Disease detection, prevention and control method and system based on dynamic monitoring data

Abstract

The invention discloses a disease detection, prevention and control method and system based on dynamic monitoring data. The method comprises the steps that traffic dust, heat-electricity, cable ozone, micrometeorology and individual respiration exposure data are collected through roadside multiple sensors, a photovoltaic inverter power monitoring module, a cable connector arc sensor, a distributed micro-meteorological station and wearable respiration monitoring equipment, so that initial data are obtained; performing preprocessing and quality control on the initial data, extracting and fusing respiratory tract stress features by using a self-attention mechanism, a genetic algorithm and a health degree mapping function, and generating a health risk score; and triggering graded early warning according to the health risk score, and predicting a doctor seeing peak. By implementing the method, the risk of the respiratory disease can be accurately early warned and closed-loop intervention can be implemented by fusing dynamic monitoring data of infrastructures, the method has the advantages of being accurate in monitoring, timely in early warning, cooperative in prevention and control and the like, and the defects of traditional disease monitoring, prevention and control are effectively overcome.

Description

Technical Field

[0001] This invention relates to big data, and more specifically to a method and system for disease detection and prevention based on dynamic monitoring data. Background Technology

[0002] With the acceleration of urbanization, urban residents face numerous health challenges, among which acute respiratory diseases have become a significant public health issue. Traditional methods for monitoring and controlling respiratory diseases mainly rely on hospital case reporting and limited environmental monitoring stations, which suffer from significant delays and insufficient spatial coverage, making it difficult to meet the needs for timely and precise prevention and control.

[0003] Currently, there is a lack of technologies for collaborative monitoring and intervention of respiratory disease risks caused by the operation of infrastructure such as transportation, energy, and electricity in urban environments. Integrating these scattered monitoring technologies and data, which serve different engineering fields, into early warning indicators for respiratory disease risks, and achieving closed-loop intervention through multi-departmental collaboration, is a pressing technical challenge and a key to improving urban public health prevention and control capabilities.

[0004] Therefore, it is necessary to design a new method to accurately predict respiratory disease risks and implement closed-loop interventions by integrating dynamic monitoring data from infrastructure. This method has the advantages of accurate monitoring, timely early warning, and coordinated prevention and control, effectively making up for the shortcomings of traditional disease monitoring and prevention. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a disease detection and prevention method and system based on dynamic monitoring data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a disease detection and prevention method based on dynamic monitoring data, comprising:

[0007] Initial data were obtained by collecting traffic dust, heat-electricity, cable ozone, micro-meteorological and individual respiratory exposure data through roadside multi-sensors, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-meteorological stations and wearable respiratory monitoring devices.

[0008] The initial data is preprocessed and quality controlled. Respiratory stress features are extracted and fused using a self-attention mechanism, genetic algorithm, and health degree mapping function to generate a health risk score.

[0009] Based on the aforementioned health risk score, a tiered early warning system is triggered to predict peak periods for medical visits.

[0010] The further technical solution is as follows: after triggering a graded early warning based on the health risk score and predicting the peak of medical visits, it also includes:

[0011] Based on the level of the warning, implement intervention measures at the transportation, energy, environmental, and individual levels;

[0012] Data collected after intervention is fed back, and incremental learning is used to optimize the model, evaluate the intervention effect, and dynamically adjust the strategy.

[0013] The further technical solution is as follows: Preprocessing and quality control of the initial data, extracting and fusing respiratory stress features using a self-attention mechanism, genetic algorithm, and health mapping function to generate a health risk score, including:

[0014] The initial data is Z-score standardized to eliminate dimensional differences, and outliers are removed based on physical boundaries to form a five-dimensional index.

[0015] Based on the five-dimensional indicators, a current respiratory exposure vector is constructed, and a difference vector is calculated by combining historical data. The weight of each dimension is determined by using a self-attention mechanism to highlight key risk factors and determine the feature vector related to heat exposure.

[0016] Based on the heat exposure-related feature vectors, an optimization objective function containing node heat exposure terms and power loss terms is constructed. A genetic algorithm is used to optimize and calculate the heat exposure target value, and the switching state of the photovoltaic feeder segments is adjusted to reduce the risk of thermal-electric exposure, so as to obtain the optimized heat exposure feature vector.

[0017] The optimized heat exposure feature vector is input into a fully connected layer for deep fusion to generate a respiratory stress feature vector. Then, it is converted into a continuous risk score through a Softmax layer and discretized into four risk levels to obtain a health risk score.

[0018] The further technical solution is as follows: Based on the five-dimensional indicators, a current respiratory exposure vector is constructed; combined with historical data, a difference vector is calculated; and a self-attention mechanism is used to determine the weights of each dimension, highlighting key risk factors, and determining heat exposure-related feature vectors, including:

[0019] The five-dimensional indicators are integrated into a vector to represent the respiratory exposure status at the current moment, so as to obtain the current respiratory exposure vector;

[0020] Collect five-dimensional index data from the same period of the past 7 days, calculate their mean, and construct a historical exposure vector to reflect the level of respiratory exposure in the same period of history.

[0021] The difference vector is obtained by comparing the current respiratory exposure vector with the historical exposure vector;

[0022] The historical exposure vector is input into the feature projection layer to extract the severity features of each dimension, thus obtaining the historical feature vector. Specifically, the differences of each factor in the historical data are calculated to obtain the historical factor change vector. The historical factor change vector is then used to calculate the inner product to obtain the inner product result. Finally, the weight values ​​of each dimension are calculated through the fully connected layer and the Softmax layer.

[0023] The difference vector is weighted based on the weight values ​​calculated by the self-attention mechanism. The dimensions related to heat exposure are then weighted to form a feature vector related to heat exposure.

[0024] The further technical solution is as follows: Based on the heat exposure-related feature vector, an optimization objective function containing node heat exposure terms and power loss terms is constructed. A genetic algorithm is used to optimize and calculate the heat exposure optimization objective value, and the switching state of the photovoltaic feeder segments is adjusted to reduce the risk of heat-electric exposure, thereby obtaining the optimized heat exposure feature vector, including:

[0025] Based on the heat exposure-related feature vectors, an optimization objective function containing node heat exposure terms and power loss terms is constructed.

[0026] Set the initial parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, and mutation probability;

[0027] An initial population is randomly generated, with each individual representing a possible state of a set of photovoltaic feeder segment switches;

[0028] Substitute each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness.

