information. Use a support vector machine classifier to classify the odor feature vector, where the support vector machine classifier uses a radial basis function kernel; optimize the kernel parameters and penalty factors of the radial basis function kernel through cross-validation; if the output result of the support vector machine classifier matches the preset category, then determine the type of medical waste to which the odor source belongs. Establish a preset threshold database, where the preset threshold database contains multiple levels of odor concentration thresholds; use the cubic spline interpolation algorithm to convert the original sensor signal into a standardized odor concentration value; determine whether the standardized odor concentration value exceeds the corresponding threshold in the preset threshold database. If the standardized odor concentration value exceeds the preset threshold, trigger an alarm at the corresponding level according to the degree of over-standard.

[0023] Exemplarily, a multi-channel gas sensor array is used to collect odor data, including metal oxide, electrochemical, and photoionization sensors. Data fusion is performed by the weighted average method, and the weights are dynamically adjusted according to the accuracy of each sensor in historical data. The fused data is standardized, and moving average filtering is used to eliminate environmental temperature and humidity interference, obtaining odor feature vectors in a unified format. The processed odor feature vectors are passed to the subsequent steps. Feature templates are extracted from a pre-established medical waste odor component database, which contains typical odor components and concentration ranges of different types of medical waste. Singular value decomposition (SVD) is used to reduce the dimensionality of the odor feature vectors, and the first N principal components with a cumulative contribution rate reaching 95% are selected as the key odor component information. New odor samples are collected regularly to update the feature templates and concentration ranges in the database. A support vector machine (SVM) classifier is used to classify the odor feature vectors to determine the type of medical waste from which the odor originates. The SVM classifier uses a radial basis function (RBF) kernel, and the kernel parameters and penalty factors are optimized through cross-validation. In the training phase, the grid search method is used to determine the optimal parameter combination to improve the classification accuracy. According to historical data statistics, a preset threshold database is established and updated regularly. For each type of medical waste, multi-level odor concentration thresholds are set, including warning values and danger values. The cubic spline interpolation algorithm is used to convert the original sensor signals into standardized odor concentration values, and the interpolation points are selected at the key inflection points of the sensor response curves. By comparing the calculated odor concentration with the preset thresholds, it is determined whether the odor concentration exceeds the standard, and different levels of alarms are triggered according to the degree of exceeding the standard. In the multi-channel gas sensor array, metal oxide sensors detect volatile organic compounds, electrochemical sensors detect ammonia and hydrogen sulfide, and photoionization sensors detect trace organic compounds. In the weighted average method, the initial weights are set to 0.4, 0.3, and 0.3, and the weights are automatically adjusted every 24 hours according to the accuracy of each sensor. The standardization process uses the z-score method to convert the data into a distribution with a mean of 0 and a standard deviation of 1. The moving average filtering window size is set to 60 seconds to eliminate the influence of environmental temperature and humidity fluctuations in real time. The medical waste odor component database contains the typical odor characteristics of five major categories of waste: infectious, injurious, pathological, chemical, and pharmaceutical. During the SVD dimensionality reduction process, the first 5 principal components are selected, and the cumulative contribution rate reaches 96.8%. 100 new odor samples are automatically collected every week to update the feature templates and concentration ranges in the database. The SVM classifier uses an RBF kernel, and the parameters are optimized through 5-fold cross-validation. The range of the kernel parameter γ is [0.001, 0.1], and the range of the penalty factor C is [1, 100], and the grid search step size is 0.1. The final optimal parameter combination is determined as γ = 0.01 and C = 10, and the classification accuracy reaches 94.5%. In the preset threshold database, 3 levels of thresholds are set for each waste type as warning values, danger values, and emergency values, corresponding to the 70%, 85%, and 95% quantiles of the odor concentration, respectively.The cubic spline interpolation algorithm selects the 0%, 25%, 50%, 75%, and 100% response points of the sensor response curve as key inflection points to construct a smooth concentration conversion curve. If the calculated odor concentration exceeds the preset threshold, an alarm of the corresponding level is triggered, and a real-time warning message is pushed to the waste management personnel.

[0024] Step S103, if the odor concentration exceeds the standard, obtain microbial information based on the correlation between the odor component information and the microorganisms, and evaluate the risk of microbial growth in combination with the temperature and humidity control parameters of the air conditioner.

[0025] Obtain the odor concentration detection result, and judge whether it exceeds the preset threshold according to the detection result. If the detection result exceeds the preset threshold, trigger the subsequent evaluation process; use the Apriori algorithm to analyze the correlation between the odor component information and the microorganisms, the support threshold of the Apriori algorithm is a preset value, and the confidence threshold of the Apriori algorithm is a preset value; screen out the microbial species highly related to the odor components according to the correlation analysis result to obtain the growth conditions and reproduction speed information of the microbial species; establish a microbial growth prediction model, and the prediction model calculates the potential growth rate of the microorganisms according to the growth conditions, the reproduction speed information, the temperature parameters and the humidity parameters; judge whether the potential growth rate exceeds the preset risk level threshold. If the potential growth rate exceeds the preset risk level threshold, determine the microbial growth risk level; the risk level includes low risk, medium risk and high risk.