[0029] Based on the fitness of individuals, a selection strategy is used to select individuals with high fitness to enter the breeding pool. Individuals with higher fitness have a greater probability of being selected.

[0030] Crossover is performed on individuals in the breeding pool. Two individuals are randomly selected as parents, and some of their genes are exchanged according to the crossover probability to generate new offspring individuals.

[0031] The genes of offspring individuals are mutated with a certain probability of mutation.

[0032] The new individuals produced through crossover and mutation replace some individuals in the old population to form a new generation of population;

[0033] Determine if the termination condition of the genetic algorithm has been met, such as the maximum number of iterations or the fitness convergence threshold. If the termination condition is met, proceed to the next step; otherwise, return to the step of substituting each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness, and continue iterating.

[0034] The individual with the best fitness is selected from the final population, and the corresponding combination of on / off states is the target value for heat exposure optimization.

[0035] Based on the switching state combination corresponding to the thermal exposure optimization target value, the actual switching state of the feeder segments of the photovoltaic system is adjusted to reduce the thermal-electric exposure risk.

[0036] Under the adjusted photovoltaic system conditions, the heat exposure-related feature vectors are recalculated to obtain the optimized heat exposure feature vectors.

[0037] The further technical solution is as follows: the optimized heat exposure feature vector is input into a fully connected layer for deep fusion to generate a respiratory stress feature vector, which is then converted into a continuous risk score through a Softmax layer and discretized into four risk levels to obtain a health risk score, including:

[0038] The optimized thermal exposure feature vector is used as input to obtain other non-thermal exposure related feature vectors.

[0039] Design a fully connected neural network where the number of neurons in the input layer matches the total dimension of the feature vector, the hidden layer uses an appropriate activation function, and the number of neurons in the output layer matches the dimension of the respiratory stress feature vector.

[0040] The spliced ​​and optimized thermal exposure feature vector and other non-thermal exposure feature vectors are input into the fully connected layer network;

[0041] Forward propagation computation is performed through a fully connected layer to achieve deep feature fusion and obtain the respiratory stress feature vector;

[0042] A Softmax layer is constructed based on the output of the fully connected layer. The input of the Softmax layer is the respiratory stress feature vector, and the output is a probability distribution representing the probability of different risk levels.

[0043] The respiratory stress feature vector is transformed into a continuous risk score between 0 and 1 using a Softmax layer, which represents the probability of respiratory disease occurrence.

[0044] Based on the magnitude of the continuous risk score, it is discretized into four risk levels to obtain the health risk score.

[0045] Its further technical solution is as follows: Based on the warning level, intervention measures are implemented at the transportation, energy, environmental, and individual levels, including:

[0046] The current warning level is determined based on the risk level range of the generated health risk score;

[0047] If the warning level is orange or red, a signal is sent to the intelligent transportation system to dynamically adjust the traffic lights and reduce the speed limit to 20km / h.

[0048] When the warning level reaches orange or red, the energy management system is triggered to transfer the unstable feeder load to the adjacent transformer, reducing air conditioning downtime and lowering the risk of thermal and electrical exposure.

[0049] If the warning level is orange or red, environmental intervention measures will be initiated, and drones will be controlled to automatically spray 5% Na2S2O3 mist droplets to quickly reduce the local ozone concentration and reduce the respiratory irritation caused by ozone pollution from cable arcing.

[0050] Based on the warning level, corresponding health alerts will be pushed to the wearable devices of high-risk groups.

[0051] This invention also provides a disease detection and prevention system based on dynamic monitoring data, comprising:

[0052] The acquisition unit is used to collect traffic dust, heat-electricity, cable ozone, micro-meteorology and individual respiratory exposure data through roadside multi-sensors, photovoltaic inverter power monitoring module, cable joint arc sensor, distributed micro-weather station and wearable respiratory monitoring device to obtain initial data;

[0053] The generation unit is used for preprocessing and quality control of the initial data, and extracts and fuses respiratory stress features using a self-attention mechanism, a genetic algorithm and a health degree mapping function to generate a health risk score;

[0054] The early warning unit is used to trigger tiered early warnings based on the health risk score and predict peak periods for medical visits.

[0055] The present invention also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the above-described method.

[0056] The present invention also provides a storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0057] The advantages of this invention compared to existing technologies are as follows: This invention integrates traffic dust, heat-electricity, cable ozone, micro-meteorological, and individual respiratory exposure data collected by roadside multi-sensor systems, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-meteorological stations, and wearable respiratory monitoring devices. After preprocessing and quality control, it extracts and fuses respiratory stress features using a self-attention mechanism, genetic algorithm, and health mapping function to generate a health risk score. Based on this score, it triggers tiered early warnings and predicts peak medical visits, achieving precise early warning and closed-loop intervention for respiratory disease risks. This process fully utilizes the advantages of dynamic infrastructure monitoring data, possessing characteristics such as accurate monitoring, timely early warning, and coordinated prevention and control, effectively compensating for the shortcomings of traditional disease monitoring and prevention.

[0058] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0059] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart illustrating the disease detection and prevention method based on dynamic monitoring data provided in an embodiment of the present invention.

[0061] Figure 2 A schematic block diagram of a disease detection and prevention system based on dynamic monitoring data provided in an embodiment of the present invention;

[0062] Figure 3 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0065] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0066] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0067] Please see Figure 1 , Figure 1 This is a flowchart illustrating a disease detection and prevention method based on dynamic monitoring data provided in an embodiment of the present invention. This method, applied to a server, integrates dynamic data collected from multiple roadside sensors, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-weather stations, and wearable respiratory monitoring devices. It utilizes a self-attention mechanism, genetic algorithm, and health mapping function to extract and fuse respiratory stress characteristics, generating an accurate health risk score. Based on this score, it triggers tiered early warnings to predict peak patient visits. Furthermore, based on the warning level, it implements closed-loop intervention measures encompassing traffic, energy, environment, and individual levels, such as dynamically adjusting traffic lights, optimizing the segmented switching status of photovoltaic feeders, and using drones to spray and reduce ozone concentration. This achieves accurate monitoring, timely early warning, and effective prevention and control of disease risks, overcoming problems such as delayed response and lack of coordination in traditional disease monitoring and prevention methods.

[0068] Figure 1 This is a flowchart illustrating the disease detection and prevention method based on dynamic monitoring data provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S130.

[0069] S110 uses roadside multi-sensor, photovoltaic inverter power monitoring module, cable joint arc sensor, distributed micro-weather station and wearable respiratory monitoring device to collect traffic dust, heat-electricity, cable ozone, micro-weather and individual respiratory exposure data to obtain initial data.

[0070] In this embodiment, initial data refers to raw, unprocessed data collected by various monitoring devices. This data directly reflects various exposure factors related to respiratory diseases in the urban environment, as detailed below:

[0071] Traffic dust data: collected by multiple roadside sensors, including data on vehicle emergency braking and acceleration events, used to calculate the emergency braking rate (HBr) and traffic flow speed information. This data will be used for subsequent analysis of the impact of traffic disturbances on road dust, and in turn to assess the level of particulate matter exposure caused by dust.

[0072] Thermal-electric exposure data: from the photovoltaic inverter power monitoring module, including active power (P) and reactive power (Q) data of the photovoltaic system, used to detect photovoltaic power drop events and calculate the depth of power loss (PLD) to determine the possible shutdown of the air conditioning system and the degree of deterioration of indoor heat index and ventilation volume.

[0073] Cable ozone data: acquired through cable joint arc sensors, reflecting the arc energy (E_arc) at the cable joint, used to estimate the local ozone increment (ACE_O3) caused by the arc, revealing the impact of ozone and other harmful gases generated by cable arcs on air quality.

[0074] Micrometeorological data: collected by distributed micrometeorological stations, covering meteorological parameters such as ambient temperature (T) and relative humidity (RH), used to calculate airway dryness (VPD) and assess the potential irritant effect of meteorological conditions on the respiratory tract.

[0075] Individual respiratory data: Wearable respiratory monitoring devices collect an individual's respiratory rate (Br) to indicate whether the individual is in a state of high respiratory load (Br>20 breaths / min), thereby reflecting the individual's respiratory stress in the current environment.

[0076] These initial data cover key factors influencing respiratory diseases in urban environments, providing foundational information for subsequent health risk assessments and interventions.

[0077] Specifically, integrated monitoring terminals are deployed along main urban roads and around communities with a high incidence of asthma, integrating a 6-axis accelerometer, a high-definition camera, and a micro particulate matter sensor.

[0078] Accelerometers capture vehicle emergency braking and acceleration events to calculate the emergency braking rate (HBr).

[0079] Utilizing video AI to analyze traffic flow speed, combined with the HBr input re-suspension model: PM 2.5 = 5.4·HBr + 0.7·V -1 Output the Traffic Dust Index (TDI_PM).

[0080] Power monitoring modules are embedded in photovoltaic inverters in communities and commercial buildings to collect active power P and reactive power Q in real time.

[0081] When a sudden drop in photovoltaic power is detected (power loss depth PLD = (P0 - P) tIf the temperature reaches 15% (P0 > 15%), a heat exposure warning is triggered, indicating that the air conditioning system may shut down and the indoor heat index and ventilation volume may deteriorate.

[0082] An arc energy sensor is deployed at the underground cable joint to monitor the arc energy E_arc in real time.

[0083] Based on the experimental calibration curve (ΔO3=0.32·E_arc) 0.8 This allows for the rapid estimation of the local ozone increase (ACE_O3) caused by electric arc.

[0084] Deploy mini weather stations to monitor ambient temperature T and relative humidity RH, and calculate airway dryness (VPD = es(T) - ea).

[0085] High-risk individuals are equipped with chest strap wearable devices to monitor respiratory rate (Br) and mark high respiratory load states (Br>20 breaths / min).

[0086] S120. The initial data is preprocessed and quality controlled. Respiratory stress features are extracted and fused using a self-attention mechanism, genetic algorithm and health degree mapping function to generate a health risk score.

[0087] In this embodiment, the health risk score refers to a continuous value between 0 and 1 obtained by integrating the optimized heat exposure feature vector and other non-heat exposure related feature vectors, and transforming them through a fully connected layer and a Softmax layer. This value is used to quantify the potential risk of respiratory diseases and discretize them into four risk levels.

[0088] In one embodiment, step S120 described above may include steps S121 to S124.

[0089] S121. Perform Z-score standardization on the initial data to eliminate dimensional differences and remove outliers based on physical boundaries to form a five-dimensional index.

[0090] First, the collected data (including traffic dust index TDI_PM, heat-electric exposure index TEE_PLD, cable ozone exposure index ACE_O3, airway dryness VPD, and individual respiratory rate IV_Br) were Z-score standardized to eliminate dimensional differences and make the different indicators comparable. Next, outliers were removed based on the physical boundaries of each indicator; for example, airway dryness VPD was set to not exceed 5 kPa, and data exceeding this range were considered invalid or missing.

[0091] S122. Construct the current respiratory exposure vector based on the five-dimensional indicators, calculate the difference vector by combining historical data, and use the self-attention mechanism to determine the weight of each dimension, highlight key risk factors, and determine the feature vector related to heat exposure.

[0092] In this step, the current respiratory exposure vector E is constructed based on five-dimensional indicators. x (t), and simultaneously obtain the average value of the same period over the past 7 days to construct the historical exposure vector E. x (h). The difference vector ΔE is obtained by calculating the difference. x The self-attention mechanism is used to determine the weights of each dimension, highlighting key risk factors. Specific steps include:

[0093] The five-dimensional indicators are integrated to form the current respiratory exposure vector.

[0094] Construct a historical exposure vector to reflect the level of the same historical period.

[0095] Calculate the difference vector to quantify the changes between the present and the past.

[0096] The self-attention mechanism is used to calculate the weights of each dimension, especially to identify feature vectors related to heat exposure.

[0097] In one embodiment, step S122 may include steps S1221 to S1225.

[0098] S1221. Integrate the five-dimensional indicators into a vector to represent the respiratory exposure status at the current moment, so as to obtain the current respiratory exposure vector.

[0099] In this embodiment, traffic dust index (TDI_PM), thermal-electric exposure index (TEE_PLD), cable ozone exposure index (ACE_O3), airway dryness (VPD), and individual respiratory rate (IV_Br) data were obtained after Z-score normalization and outlier removal.

[0100] These five preprocessed indicators are arranged in a certain order and integrated into a vector, denoted as the current respiratory exposure vector E. x (t) = [TDI_PM, TEE_PLD, ACE_O3, VPD, IV_Br], which comprehensively reflects the respiratory exposure status within the monitoring area at the current moment.