[0026] Exemplarily, based on the odor concentration detection result, it is judged whether it exceeds a preset threshold, and the preset threshold is set based on the 95th percentile of historical data. If it exceeds the threshold, the subsequent evaluation process is triggered, and the odor component information detected currently is extracted from the odor component database, including the concentration and proportion of various chemical substances. The odor component database adopts a key-value pair structure, where the key is the chemical substance name and the value is the concentration range, which is automatically updated once a week. The Apriori algorithm is used to analyze the correlation between the odor component information and microorganisms, with the support threshold set to 0.1 and the confidence threshold set to 0.7. The types of microorganisms highly correlated with the current odor components are screened out from the microorganism database, and the key information such as the growth conditions and reproduction rate of these microorganisms is obtained. The screened microorganism information is transmitted to the next prediction model in JSON format. The real-time temperature and humidity data of the air conditioning system are collected, and combined with the growth conditions of microorganisms, a microorganism growth prediction model based on the Gompertz equation is established. This model considers multiple influencing factors such as temperature, humidity, and odor component concentration, and calculates the potential growth rate of microorganisms in the current environment. According to the prediction result, the temperature and humidity parameters of the air conditioning system are automatically adjusted to make the environmental conditions unfavorable for microorganism growth. According to the output result of the microorganism growth prediction model, combined with the preset risk level standard, the risk of microorganism breeding is evaluated. The risk level is divided into three levels: low, medium, and high. Low risk corresponds to a predicted growth rate of less than 0.1 / hour, medium risk corresponds to 0.1 - 0.5 / hour, and high risk corresponds to greater than 0.5 / hour. A risk assessment report is generated, including the risk level, the main risk microorganism types, and the recommended prevention and control measures. At the same time, a microorganism growth trend chart and a risk heat map are generated to visually display the risk distribution. In the medical waste treatment area, the odor concentration detection system monitors the odor concentration in the environment in real time, and samples every 5 minutes using an electrochemical sensor. Based on the historical data of the past 30 days, the 95th percentile is calculated as the preset threshold, which is currently set to 50 ppm. The detected odor concentration reaches 52 ppm, exceeding the preset threshold, and the evaluation process is triggered. The odor component information detected currently is extracted from the odor component database, including 0.8 ppm of formaldehyde, 15 ppm of ethanol, 5 ppm of acetone, etc. The Apriori algorithm analyzes the correlation between the odor components and microorganisms, with the support threshold of 0.1 and the confidence threshold of 0.7, and identifies three microorganisms highly correlated with the current odor as Escherichia coli, Staphylococcus aureus, and Pseudomonas aeruginosa. The growth conditions of these microorganisms are obtained, including the optimal growth temperature of 25 - 37 °C and the relative humidity of 60 - 80%. The real-time temperature and humidity data of the air conditioning system show that the current environmental temperature is 28 °C and the relative humidity is 75%. The growth prediction model based on the Gompertz equation calculates that the current potential growth rate of Escherichia coli is 0.3 / hour. The air conditioning temperature is adjusted to 22 °C and the relative humidity is reduced to 55% to inhibit the growth of microorganisms.The risk assessment result shows a medium risk level, with a growth rate of 0.3 per hour, which is between 0.1 - 0.5 per hour. The generated risk heat map shows that the risk is highest at the center of the treatment area and appears red. The microbial growth trend chart predicts that if no measures are taken, the concentration of Escherichia coli will increase by 7 times after 24 hours. A risk alert is sent to the mobile terminal of the management personnel, suggesting strengthening the disinfection and ventilation of the area.

[0027] According to the odor component information, query in the microbial database to determine the types of microorganisms associated with the odor components. Extract the growth suitable condition information for each identified microorganism, and obtain the current temperature and humidity control parameters of the air conditioner in real time, including temperature setting, humidity control level, and fresh air introduction ratio. Calculate the comprehensive breeding risk index for each microorganism based on the growth suitable condition information of the microorganism and the current temperature and humidity control parameters of the air conditioner.

[0028] Obtain the odor component information and calculate the cosine similarity, which is used to identify the associated microorganisms with a matching degree exceeding the preset threshold; obtain the unique identifier of the associated microorganisms and use the unique identifier to query the growth suitable condition information of the microorganisms; the growth suitable condition information includes temperature range, relative humidity range, pH value range, and oxygen demand; obtain the temperature and humidity control parameters, which include temperature setting value, humidity control level, and fresh air introduction ratio; standardize the temperature and humidity control parameters into standardized values; use a multi-factor weighted scoring model to compare the growth suitable condition information with the standardized values; preset weight values are assigned to the temperature factor, humidity factor, pH value factor, and oxygen demand factor in the multi-factor weighted scoring model; judge the harm degree of the associated microorganisms; if the harm degree of the associated microorganisms exceeds the preset threshold, calculate the comprehensive breeding risk index for the associated microorganisms; the comprehensive breeding risk index is determined by the product of the output result of the multi-factor weighted scoring model and the harm degree.

[0029] Exemplarily, according to the odor component information, a cosine similarity calculation method based on TF-IDF is used to conduct a comprehensive search in the microorganism database. Microorganisms with a matching degree exceeding 80% are identified as associated microorganisms, generating a list of relevant microorganisms, including the microorganism names and their corresponding matching degrees. The matching degree threshold is determined through historical data analysis and adjusted quarterly. The generated list of relevant microorganisms is passed to the next step in JSON format. For each identified microorganism, a B+ tree index structure is used to quickly extract its detailed growth suitability condition information from the microorganism database, including the optimal growth temperature range, optimal relative humidity range, optimal pH value range, and oxygen demand. In the B+ tree index structure, the microorganism name is used as the key, and the growth condition information is used as the value to achieve efficient query and data extraction. The temperature and humidity control parameters of the current air conditioning system are obtained in real time through the air conditioning control system interface, including the temperature setting value, humidity control level, and fresh air intake ratio, and these parameter values are normalized to the normalized values between 0 and 1. The impact of the fresh air intake ratio on microorganism growth is quantified by introducing the air quality index to calculate the inhibitory effect of fresh air on microorganism reproduction. Using a multi-factor weighted scoring model, the microorganism growth suitability condition information is compared with the temperature and humidity control parameters of the current air conditioning system to calculate the suitability score of each microorganism in the current environment. In the model, the weight of the temperature factor is 0.4, the weight of the humidity factor is 0.3, the weight of the pH value factor is 0.2, and the weight of the oxygen demand factor is 0.1. Combining with the microorganism hazard degree weight, the comprehensive breeding risk index of each microorganism is finally obtained. The microorganism hazard degree weight is comprehensively evaluated according to its pathogenicity, transmission ability, and drug resistance, and the analytic hierarchy process is used to determine the weight value. In the medical waste treatment area, the odor sensor detects that the formaldehyde concentration is 0.5 ppm, the ammonia concentration is 2 ppm, and the hydrogen sulfide concentration is 0.1 ppm. A search is conducted in the microorganism database using the cosine similarity calculation method based on TF-IDF, with the odor components as the query terms and the microorganism characteristics as the documents. The calculation results show that the matching degree of Escherichia coli is 85%, that of Staphylococcus aureus is 82%, and that of Pseudomonas aeruginosa is 78%. Since the set matching threshold is 80%, the system generates a list of relevant microorganisms including Escherichia coli and Staphylococcus aureus. The B+ tree index structure is used to quickly extract the growth conditions of these two microorganisms: the optimal growth temperature of Escherichia coli is 35 - 37 °C, the relative humidity is 60 - 70%, and the pH value is 6.0 - 7.0; the optimal growth temperature of Staphylococcus aureus is 30 - 37 °C, the relative humidity is 20 - 80%, and the pH value is 7.0 - 7.5. The air conditioning control system shows that the current temperature is 25 °C, the relative humidity is 55%, and the fresh air intake ratio is 30%. These parameters are normalized to obtain a temperature of 0.71, a humidity of 0.69, and a fresh air ratio of 0.3. In the multi-factor weighted scoring model, the weight of the temperature factor is 0.4, the weight of the humidity is 0.3, the weight of the pH value is 0.2, and the oxygen demand is 0.1.The calculated environmental suitability scores for Escherichia coli and Staphylococcus aureus are 0.68 and 0.72 respectively. Considering the weight of the microbial hazard level, with 0.6 for Escherichia coli and 0.8 for Staphylococcus aureus, the final comprehensive breeding risk indices are 0.41 and 0.58 respectively. A risk report can be generated, suggesting to lower the temperature and humidity and increase the proportion of fresh air introduction to reduce the breeding risk.