[0101] S1222. Collect five-dimensional index data for the same period in the past 7 days, calculate their mean, and construct a historical exposure vector to reflect the respiratory exposure level in the same period in history.

[0102] In this embodiment, five-dimensional index data of the same time points in the past 7 days and the current time are collected from the data storage module, including traffic dust index, heat-electricity exposure index, cable ozone exposure index, airway dryness and individual respiratory rate.

[0103] The average value of each of the five indicators collected over the past 7 days was calculated to obtain the historical average value of each indicator.

[0104] The calculated averages of the five indicators are arranged in the same order as the current respiratory exposure vector to construct the historical exposure vector E. x (h) = [TDI_PM_h, TEE_PLD_h, ACE_O3_h, VPD_h, IV_Br_h], where the subscript h represents the historical mean. This vector reflects the level of respiratory exposure during the same historical period.

[0105] S1223. Obtain the difference vector by using the difference between the current respiratory exposure vector and the historical exposure vector.

[0106] In this embodiment, the current respiratory exposure vector E is... x (t) and historical exposure vector E x The corresponding index of (h) is subtracted, that is, TDI_PM is subtracted from TDI_PM_h, TEE_PLD is subtracted from TEE_PLD_h, and so on, to obtain the difference of each index.

[0107] Arrange the differences of the five indicators obtained from the above calculations in the same order to construct the difference vector ΔE. x =E x (t)-E x (h) = [ΔTDI_PM, ΔTEE_PLD, ΔACE_O3, ΔVPD, ΔIV_Br], which quantifies the changes in respiratory exposure compared to the same period in history.

[0108] S1224. Input the historical exposure vector into the feature projection layer, extract the severity features of each dimension, and obtain the historical feature vector. Specifically calculate the differences of each factor in the historical data to obtain the historical factor change vector. Perform inner product calculation on the historical factor change vector to obtain the inner product result, and calculate the weight values ​​of each dimension through the fully connected layer and the Softmax layer.

[0109] In this embodiment, the historical exposure vector E x (h) is input to the feature projection layer, which is a neural network layer used to extract features from the data.

[0110] Feature projection layer on historical exposure vector E x (h) is processed to extract the severity features of each dimension (i.e. each indicator) to obtain the historical feature vector.

[0111] Further analysis of historical data is conducted to calculate the differences among various factors in the historical data, resulting in a vector of historical factor changes. This vector reflects the changes of each factor in the historical data.

[0112] The inner product is calculated by multiplying the vector of historical factor changes with itself. The inner product operation can measure the correlation and difference between the elements in the vector and obtain the inner product result.

[0113] The inner product result is input into a fully connected layer, where neurons are connected to each neuron in the previous layer, undergoing linear transformation and nonlinear activation. Then, a Softmax layer is used to convert the output of the fully connected layer into a probability distribution, yielding weight values ​​for each dimension. These weight values ​​represent the relative importance of each indicator in respiratory disease risk assessment.

[0114] S1225. Based on the weight values ​​calculated by the self-attention mechanism, the difference vector is weighted, and the dimensions related to heat exposure are adjusted to form a feature vector related to heat exposure.

[0115] Obtain the weight values ​​of each dimension calculated through the self-attention mechanism from the Softmax layer.

[0116] The difference vector ΔE x Each dimension is multiplied by its corresponding weight value to obtain a weighted difference vector.

[0117] Dimensions related to heat exposure (such as the heat-electric exposure index TEE_PLD) were identified, and their corresponding parts were extracted from the weighted difference vector. After further weight adjustment and feature integration, a heat exposure-related feature vector was formed. This vector highlights the importance of heat exposure-related indicators in current changes in respiratory exposure, providing a feature basis for subsequent optimization steps targeting heat-electric exposure risk.

[0118] S123. Based on the heat exposure-related feature vector, construct an optimization objective function containing node heat exposure terms and power loss terms, use a genetic algorithm to optimize and calculate the heat exposure optimization objective value, and adjust the segmented switching state of the photovoltaic feeder to reduce the risk of heat-electric exposure, so as to obtain the optimized heat exposure feature vector.

[0119] Based on the heat exposure-related feature vectors obtained from the above steps, an optimization objective function containing node heat exposure and power loss terms is constructed. A genetic algorithm is then used to find the optimal solution to adjust the segmented switching state of the photovoltaic feeder and reduce the risk of thermal and electrical exposure. This process involves setting genetic algorithm parameters, generating an initial population, calculating fitness, selecting a breeding pool, performing crossover and mutation operations until the termination condition is met, and finally selecting the optimal individual from the final population as the solution.

[0120] In this embodiment, the optimized thermal exposure feature vector refers to the feature vector recalculated after optimizing and adjusting the photovoltaic feeder segment switching state by applying a genetic algorithm to the initial thermal exposure-related feature vector. This aims to effectively reduce the risk of thermal-electrical exposure and improve the operational safety of the system.

[0121] In one embodiment, step S123 described above may include steps S1231 to S12312.

[0122] S1231. Based on the heat exposure-related feature vector, construct an optimization objective function that includes a node heat exposure term and a power loss term.

[0123] In this embodiment, the heat exposure-related feature vector obtained from step S1225 is analyzed in depth to clarify the key information related to heat exposure contained therein, such as node heat exposure level and power loss.

[0124] Based on the needs of respiratory disease risk prevention and control, the optimization objectives are determined to reduce the risk of heat exposure and reduce power loss, so as to improve the operating efficiency and safety of photovoltaic systems, while reducing the adverse effects of heat-electric exposure on respiratory health.

[0125] Based on the heat exposure-related feature vectors, an optimization objective function is constructed that includes a node heat exposure term and a power loss term. This function can be expressed as minimizing the weighted sum of heat exposure risk and power loss, for example: minimize(w1*node heat exposure term + w2*power loss term), where w1 and w2 are the weight coefficients of the node heat exposure term and the power loss term, respectively, used to balance their importance in the optimization process.

[0126] S1232. Set the initial parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, and mutation probability.

[0127] In this embodiment, the population size of the genetic algorithm is determined based on the complexity of the optimization problem and the limitations of computational resources. A larger population size can increase population diversity and improve search capabilities, but it will also increase computational costs. For example, a population size of 100 can be chosen.

[0128] Set the maximum number of iterations for the genetic algorithm to control its runtime and convergence. For example, set the maximum number of iterations to 100.

[0129] Choosing an appropriate crossover probability is crucial, as it determines the probability that parent individuals will crossover to generate offspring. A higher crossover probability promotes gene exchange and the generation of new individuals, but an excessively high probability may slow down algorithm convergence. A crossover probability of 0.8-0.9 is typically recommended.

[0130] Determine the mutation probability, which refers to the probability that an individual's gene will mutate. An appropriate mutation probability can prevent the algorithm from getting trapped in local optima and maintain population diversity. Generally, a mutation probability of 0.01-0.1 can be set.

[0131] S1233. Randomly generate an initial population, where each individual represents a possible state of a set of photovoltaic feeder segment switches.

[0132] In this embodiment, each possible state of the photovoltaic feeder segment switch is encoded as an individual. For example, the open and closed state of the switch can be represented by binary code, with each switch corresponding to a binary bit, where 0 indicates that the switch is open and 1 indicates that the switch is closed.

[0133] Based on the determined encoding method, multiple individuals are randomly generated to form an initial population. The population size is determined according to the population size set in step S1232, for example, generating 100 random individuals as the initial population.

[0134] S1234. Substitute each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the fitness of the individual.

[0135] In this embodiment, each individual in the population (i.e., the encoded combination of switch states) is decoded to obtain the actual state of the corresponding photovoltaic feeder segment switch.

[0136] Based on the decoded switch state, the operation of the photovoltaic system is simulated, and the node heat exposure value and power loss value are calculated under the switch state.

[0137] The calculated node heat exposure and power loss values ​​are substituted into the optimization objective function constructed in step S1231 to calculate the fitness of each individual. The fitness value reflects the individual's performance in the optimization objective function; a lower fitness indicates a better switching state for that individual.

[0138] S1235. Based on the fitness of individuals, a selection strategy is used to select individuals with high fitness to enter the breeding pool. The higher the fitness of an individual, the greater the probability of it being selected.

[0139] In this embodiment, a suitable selection strategy is chosen, such as roulette wheel selection or tournament selection. In roulette wheel selection, the probability of an individual being selected is directly proportional to its fitness; the higher the fitness of an individual, the greater the probability of it being selected. In tournament selection, several individuals are randomly selected for comparison, and the individual with the highest fitness is selected to enter the breeding pool.

[0140] Create a breeding pool to store the selected individuals.

[0141] According to the selected selection strategy, individuals with high fitness are selected from the initial population to enter the breeding pool until the breeding pool reaches a certain size, for example, the size of the breeding pool can be set to be the same as the size of the initial population.

[0142] S1236. Perform a crossover operation on the individuals in the breeding pool, randomly select two individuals as parents, exchange some of their genes according to the crossover probability, and generate new offspring individuals.

[0143] In this embodiment, two individuals are randomly selected from the breeding pool as parents.

[0144] Based on the crossover probability set in step S1232, a decision is made as to whether to perform a crossover operation on the parent individuals. If the crossover condition is met, parts of their genes are exchanged to generate two new offspring individuals. For example, in the case of binary encoding, a crossover point can be selected, and the gene portions of the parent individuals after the crossover point can be exchanged.

[0145] New offspring individuals are obtained through crossover operations, and these offspring individuals inherit some of the genetic characteristics of the parent individuals.

[0146] S1237. Mutate the genes of offspring individuals with a certain probability of mutation.

[0147] In this embodiment, for each offspring individual, a determination is made as to whether to perform a mutation operation based on the mutation probability set in step S1232.

[0148] If the mutation condition is met, one or more gene positions in the offspring are randomly selected for mutation. In binary encoding, a 0 in a gene position is changed to a 1, or a 1 is changed to a 0.

[0149] The resulting offspring individuals, after undergoing mutation, possess new genetic characteristics, thus increasing the diversity of the population.

[0150] S1238. Replace some individuals in the old population with new individuals produced by crossover and mutation to form a new generation of population.

[0151] Determine a population renewal strategy, such as an elite retention strategy, which involves retaining a portion of the most fit individuals in the new generation of the population, while the remaining individuals are replaced by new individuals resulting from crossover and mutation.

[0152] Following a defined population renewal strategy, some individuals in the old population are replaced by new individuals generated through crossover and mutation, forming a new generation population. This new generation population retains some superior individuals while introducing new genetic traits, thus improving the overall fitness of the population.

[0153] S1239. Determine whether the termination condition of the genetic algorithm has been met, such as the maximum number of iterations or the fitness convergence threshold. If the termination condition is met, proceed to the next step; otherwise, return to the step of substituting each individual into the optimization objective function, calculating its corresponding heat exposure risk and power loss value to obtain the individual's fitness, and continue iterating;

[0154] Check whether the genetic algorithm has reached the set termination conditions, including whether it has reached the maximum number of iterations (e.g., 100 times) or whether the fitness has converged (i.e., the change in fitness value over several consecutive generations is less than the set convergence threshold, e.g., 0.001).

[0155] If the termination condition is not met, return to step S1234 and continue to perform fitness calculation, selection, crossover, mutation and other operations on the individuals in the new generation population to start a new round of iteration.

[0156] When the termination condition is met, the iterative process of the genetic algorithm ends and proceeds to the next step.

[0157] S12310. Select the individual with the best fitness from the final population. The corresponding combination of switching states is the target value for heat exposure optimization.

[0158] The fitness of all individuals in the final population is compared to identify the individual with the optimal fitness. The individual with the optimal fitness is the one that minimizes the objective function, and the corresponding combination of switching states can effectively reduce the risk of heat exposure and power loss.

[0159] The combination of switching states corresponding to the individual with the best fitness is determined as the target value for thermal exposure optimization. This target value provides a basis for adjusting the switching states of the photovoltaic feeder segments.

[0160] S12311. Based on the switching state combination corresponding to the thermal exposure optimization target value, actually adjust the switching state of the feeder segments of the photovoltaic system to reduce the risk of thermal-electric exposure.

[0161] The switch state combination corresponding to the thermal exposure optimization target value is decoded to obtain the specific opening and closing state of the photovoltaic feeder segment switch.

[0162] Based on the decoded switch status, the feeder sectionalizing switches of the photovoltaic system are operated and adjusted to their optimized states. By adjusting the switch status, the operating mode of the photovoltaic system can be changed, reducing the risk of heat exposure and power loss, improving the system's operating efficiency and safety, thereby reducing the adverse effects of thermal-electrical exposure on respiratory health.

[0163] S12312. Under the adjusted photovoltaic system state, recalculate the heat exposure-related feature vector to obtain the optimized heat exposure feature vector.