[0030] Step S104: Simulate the air flow distribution of the air conditioning system, combine with the layout of the air outlet positions of the air conditioner, predict the diffusion path and influence range of odors in different areas of the hospital, and generate an odor diffusion heat map.

[0031] Obtain the hospital building information model, construct a three-dimensional digital model according to the building information model; divide the three-dimensional digital model into grid cells, record the spatial coordinates, temperature, humidity and air flow information of the grid cells; perform computational fluid dynamics simulation on the grid cells to obtain air flow field distribution data; store the air flow field distribution data in the form of a three-dimensional matrix; receive the odor source position and initial concentration information; according to the odor source position, initial concentration information and air flow field distribution data, use the Euler-Lagrange hybrid model to calculate the diffusion process of odor molecules; obtain the time-series concentration distribution data; analyze the relationship between the air flow field distribution data and the time-series concentration distribution data; optimize the air outlet position and angle according to the relationship; generate a continuous concentration distribution field based on the optimized air outlet layout; map the continuous concentration distribution field to the color space to generate an odor diffusion heat map; superimpose the odor diffusion heat map on the hospital floor plan to display the odor diffusion path and influence range.

[0032] Exemplarily, according to the hospital Building Information Model (BIM), the hospital floor plan and the air conditioning system design drawing are integrated to construct an accurate three-dimensional digital model. The space is divided into grid cells, and the size of each cell is set to 0.1 m × 0.1 m × 0.1 m to improve the simulation accuracy. The spatial coordinates, temperature, humidity, air flow velocity, and direction of each grid point are recorded. The Navier-Stokes equations are used for computational fluid dynamics simulation to calculate the air flow velocity vector and pressure value of each grid point. Considering the temperature stratification effect, the space is vertically divided into multiple temperature layers, and different initial temperatures are set for each layer. The humidity distribution is represented by a relative humidity field function and is calculated in coupling with the temperature field. The air flow field distribution data of the entire space are obtained and stored in the form of a three-dimensional matrix as the input for odor diffusion simulation. Using the real-time data collected by odor sensors, the odor source location and initial concentration are input into the air flow field model. An Euler-Lagrange hybrid model is adopted, combined with the air flow field data, to simulate the movement and diffusion process of odor molecules under the action of air flow. The Euler method calculates the continuous concentration field, and the Lagrange method tracks the trajectories of individual odor particles. The odor concentration values of each grid cell at different time points are calculated to generate time-series concentration distribution data. According to the layout of the air conditioning outlet positions, the relationship between the air flow distribution and the odor diffusion path is analyzed. The genetic algorithm is used to optimize the outlet positions and angles to minimize the odor accumulation in key areas. Based on the optimized layout, the Kriging interpolation method is used to generate a continuous concentration distribution field. The thermal mapping visualization technique is adopted to map different concentration values to the color space, generate an odor diffusion thermal map, and overlay it on the hospital floor plan to visually display the odor diffusion path and the influence range. In the operating area of a three-story hospital, the building information model integrates the 1:100 scale floor plan and the air conditioning system design drawing to construct an accurate three-dimensional digital model of 20 m × 30 m × 9 m. The space is divided into 200×300×90 grid cells, each of which is 0.1 m × 0.1 m × 0.1 m. In the Navier-Stokes equation simulation, the boundary conditions are set as no-slip on the wall, an inlet velocity of 2 m / s, and an outlet pressure of standard atmospheric pressure. The temperature field is initially set to three layers, 22 °C from the ground to 1.5 m, 24 °C from 1.5 m to 4 m, and 26 °C above 4 m. The initial value of the relative humidity field is set to 50%, and it is calculated in coupling with the temperature field. The odor source is set at the center of the operating room, and the initial concentration is 100 ppm. In the Euler-Lagrange hybrid model, the Euler grid is consistent with the spatial grid, and the number of Lagrange particles is set to 10,000. The simulation time step is 0.1 s, and the total duration is 300 s. When the genetic algorithm optimizes the outlet positions, the population size is set to 100, the number of iterations is 50, the crossover probability is 0.8, and the mutation probability is 0.1. The objective function is to minimize the average odor concentration in the key area, within 3 m around the operating table. The Kriging interpolation uses an exponential variogram, and the search radius is 2 m.In the heat map visualization, concentration values from 0 to 10 ppm are mapped to blue, 10 to 50 ppm to green, 50 to 100 ppm to yellow, and above 100 ppm to red. The finally generated heat map shows that under the optimized air outlet layout, the odor mainly diffuses along the ceiling towards the exhaust vent, and the concentration within the surgical area remains below 20 ppm, meeting the requirements of the medical environment.

[0033] In step S105, based on the odor diffusion heat map, considering factors such as the replacement frequency of the air conditioner filter and the pipeline cleaning cycle, determine the areas that need to be disinfected with emphasis. At the same time, select a disinfection plan for dealing with such odors and microorganisms in combination with the risk of microbial growth.

[0034] Obtain the data of the odor diffusion heat map, set the odor concentration threshold according to the historical data quantile, and divide the odor pollution area; through rasterization processing, correspond the heat map data with the actual spatial position to obtain a list of key areas of concern. Based on the filter replacement frequency and pipeline cleaning cycle data, establish a filter efficiency decay model and a pipeline pollution accumulation model; the pipeline pollution accumulation model uses a logarithmic growth function as P(t) = a ln(t + 1), where P(t) is the pollution degree at time t and a is the pollution rate coefficient. Calculate the remaining efficiency of the current filter and the pipeline pollution degree, and associate the results with the list of key areas of concern. Use a multi-factor weighted scoring method to score the odor concentration, filter efficiency, pipeline pollution degree, and microbial growth risk; if the scoring result is within the pre-set percentage, it is determined as the key disinfection area. Select a suitable disinfection plan according to the odor type and microbial species in the key disinfection area; construct a decision tree, the root node of the decision tree is the odor type, the second-level node is the main microbial species, and the third-level node is the environmental sensitivity. Select the optimal disinfection plan through the decision tree, and generate a disinfection guidance plan including the name of the disinfectant, the use concentration, and the application method.

[0035] Exemplarily, according to the odor diffusion heat map, the low, medium, and high odor concentration thresholds are set using the 90%, 95%, and 99% quantiles of historical data to divide the odor pollution area. Combining the spatial coordinate information, the heat map data is corresponding to the actual spatial position through rasterization processing to generate a list of key areas of concern, including area numbers, location descriptions, and odor concentration values. Extract the filter replacement frequency and pipeline cleaning cycle data from the air conditioner maintenance record database, and use the exponential decay function to establish the filter efficiency decay model as E(t) = E0e^(-kt), where E(t) is the efficiency at time t, E0 is the initial efficiency, and k is the decay coefficient. The pipeline pollution accumulation model uses the logarithmic growth function as P(t) = a ln(t + 1), where P(t) is the pollution degree at time t, and a is the pollution rate coefficient. Calculate the remaining efficiency of the current filter and the pipeline pollution degree, associate the results with the list of key areas of concern, and pass them to the next step. Use the multi-factor weighted scoring algorithm, comprehensively considering the odor concentration weight of 0.4, the filter efficiency weight of 0.2, the pipeline pollution degree weight of 0.2, and the microbial growth risk weight of 0.2. The microbial growth risk data is obtained from the aforementioned microbial assessment module and normalized to 0-100 points. The disinfection priority score is calculated for each area, and the calculation formula is

[0036] Score = 0.4O + 0.2F + 0.2P + 0.2M, where O, F, P, and M are the standardized odor concentration, filter efficiency, pipeline pollution level, and microbial risk score, respectively. The top 20% of the areas with the highest scores are identified as key disinfection areas. According to the odor types and microbial species in the key disinfection areas, applicable disinfection schemes are screened from the disinfectant database. The decision tree algorithm is used to select the optimal disinfection scheme. The construction of the decision tree includes: setting the root node as the odor type, the second-layer node as the main microbial species, the third-layer node as the environmental sensitivity, and the leaf node as the recommended disinfectant. The optimal splitting feature is selected based on the information gain ratio, and the pruning algorithm is used to avoid overfitting. Finally, a disinfection guidance scheme including the disinfectant name, usage concentration, and application method is generated. In the operating room area of a certain hospital, the odor diffusion heat map shows the formaldehyde concentration distribution. Based on historical data analysis, the low, medium, and high thresholds are set at 0.05 ppm, 0.08 ppm, and 0.1 ppm, respectively. Through rasterization, a 20m × 30m area is divided into 400 cells of 0.5m × 0.5m, and each cell corresponds to a concentration value. 15 key areas of concern are identified, and the concentration in Area 3 (coordinates 5,7) reaches 0.09 ppm. The air conditioner maintenance record shows that the filter was last replaced 30 days ago and the pipeline was cleaned 90 days ago. Using the filter efficiency decay model E(t) = 100e^(-0.02t), the current filter efficiency is calculated to be 54.8%. The pipeline pollution accumulation model P(t) = 10ln(t + 1) calculates the pollution level to be 45.2. The microbial assessment report shows that the Escherichia coli risk index in this area is 75 out of 100. The multi-factor weighted scoring algorithm calculates the score of Area 3 to be 0.4×90 + 0.2×54.8 + 0.2×(100 - 45.2) + 0.2×75 = 73.92, ranking second, and it is identified as a key disinfection area. The decision tree algorithm selects the disinfection scheme based on the odor type being formaldehyde, the microbial species being Escherichia coli, and the environmental sensitivity being high in the operating room. At the first layer node of the tree, the "odor type" selects the "aldehyde" branch, at the second layer, the "microbial species" selects the "bacteria" branch, and at the third layer, the "environmental sensitivity" selects the "high" branch. Finally, it is recommended to use a 0.5% peracetic acid solution, atomize and spray, with an action time of 30 minutes.

[0037] Step S106, according to the selected disinfection scheme, extract the chemical components of the disinfectant used in the disinfection scheme. Through chemical reaction simulation, evaluate the potential reaction of the chemical components with the known odor components in the air. According to the evaluation results, if toxic and harmful substances will be produced, optimize the disinfection scheme to avoid secondary pollution.

[0038] Obtain information on the selected disinfection plan in the disinfection plan database, where the information includes the name of the disinfectant, chemical composition, concentration, and usage method; construct a three-dimensional molecular model that contains the chemical composition and odor components of the disinfectant; run a chemical reaction simulation based on the three-dimensional molecular model to obtain reaction product information; match the reaction product information with the toxic and hazardous substances database to determine whether there are potential toxic and hazardous substances; if there are potential toxic and hazardous substances, calculate their generated concentration and compare it with a preset safety threshold; obtain a hazard degree assessment result based on the comparison result, where the hazard degree assessment result includes the substance name, predicted concentration, and exceeded standard multiple; if the hazard degree assessment result determines that toxic and hazardous substances exceeding the safety threshold will be generated, start a disinfection plan optimization program based on the genetic algorithm; the disinfection plan optimization program generates new alternative disinfection plans by adjusting the type, concentration, or usage method of the disinfectant.