[0164] After adjusting the status of the photovoltaic feeder segment switches, the operating status of the photovoltaic system is monitored to obtain new data related to heat exposure, such as node heat exposure values ​​and power loss values.

[0165] Based on the new system status data, the heat exposure-related feature vector was recalculated. This optimized heat exposure feature vector reflects the changes in heat exposure-related indicators under the adjusted photovoltaic system status, verifying the effectiveness of the optimization measures and providing updated feature information for subsequent health risk assessment.

[0166] S124. The optimized heat exposure feature vector is input into a fully connected layer for deep fusion to generate a respiratory stress feature vector. Then, it is converted into a continuous risk score through a Softmax layer and discretized into four risk levels to obtain a health risk score.

[0167] The final step involves integrating the optimized thermal exposure feature vector with other non-thermal exposure feature vectors. These features are deeply fused using a fully connected neural network, and then a softmax layer is used to convert them into continuous risk scores, which are then discretized into four risk levels: green (≤0.25), yellow (0.25–0.5), orange (0.5–0.75), and red (>0.75). This step enables a quantitative assessment of respiratory disease health risks, providing an intuitive risk indication.

[0168] The entire process combines advanced data analysis techniques (such as self-attention mechanisms and genetic algorithms) and machine learning models (such as fully connected layers and Softmax layers) through a systematic approach, aiming to improve the accuracy of disease monitoring, the timeliness of early warning, and the effectiveness of prevention and control measures.

[0169] In one embodiment, step S124 described above may include steps S1241 to S1247.

[0170] S1241. Using the optimized heat exposure feature vector as input, obtain other non-heat exposure related feature vectors;

[0171] S1242. Design a fully connected neural network where the number of neurons in the input layer matches the total dimension of the feature vector, the hidden layer uses an appropriate activation function, and the number of neurons in the output layer matches the dimension of the respiratory stress feature vector.

[0172] S1243. Input the spliced ​​and optimized thermal exposure feature vector and other non-thermal exposure feature vectors into the fully connected layer network;

[0173] S1244. Forward propagation calculation is performed through a fully connected layer to achieve deep fusion of features and obtain the respiratory stress feature vector.

[0174] S1245. Construct a Softmax layer based on the output of the fully connected layer. The input of the Softmax layer is the respiratory stress feature vector, and the output is a probability distribution representing the probability of different risk levels.

[0175] S1246. Use the Softmax layer to convert the respiratory stress feature vector into a continuous risk score, which is between 0 and 1 and represents the probability of respiratory disease.

[0176] S1247. Based on the magnitude of the continuous risk score, discretize it into four risk levels to obtain the health risk score.

[0177] Specifically, the thermal exposure feature vector optimized in step S123 is used as input. This vector contains new thermal exposure-related feature information after adjusting the switching state of the photovoltaic feeder segments.

[0178] Other non-thermal exposure-related feature vectors are obtained from step S121, including traffic dust index (TDI_PM), cable ozone exposure index (ACE_O3), airway dryness (VPD), and individual respiratory rate (IV_Br).

[0179] Design a fully connected neural network where the number of neurons in the input layer matches the total dimension of the feature vector, the hidden layer uses an appropriate activation function (such as ReLU), and the number of neurons in the output layer matches the dimension of the respiratory stress feature vector.

[0180] The concatenated and optimized thermally exposed feature vectors and other non-thermally exposed feature vectors are input into the fully connected layer network.

[0181] Forward propagation computation is performed through a fully connected layer to achieve deep feature fusion and obtain the respiratory stress feature vector.

[0182] A Softmax layer is constructed based on the output of the fully connected layer. The input of the Softmax layer is the respiratory stress feature vector, and the output is a probability distribution representing the probability of different risk levels.

[0183] The Softmax layer transforms the respiratory stress feature vector into a continuous risk score, which is between 0 and 1 and represents the probability of respiratory disease occurring.

[0184] Based on the magnitude of the continuous risk score, it is discretized into four risk levels: green (≤0.25), yellow (0.25–0.5), orange (0.5–0.75), and red (>0.75), to obtain the final health risk score.

[0185] S130. Trial warning is triggered based on the health risk score to predict peak visit times.

[0186] Obtain the health risk score generated in step S124, which is between 0 and 1 and reflects the likelihood of developing respiratory diseases.

[0187] Based on the range of the health risk score, a corresponding graded warning is triggered.

[0188] A green alert is triggered when the health risk score is ≤0.25, indicating that the current risk of respiratory diseases is low and no special intervention is required, but continuous monitoring is necessary.

[0189] A yellow alert is triggered when the health risk score is between 0.25 and 0.5, alerting relevant departments and the public to pay attention to the risk of respiratory diseases and recommending that outdoor activities be reduced, especially for children, the elderly and people with respiratory diseases.

[0190] An orange alert is triggered when the health risk score is between 0.5 and 0.75, indicating a high risk of respiratory diseases that may have a significant impact on public health. At this time, appropriate intervention measures should be initiated, such as increasing the frequency of watering urban roads to reduce dust, adjusting the operation of photovoltaic systems to reduce the risk of thermal and electrical exposure, and spraying neutralizing agents in polluted areas to reduce ozone concentrations.

[0191] A health risk score >0.75 triggers a red alert, indicating an extremely high risk of respiratory illnesses and potentially a large-scale respiratory outbreak. Immediate and enhanced intervention measures should be implemented, such as restricting motor vehicle use to reduce traffic dust and exhaust emissions, further optimizing energy allocation to reduce thermal and electrical exposure risks, and expanding the coverage of neutralizing agent spraying in contaminated areas.

[0192] Historical respiratory disease emergency room data were collected, and a time-series regression model was established. The current health risk score and the actual number of emergency room visits on that day were used as input variables to predict the number of respiratory emergency room visits the following day. Based on the prediction results, medical resources were allocated in advance to prepare for peak visitation times, such as increasing the number of medical staff on duty and ensuring sufficient stock of relevant medicines and equipment.

[0193] In one embodiment, the above method further includes:

[0194] S140. Based on the warning level, implement intervention measures at the transportation, energy, environmental, and individual levels; specifically including:

[0195] Based on the risk level range of the generated health risk score, the current warning level is determined. If the warning level is orange or red, a signal is sent to the intelligent transportation system to dynamically adjust traffic lights and reduce the vehicle speed limit to 20 km / h. When the warning level reaches orange or red, the energy management system is triggered to transfer the unstable feeder load to an adjacent transformer, reducing air conditioning downtime and lowering the risk of thermal and electrical exposure. If the warning level is orange or red, environmental intervention measures are initiated, controlling drones to automatically spray 5% Na2S2O3 mist droplets to quickly reduce the local ozone concentration and alleviate the respiratory irritation caused by ozone pollution from cable arcing. Based on the warning level, corresponding health tips are pushed to the wearable devices of high-risk groups.