[0039] Exemplarily, detailed information of the selected disinfection scheme is extracted from the disinfection scheme database, including the name of the disinfectant, chemical composition, concentration, and usage method. For complex disinfectant mixtures, high-performance liquid chromatography-mass spectrometry is used for component decomposition and quantitative analysis. At the same time, information on known odor components in the air is obtained from the odor component database to generate a list of chemical components to be evaluated. Using the GROMACS molecular dynamics simulation software, a three-dimensional molecular model containing the chemical components of the disinfectant and odor components is constructed. The simulation environment parameters are set, including a temperature of 25 °C, a relative humidity of 60%, and a standard atmospheric pressure. The chemical reaction simulation program is run, and the ReaxFF reaction force field is used to calculate the reaction probability between components and the possible reaction products. The simulation duration is set to 100 nanoseconds, and the time step is 0.1 femtosecond. The reaction product information obtained from the simulation is matched with the toxic and harmful substances database to determine whether there are potential toxic and harmful substances. If so, its generated concentration is calculated and compared with the preset safety threshold. The safety threshold refers to the World Health Organization's indoor air quality guidelines and national occupational health standards, and the more stringent standard value is selected. The hazard degree assessment result is obtained, including the substance name, predicted concentration, and exceeded standard multiple. The assessment result is transmitted to the next step in JSON format. According to the hazard degree assessment result, if it is judged that toxic and harmful substances exceeding the safety threshold will be generated, a disinfection scheme optimization program based on the genetic algorithm is started. By adjusting the type, concentration, or usage method of the disinfectant, a new alternative disinfection scheme is generated. The optimization objective function considers three factors: disinfection effect, risk of harmful substance generation, and economic cost. The population size is set to 100, the number of iterations is 50, the crossover probability is 0.8, and the mutation probability is 0.1. The optimal scheme of each round of iteration is returned for evaluation until the optimal disinfection scheme that does not generate harmful substances is found. In the evaluation of the disinfection scheme for an operating room in a hospital, the selected 0.5% peracetic acid solution disinfection scheme was extracted. Analysis by high-performance liquid chromatography-mass spectrometry showed that the main components of this solution were 0.5% peracetic acid, 0.3% acetic acid, and 0.2% hydrogen peroxide. The odor component database showed that there were 0.05 ppm of formaldehyde and 0.01 ppm of benzene in the air. The GROMACS software constructed a three-dimensional molecular model containing these components and performed a 100-nanosecond simulation at 25 °C, 60% relative humidity, and 1 standard atmospheric pressure. The ReaxFF force field calculation showed that the reaction probability between peracetic acid and formaldehyde was 0.15, and formic acid and acetic acid might be generated. The simulated predicted concentration of formic acid could reach 0.02 ppm. Comparing with the World Health Organization's indoor air quality guidelines, the safety threshold of formic acid was 0.01 ppm. The assessment result was transmitted in JSON format as {"substance": "formic acid

[0040] ","concentration":0.02,"threshold":0.01,"exceed_ratio":2}. The genetic algorithm optimization program starts with a population size of 100 and 50 iterations. The objective function is set as f = 0.5 disinfection effect - 0.3 harmful substance risk - 0.2 * cost. The optimal solution found in the 27th iteration is a 0.3% peracetic acid solution combined with ultraviolet irradiation. This solution is simulated again by GROMACS, predicting that the formic acid concentration drops to 0.008 ppm, which is lower than the safety threshold, while maintaining a 95% disinfection effect. A new disinfection guidance plan is generated, including solution preparation and ultraviolet irradiation parameters.

[0041] Analyze the selected disinfection plan, extract the active chemical components in all disinfectants, and based on the odor component information, obtain a list of chemical substances of all odor components known to exist in the air. Predict all chemical reactions that can occur between the disinfectant components and the odor components, and analyze and identify potential toxic and harmful substances.

[0042] Obtain the chemical composition information of the disinfectant and the odor component information known to exist in the air; generate a comprehensive chemical substance list according to the chemical composition information and the odor component information. Determine the chemical reactions that can occur between the disinfectant components and the odor components according to the molecular structures and reaction activities of the components in the comprehensive chemical substance list, and generate a list of potential reaction products including reactants, products, and reaction conditions. Use the quantitative structure-activity relationship model to conduct toxicity assessments for each compound in the list of potential reaction products; classify the potential reaction products according to the toxicity assessment results to obtain a list of potential toxic and harmful substances. Based on the computational fluid dynamics model, combined with the disinfectant dosage, odor component concentration, and environmental parameters, quantitatively predict the generation concentration of potential toxic and harmful substances; simulate the diffusion process of toxic and harmful substances in the air and calculate the equilibrium concentration of toxic and harmful substances at each point in space. Compare the equilibrium concentration with the preset toxicological threshold; if the equilibrium concentration exceeds the toxicological threshold, it is determined as a high-risk toxic and harmful substance; conduct uncertainty analysis and sensitivity analysis on the high-risk toxic and harmful substances.

[0043] Exemplarily, according to the selected disinfection protocol, the chemical composition information of the disinfectant is extracted from the disinfectant database, and the high performance liquid chromatography-mass spectrometry technique is used to analyze the components of the complex disinfectant mixture to identify and extract the active chemical components. At the same time, all the odor component information known to exist in the air is obtained to generate a comprehensive chemical substance list containing the disinfectant active components and odor components. Density functional theory calculations are used to predict all possible chemical reactions between the disinfectant components and odor components based on the molecular structures and reaction activities of the components in the comprehensive chemical substance list. Considering the reaction kinetics and thermodynamics factors, a list of potential reaction products is generated, including reactants, products, reaction conditions, reaction probabilities, and Gibbs free energy changes. The list of potential reaction products is passed to the next step in JSON format. A quantitative structure-activity relationship model is used to evaluate the toxicity of each compound in the list of potential reaction products, including acute toxicity, carcinogenicity, teratogenicity, and mutagenicity. According to the prediction results, the reaction products are classified into hazard levels to generate a list of potential toxic and harmful substances. A random forest algorithm is used to verify and optimize the prediction results to improve the prediction accuracy. Based on the computational fluid dynamics model, combined with the disinfectant dosage, the concentration of odor components in the air, and environmental parameters, the generation concentration of potential toxic and harmful substances is quantitatively predicted. The diffusion process of toxic and harmful substances in the air is simulated, and their equilibrium concentrations at each point in space are calculated. The predicted concentrations are compared with the relevant toxicological thresholds, and uncertainty analysis and sensitivity analysis are performed to finally determine the high-risk toxic and harmful substances that need to be focused on. In the evaluation of the disinfection protocol for an operating room in a hospital, a 0.5% peracetic acid solution disinfectant was analyzed. High performance liquid chromatography-mass spectrometry analysis showed that the main components of this solution were 0.5% peracetic acid, 0.3% acetic acid, and 0.2% hydrogen peroxide. The odor component database showed that there were 0.05 ppm of formaldehyde and 0.01 ppm of benzene in the air. Density functional theory calculations using the B3LYP / 6-31G(d) basis set predicted that the Gibbs free energy change for the reaction between peracetic acid and formaldehyde was -15.3 kJ / mol, and the reaction probability was 0.78. The main reaction product was formic acid, and the predicted concentration was 0.03 ppm. Evaluation by the quantitative structure-activity relationship model showed that the acute toxicity LD50 of formic acid was 1100 mg / kg, which was lower than the high-risk level threshold of 50 mg / kg. The random forest algorithm, with 500 trees and a maximum depth of 10, verified the prediction results with an accuracy of 95%. The computational fluid dynamics model used the k-ε turbulence model to simulate an operating room space of 20 m × 15 m × 3 m. The results showed that under ventilation conditions with 6 air change rates per hour, the maximum concentration of formic acid occurred within 1.5 m of the disinfection point and was 0.02 ppm, which was lower than the occupational exposure limit of 5 ppm. Sensitivity analysis showed that the ventilation conditions had the greatest impact on the concentration distribution, and increasing the air change rate to 8 times per hour could reduce the maximum concentration to 0.015 ppm.Formic acid is finally determined to be a substance with medium risk, and it is recommended to strengthen ventilation during disinfection.

[0044] Step S107: According to the optimized disinfection plan, control the air-conditioning system to enter the disinfection mode, adjust the fresh air intake ratio and the filtration efficiency level, and spray disinfectant in key areas, and monitor the disinfection effect in real time.

[0045] Switch the air-conditioning system to the disinfection mode, where the disinfection mode includes the fresh air intake ratio and the filter efficiency level; among them, the fresh air intake ratio is obtained by comparing the indoor and outdoor air quality indices, and the filter efficiency level is adjusted to a preset level. Receive the spraying instruction for the key area, and spray the disinfectant according to the spraying instruction at preset time intervals and doses; obtain environmental parameter data, which is collected by a sensor network distributed in each area; among them, the sensor network includes temperature and humidity sensors, VOC sensors, particulate matter sensors, and special disinfectant concentration sensors. Calculate the disinfection effect index, which is based on the real-time environmental parameter data and a pre-established disinfection effect evaluation model. Determine whether the disinfection effect index meets the preset conditions; if the disinfection effect index does not meet the preset conditions, then adjust the air-conditioning parameters and the disinfectant spraying strategy; among them, the air-conditioning parameters include the fresh air intake ratio and the filter efficiency level, and the disinfectant spraying strategy includes the spraying time interval and the dose.

[0046] Exemplarily, according to the optimized disinfection plan, send an instruction to the air-conditioning control system to switch the air-conditioning system to the disinfection mode. Based on the comparison of the indoor and outdoor air quality indices, calculate the fresh air intake ratio. When the outdoor air quality index is lower than the indoor one, the fresh air ratio is increased to 80%, and vice versa to 20%. At the same time, raise the filter efficiency level to MERV13 to achieve an 85% filtration rate for particles of 0.3 - 1.0 microns, balancing air purification and disinfectant diffusion. Through remote control, activate the automatic spraying device in the key area, and spray the disinfectant at a preset time interval of every 30 minutes and a dose of 5 ml / m 2, accurately spray disinfectants. Spraying information is transmitted to the central control unit in real time for subsequent environmental parameter analysis and disinfection effect evaluation. Using the sensor network distributed in various areas, environmental parameter data is collected in real time. The sensors include temperature and humidity sensors, VOC sensors, particulate matter sensors and special disinfectant concentration sensors, with one set installed every 100 square meters. The collected data include temperature, humidity, air flow velocity, VOC concentration, PM2.5 concentration and disinfectant concentration. The data is collected every 5 seconds and transmitted to the central processing unit. The long short-term memory network (LSTM) algorithm is used to dynamically calculate the disinfection effect index by combining real-time environmental parameter data and a pre-established disinfection effect evaluation model. The disinfection effect evaluation model is trained based on historical data and considers the relationship between environmental parameters and microbial survival rate. At the same time, air microbial samples are collected every hour through an automatic sampling device, and the microbial content is detected using rapid PCR technology to directly evaluate the disinfection effect. According to the calculation results, the air conditioning parameters and disinfectant spraying strategy are automatically adjusted to achieve closed-loop control of the disinfection process. During the disinfection process of an operating room in a hospital, the air conditioning control system receives the optimized disinfection scheme instruction and switches to disinfection mode. Real-time monitoring shows that the outdoor air quality index is 75 and the indoor air quality index is 120. The fresh air introduction ratio is adjusted to 80%. At the same time, the filter efficiency is upgraded to MERV13, and the filtration rate of 0.3-1.0 micron particles reaches 85%. Then the automatic spraying device in the key areas is activated, spraying 0.5% peracetic acid solution every 30 minutes, with a dosage of 5ml / m 2 . Five sets of sensors distributed in the 100-square-meter operating room collected data in real time, showing a temperature of 22°C, a relative humidity of 60%, and a VOC concentration of 0.3 mg / m 3 , PM2.5 concentration 15μg / m 3 , disinfectant concentration 0.2mg / m 3 . These data are transmitted to the central processing unit every 5 seconds. Based on these parameters and the pre-trained model, the LSTM algorithm calculates the current disinfection effect index to be 0.85, with a full score of 1.0. At the same time, the automatic sampling device collects air samples, and the rapid PCR test results show that the total number of bacteria is 50 CFU / m 3 , lower than 500CFU / m before disinfection 3 Based on these results, the air conditioning temperature was automatically lowered by 1°C and the relative humidity was increased by 5% to optimize the disinfectant effect; the final disinfection effect index stabilized at 0.92 and the total bacterial count dropped to 20 CFU / m 3 , meeting the disinfection standards of operating rooms.

[0047] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than limiting it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for the whole-process supervision of medical waste based on an intelligent medical waste monitoring platform, characterized in that The method includes: according to the classification of medical waste types, using a distributed sensor network to conduct odor sampling on the waste transfer paths in different areas to obtain odor data, including odor concentration data and odor component information; matching the collected odor data with the pre-established medical waste odor component data to determine the type of medical waste to which the odor source belongs and determining whether the odor concentration exceeds a preset threshold; if the odor concentration exceeds the standard, obtaining microbial information based on the correlation between the odor component information and microorganisms, and combining the temperature and humidity control parameters of the air conditioner to evaluate the risk of microbial growth; simulating the air flow distribution of the air conditioning system, combining the layout of the air conditioner outlet positions, predicting the diffusion path and influence range of the odor in different areas of the hospital, and generating an odor diffusion heat map; according to the odor diffusion heat map, combining factors such as the air conditioner filter replacement frequency and the pipeline cleaning cycle, determining the areas that need to be disinfected key points, and at the same time selecting a disinfection plan for treating the odor and microorganisms in the areas that need to be disinfected key points based on the risk of microbial growth; according to the selected disinfection plan, extracting the chemical components of the disinfectant used in the disinfection plan, simulating through chemical reactions, evaluating the potential reaction of the chemical components with the known odor components in the air, and according to the evaluation results, optimizing the disinfection plan if toxic and harmful substances will be generated to avoid secondary pollution; according to the optimized disinfection plan, controlling the air conditioning system to enter the disinfection mode, adjusting the fresh air intake ratio and the filtration efficiency level, and spraying disinfectant in key areas, and monitoring the disinfection effect in real time.