[0196] S150. Collect data after intervention and provide feedback, use incremental learning to optimize the model, evaluate the intervention effect and dynamically adjust the strategy.

[0197] Specifically, a continuously operating monitoring system collects data such as health risk scores and respiratory disease incidence after the intervention is implemented. This data should have the same format and time granularity as the pre-intervention data to facilitate comparative analysis.

[0198] Record information such as the implementation time, intensity, and coverage of various interventions (e.g., traffic speed limits, photovoltaic system adjustments, environmental spraying, individual health tips) in order to assess the relationship between interventions and changes in health risks.

[0199] The collected post-intervention data will be promptly fed back to the disease monitoring and control system to enrich the system's data sample and provide the latest information for the continuous optimization of the model.

[0200] Employ online learning algorithms suitable for model updates, such as online gradient descent and incremental learning of random forests. These algorithms can update the model with new data without retraining the entire model.

[0201] The data after intervention is input into the incremental learning algorithm to update the parameters of the health risk assessment model and the peak visit prediction model, so that the model can adapt to changes in the environment and population, and improve the accuracy and reliability of the prediction.

[0202] The effectiveness of the intervention was quantitatively assessed by comparing the distribution of health risk scores before and after the intervention, changes in the number of respiratory emergency visits, improvements in environmental indicators (such as traffic dust index, heat-electricity exposure index, cable ozone exposure index, airway dryness, etc.) and improvements in individual health indicators (such as respiratory rate).

[0203] Calculate intervention effectiveness evaluation indicators, such as the decrease in health risk scores before and after intervention, the reduction in the number of respiratory emergency visits, and the degree of improvement in environmental indicators, to intuitively reflect the actual effect of the intervention measures.

[0204] Based on the results of the intervention effectiveness evaluation indicators, analyze the advantages and disadvantages of the current intervention strategy, and determine the aspects that need to be adjusted, such as the intervention threshold, the intensity of intervention measures, and the combination of intervention measures.

[0205] Based on the analysis results, intervention strategies are dynamically adjusted. For example, if it is found that traffic speed limits in certain areas are significantly effective in reducing health risks, the speed limit areas can be appropriately expanded or the speed limit can be further reduced; if a certain environmental spraying measure is not effective in reducing ozone concentration, the spraying agent can be changed or the spraying frequency can be adjusted.

[0206] The adjusted intervention strategies are applied to actual disease prevention and control work, and their effects are continuously monitored and evaluated to form a closed-loop feedback and optimization mechanism, thereby continuously improving the effectiveness and precision of intervention measures.

[0207] For example, suppose a complete monitoring system is deployed in the commercial center of a city, including roadside multi-sensors, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-weather stations, and wearable respiratory monitoring devices.

[0208] Data collection:

[0209] Traffic dust data: Multiple roadside sensors collected data on vehicle emergency braking and acceleration events, and the emergency braking rate (HBr) was calculated to be 0.4 times / km. At the same time, the traffic flow speed was analyzed, and combined with the re-suspension model, the traffic dust index (TDI_PM) was obtained to be 0.6.

[0210] Thermal-electric exposure data: The photovoltaic inverter power monitoring module shows an active power (P) of 80kW and a reactive power (Q) of 60kvar. A sudden drop in photovoltaic power was detected, and the depth of power loss (PLD) was calculated to be 20%.

[0211] Cable ozone data: The arc energy (E_arc) detected by the arc sensor at the cable joint is 50J. Based on the calibration curve, the ozone increment (ACE_O3) of the cable is estimated to be 15ppb.

[0212] Micrometeorological data: The distributed micrometeorological station measured the ambient temperature (T) to be 30℃ and the relative humidity (RH) to be 40%, and calculated the airway dryness (VPD) to be 1.2 kPa.

[0213] Individual respiratory data: The average respiratory rate (Br) of individuals collected by the wearable respiratory monitoring device was 22 breaths / min.

[0214] The initial data was standardized using Z-scores to remove outliers, resulting in a five-dimensional index.

[0215] The standardized five-dimensional indicators are integrated into the current respiratory exposure vector E. x (t)=[0.6,0.2,0.15,1.2,0.22].

[0216] Collect five-dimensional indicator data from the same period over the past 7 days, calculate the mean value, and derive the historical exposure vector E. x (h)=[0.4,0.1,0.1,1.0,0.18].

[0217] ΔE x =E x (t)-E x (h)=[0.2,0.1,0.05,0.2,0.04].

[0218] Historical exposure vector E x (h) Input the feature projection layer and calculate the weight values ​​for each dimension, which are [0.3, 0.25, 0.15, 0.15, 0.15]. The dimension related to heat exposure (assumed to be the second dimension) has a weight of 0.25.

[0219] After weighting the difference vector, the feature vector related to heat exposure is [0.2*0.3,0.1*0.25,0.05*0.15,0.2*0.15,0.04*0.15]=[0.06,0.025,0.0075,0.03,0.006].

[0220] Based on the feature vectors related to heat exposure, an optimization objective function is constructed that includes node heat exposure terms and power loss terms, for example: minimize(0.6*node heat exposure term + 0.4*power loss term).

[0221] The population size is set to 100, the maximum number of iterations is 50, the crossover probability is 0.85, and the mutation probability is 0.05.

[0222] 100 initial populations are randomly generated, with each individual representing a possible state (binary code) of a set of photovoltaic feeder segment switches.

[0223] Substitute each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value, and obtain the fitness of the individual.

[0224] Individuals are selected based on their fitness and placed into a breeding pool for crossover and mutation operations to generate a new generation of population.

[0225] After multiple iterations, the maximum number of iterations is reached.

[0226] The individual with the best fitness is selected from the final population, and the corresponding combination of switching states is taken as the target value for thermal exposure optimization. After adjusting the switching states of the feeder segments of the photovoltaic system according to this target value, the optimized thermal exposure feature vector is recalculated as [0.04, 0.02, 0.006, 0.025, 0.005].