2. The method according to claim 1, wherein The step of, according to the classification of medical waste types, using a distributed sensor network to conduct odor sampling on the waste transfer paths in different areas to obtain odor data, including odor concentration data and odor component information, includes: obtaining the sampling area information divided according to the pre-established medical waste classification standard, setting multiple odor sampling points in the sampling area to form a distributed sensor network covering the waste transfer path; using a gas sensor array to conduct real-time detection on the odor sampling points to obtain the odor concentration and component data of each sampling point; fusing different types of sensor data according to the weighted average algorithm to determine the comprehensive odor data; performing preprocessing on the comprehensive odor data to eliminate environmental noise interference through a filtering algorithm. Reducing the dimension of the multi-dimensional odor feature data and extracting an odor fingerprint feature vector containing key parameters such as the gas component concentration ratio, odor intensity, and duration.

3. The method according to claim 1, wherein, Matching the collected odor data with the pre-established medical waste odor component data to determine the type of medical waste to which the odor source belongs and to determine whether the odor concentration exceeds a preset threshold, including: obtaining the odor data collected by a multi-channel gas sensor array, where the odor data includes the outputs of metal oxide, electrochemical, and photoionization sensors; dynamically adjusting weights according to the accuracy of each sensor in historical data, and using the weighted average method to fuse the odor data; performing normalization processing on the fused data to obtain an odor feature vector in a unified format; extracting a feature template from the pre-established medical waste odor component database, where the feature template contains the typical odor components and concentration ranges of different types of medical waste; performing dimensionality reduction processing on the odor feature vector, and selecting the first N principal components with a cumulative contribution rate reaching a preset value as the key odor component information; using a support vector machine classifier to classify the odor feature vector, where the support vector machine classifier uses a radial basis function kernel; optimizing the kernel parameters and penalty factor of the radial basis function kernel through cross-validation; if the output result of the support vector machine classifier matches the preset category, then determining the type of medical waste of the odor source; Establishing a preset threshold database, where the preset threshold database contains multiple levels of odor concentration thresholds; using a cubic spline interpolation algorithm to convert the original sensor signal into a standardized odor concentration value; determining whether the standardized odor concentration value exceeds the corresponding threshold in the preset threshold database; if the standardized odor concentration value exceeds the preset threshold, then triggering an alarm at the corresponding level according to the degree of exceeding the standard.

4. The method according to claim 1, wherein, If the odor concentration exceeds the standard, then obtaining microbial information based on the relevance between the odor component information and microorganisms, and combining the temperature and humidity control parameters of the air conditioner to evaluate the risk of microbial growth, including: obtaining the odor concentration detection result, and determining whether it exceeds the preset threshold according to the detection result, if the detection result exceeds the preset threshold, then triggering the subsequent evaluation process; using the Apriori algorithm to analyze the relevance between the odor component information and microorganisms, where the support threshold of the Apriori algorithm is a preset value, and the confidence threshold of the Apriori algorithm is a preset value; screening out the microbial species highly relevant to the odor components according to the relevance analysis result to obtain the growth conditions and reproduction speed information of the microbial species; establishing a microbial growth prediction model, where the prediction model calculates the potential growth rate of microorganisms based on the growth conditions, the reproduction speed information, and the temperature and humidity parameters; determining whether the potential growth rate exceeds the preset risk level threshold, if the potential growth rate exceeds the preset risk level threshold, then determining the microbial growth risk level; The risk levels include low risk, medium risk, and high risk; it also includes: querying in a microbial database according to the odor component information to determine the microbial species associated with the odor components, extracting the growth suitable condition information for each identified microorganism, obtaining in real time the temperature and humidity control parameters of the current air conditioner, including the temperature setting value, the humidity control level, and the fresh air intake ratio, and calculating the comprehensive breeding risk index for each microorganism according to the microbial growth suitable condition information and the temperature and humidity control parameters of the current air conditioner.

5. The method according to claim 4, wherein, The querying in a microbial database according to the odor component information to determine the microbial species associated with the odor components, extracting the growth suitable condition information for each identified microorganism, obtaining in real time the temperature and humidity control parameters of the current air conditioner, including the temperature setting value, the humidity control level, and the fresh air intake ratio, and calculating the comprehensive breeding risk index for each microorganism according to the microbial growth suitable condition information and the temperature and humidity control parameters of the current air conditioner includes: obtaining the odor component information, calculating the cosine similarity, where the cosine similarity is used to identify the associated microorganisms with a matching degree exceeding a preset threshold; obtaining the unique identifier of the associated microorganisms, and querying the microbial growth suitable condition information using the unique identifier; the growth suitable condition information includes the temperature range, the relative humidity range, the pH value range, and the oxygen demand; obtaining the temperature and humidity control parameters, where the temperature and humidity control parameters include the temperature setting value, the humidity control level, and the fresh air intake ratio; normalizing the temperature and humidity control parameters to a normalized value; using a multi-factor weighted scoring model to compare the growth suitable condition information with the normalized value; in the multi-factor weighted scoring model, preset weight values are assigned to the temperature factor, the humidity factor, the pH value factor, and the oxygen demand factor respectively. Judging the harm degree of the associated microorganisms. If the harm degree of the associated microorganisms exceeds the preset threshold, then calculating the comprehensive breeding risk index for the associated microorganisms. The comprehensive breeding risk index is determined by the product of the output result of the multi-factor weighted scoring model and the harm degree.

6. The method according to claim 1, wherein Simulating the air flow distribution of the air conditioning system, combining with the layout of the air conditioner outlet positions, predicting the diffusion path and influence range of the odor in different areas of the hospital, and generating an odor diffusion heat map, including: obtaining the hospital building information model, constructing a three-dimensional digital model according to the building information model; dividing the three-dimensional digital model into grid cells, recording the spatial coordinates, temperature, humidity, and air flow information of the grid cells; performing computational fluid dynamics simulation on the grid cells to obtain the air flow field distribution data; storing the air flow field distribution data in the form of a three-dimensional matrix; receiving the odor source position and initial concentration information; calculating the odor molecule diffusion process using the Euler-Lagrange hybrid model according to the odor source position, initial concentration information, and air flow field distribution data. Obtain time-series concentration distribution data; analyze the relationship between the airflow field distribution data and the time-series concentration distribution data; optimize the position and angle of the air-conditioning outlet according to the relationship; generate a continuous concentration distribution field based on the optimized outlet layout; map the continuous concentration distribution field to the color space to generate an odor diffusion heat map; superimpose the odor diffusion heat map on the hospital floor plan to display the odor diffusion path and impact range.

7. The method according to claim 1, wherein The method comprises: obtaining odor diffusion thermodynamic map data, setting odor concentration thresholds according to historical data quantiles, and dividing odor pollution areas; matching thermodynamic map data with actual spatial positions through rasterization to obtain a list of key areas of concern; establishing a filter efficiency attenuation model and a pipeline pollution accumulation model according to filter replacement frequency and pipeline cleaning cycle data; and adopting a logarithmic growth function of P(t)=aln(t+1) for the pipeline pollution accumulation model, where P(t) is the pollution degree at time t and a is the pollution rate coefficient. Calculate the current remaining filter efficiency and pipeline contamination level, and associate the results with the list of key areas of concern; use a multi-factor weighted scoring method to score odor concentration, filter efficiency, pipeline contamination level and microbial breeding risk; if the scoring result is within the preset percentage, it will be determined as a key disinfection area; select applicable disinfection plans based on the odor type and microbial species in the key disinfection area; construct a decision tree with the root node being the odor type, the second-level nodes being the main microbial species, and the third-level nodes being the environmental sensitivity; select the optimal disinfection plan through the decision tree, and generate a disinfection guidance plan including the name of the disinfectant, usage concentration and application method.

8. The method according to claim 1, wherein, The method extracts the chemical composition of the disinfectant used in the disinfection scheme according to the selected disinfection scheme, evaluates the potential reaction of the chemical composition with the known odor components in the air through chemical reaction simulation, and optimizes the disinfection scheme according to the evaluation result if toxic and harmful substances will be generated to avoid secondary pollution, including: obtaining information of the selected disinfection scheme in the disinfection scheme database, the information including the name, chemical composition, concentration and usage of the disinfectant; constructing a three-dimensional molecular model, the three-dimensional molecular model including the chemical composition and odor components of the disinfectant; running a chemical reaction simulation according to the three-dimensional molecular model to obtain reaction product information; matching the reaction product information with the toxic and harmful substance database to determine whether there are potential toxic and harmful substances; If there are potential toxic and hazardous substances, their generated concentration is calculated and compared with the preset safety threshold; based on the comparison result, a hazard level assessment result is obtained, which includes the substance name, predicted concentration and the excess multiple; if the hazard level assessment result determines that toxic and hazardous substances exceeding the safety threshold will be generated, a disinfection scheme optimization program based on a genetic algorithm is started; The disinfection plan optimization program generates new alternative disinfection plans by adjusting the type and concentration of disinfectants or the usage method; it also includes: analyzing the selected disinfection plan, extracting the active chemical components in all disinfectants, obtaining a list of chemical substances of all odor components known to exist in the air based on the odor component information, predicting all chemical reactions that can occur between the disinfectant components and the odor components, and analyzing and identifying potential toxic and harmful substances.

9. According to the method of claim 8, wherein, The analysis of the selected disinfection plan, extracting the active chemical components in all disinfectants, obtaining a list of chemical substances of all odor components known to exist in the air based on the odor component information, predicting all chemical reactions that can occur between the disinfectant components and the odor components, and analyzing and identifying potential toxic and harmful substances includes: obtaining the chemical composition information of the disinfectant and obtaining the odor component information known to exist in the air; generating a comprehensive chemical substance list according to the chemical composition information and the odor component information; According to the molecular structure and reactivity of each component in the comprehensive chemical substance list, determining the chemical reactions that can occur between the disinfectant components and the odor components, generating a list of potential reaction products including reactants, products, and reaction conditions; using the quantitative structure-activity relationship model to conduct toxicity assessments for each compound in the list of potential reaction products; classifying the hazard levels of the potential reaction products according to the toxicity assessment results to obtain a list of potential toxic and harmful substances; Based on the computational fluid dynamics model, combined with the disinfectant dosage, odor component concentration, and environmental parameters, quantitatively predicting the generation concentration of potential toxic and harmful substances; Simulating the diffusion process of toxic and harmful substances in the air, calculating the equilibrium concentration of toxic and harmful substances at each point in space; comparing the equilibrium concentration with a preset toxicological threshold; if the equilibrium concentration exceeds the toxicological threshold, it is determined as a high-risk toxic and harmful substance; conducting uncertainty analysis and sensitivity analysis on the high-risk toxic and harmful substances.

10. The method according to claim 1, wherein According to the optimized disinfection plan, control the air conditioning system to enter the disinfection mode, adjust the fresh air intake ratio and the filtration efficiency level, and spray disinfectant in key areas, and monitor the disinfection effect in real time, including: switching the air conditioning system to the disinfection mode, where the disinfection mode includes the fresh air intake ratio and the filtration efficiency level; wherein, the fresh air intake ratio is obtained by comparing the indoor and outdoor air quality indexes, and the filtration efficiency level is adjusted to a preset level; receiving a spraying instruction for key areas, and spraying disinfectant according to the spraying instruction at preset time intervals and doses; obtaining environmental parameter data, where the environmental parameter data is collected by a sensor network distributed in each area; wherein, the sensor network includes temperature and humidity sensors, VOC sensors, particulate matter sensors, and special disinfectant concentration sensors; calculating a disinfection effect index, where the disinfection effect index is based on real-time environmental parameter data and a pre-established disinfection effect evaluation model; determining whether the disinfection effect index meets a preset condition; if the disinfection effect index does not meet the preset condition, then adjust the air conditioning parameters and the disinfectant spraying strategy; wherein, the air conditioning parameters include the fresh air intake ratio and the filtration efficiency level, and the disinfectant spraying strategy includes the spraying time interval and the dose.

Images