[0227] The optimized thermal exposure feature vector is concatenated with other non-thermal exposure-related feature vectors and then input into the designed fully connected neural network.

[0228] After forward propagation calculation, the respiratory stress feature vector is obtained as [0.1, 0.15, 0.08, 0.12, 0.05].

[0229] The feature vector is converted into a continuous risk score of 0.65 by a Softmax layer and then discretized into an orange risk level.

[0230] Triggering a tiered warning: An orange warning is triggered based on a health risk score of 0.65.

[0231] Predicting peak visitation: Based on time series regression model, the number of respiratory emergency room visits is predicted to increase by 30% the following day.

[0232] Intervention measures implemented: Based on the orange alert level, a signal was sent to the intelligent transportation system to dynamically adjust the vehicle speed limit to 20km / h; the energy management system was triggered to transfer the unstable feeder load to an adjacent transformer to reduce air conditioning downtime; drones were controlled to automatically spray 5% Na2S2O3 mist droplets to reduce local ozone concentration; health tips were pushed to wearable devices of high-risk groups, advising them to reduce outdoor activities.

[0233] Data collection after intervention: After the intervention measures are implemented, continue to collect the above types of data for feedback.

[0234] Incremental learning optimization model: The health risk assessment model and the peak visit prediction model are updated using an incremental learning algorithm.

[0235] Assessing the effectiveness of the intervention: Compare indicators such as health risk scores and the number of respiratory emergency visits before and after the intervention to evaluate the effectiveness of the intervention measures. Assume that after the intervention, the health risk score decreases to 0.4, and the increase in the number of respiratory emergency visits decreases to 15%.

[0236] Dynamic adjustment strategy: Based on the assessment results, appropriately expand the traffic speed limit area and further optimize the parameters of the energy management system to continuously improve the intervention effect.

[0237] The aforementioned disease detection and prevention method based on dynamic monitoring data integrates traffic dust, heat-electricity, cable ozone, micro-meteorological, and individual respiratory exposure data collected by roadside multi-sensor systems, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-meteorological stations, and wearable respiratory monitoring devices. After preprocessing and quality control, respiratory stress features are extracted and fused using a self-attention mechanism, genetic algorithm, and health mapping function to generate a health risk score. This score then triggers tiered early warnings and predicts peak medical visits, achieving precise early warning and closed-loop intervention for respiratory disease risks. This process fully leverages the advantages of dynamic infrastructure monitoring data, featuring accurate monitoring, timely early warning, and coordinated prevention and control, effectively compensating for the shortcomings of traditional disease monitoring and prevention methods.

[0238] Figure 3 This is a schematic block diagram of a disease detection and prevention system 300 based on dynamic monitoring data provided in an embodiment of the present invention. Figure 3 As shown, corresponding to the above-described disease detection and control method based on dynamic monitoring data, the present invention also provides a disease detection and control system 300 based on dynamic monitoring data. This disease detection and control system 300 includes a unit for executing the above-described disease detection and control method based on dynamic monitoring data, and the system can be configured in a server. Specifically, please refer to... Figure 3 The disease detection and prevention system 300 based on dynamic monitoring data includes an acquisition unit 301, a generation unit 302, and an early warning unit 303.

[0239] The acquisition unit 301 is used to collect traffic dust, heat-electricity, cable ozone, micro-meteorology, and individual respiratory exposure data through roadside multi-sensors, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-meteorological stations, and wearable respiratory monitoring devices to obtain initial data; the generation unit 302 is used to preprocess and perform quality control on the initial data, extract and fuse respiratory stress features using a self-attention mechanism, genetic algorithm, and health degree mapping function to generate a health risk score; the early warning unit 303 is used to trigger graded early warnings based on the health risk score to predict peak visits to medical institutions.

[0240] Also includes:

[0241] Intervention units are used to implement intervention measures at the transportation, energy, environmental, and individual levels, based on the level of the warning.

[0242] The evaluation unit is used to collect data after the intervention and provide feedback. It uses an incremental learning optimization model to evaluate the intervention effect and dynamically adjust the strategy.

[0243] In one embodiment, the generation unit 302 is used to perform Z-score standardization on the initial data to eliminate dimensional differences and remove outliers based on physical boundaries to form a five-dimensional index; construct a current respiratory exposure vector based on the five-dimensional index, calculate a difference vector by combining historical data, and use a self-attention mechanism to determine the weight of each dimension, highlighting key risk factors and determining heat exposure-related feature vectors; construct an optimization objective function containing node heat exposure terms and power loss terms based on the heat exposure-related feature vectors, use a genetic algorithm to optimize and calculate the heat exposure optimization objective value, and adjust the segmented switching state of the photovoltaic feeder to reduce the risk of heat-electric exposure, so as to obtain an optimized heat exposure feature vector; input the optimized heat exposure feature vector into a fully connected layer for deep fusion to generate a respiratory stress feature vector, and then convert it into a continuous risk score through a Softmax layer, and discretize it into four risk levels to obtain a health risk score.

[0244] In one embodiment, the generation unit 302 is used to integrate the five-dimensional indicators into a vector to represent the respiratory exposure status at the current moment, so as to obtain the current respiratory exposure vector; collect the five-dimensional indicator data of the same period in the past 7 days, calculate its mean, and construct a historical exposure vector to reflect the respiratory exposure level of the same period in history; obtain a difference vector by the difference between the current respiratory exposure vector and the historical exposure vector; input the historical exposure vector into the feature projection layer, extract the severity features of each dimension, obtain the historical feature vector, specifically calculate the difference of each factor in the historical data, obtain the historical factor change vector, perform inner product calculation on the historical factor change vector, obtain the inner product result, and calculate the weight value of each dimension through a fully connected layer and a Softmax layer; according to the weight value calculated by the self-attention mechanism, the difference vector is weighted, wherein the dimensions related to heat exposure are weighted and adjusted to form a feature vector related to heat exposure.

[0245] In one embodiment, the generation unit 302 is used to construct an optimization objective function containing node heat exposure terms and power loss terms based on the heat exposure-related feature vectors; set initial parameters for the genetic algorithm, including population size, maximum number of iterations, crossover probability, and mutation probability; randomly generate an initial population, where each individual represents a possible state of a set of photovoltaic feeder segment switches; substitute each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness; based on the individual's fitness, use a selection strategy to select individuals with high fitness to enter the breeding pool, with individuals with higher fitness having a greater probability of being selected; perform a crossover operation on the individuals in the breeding pool, randomly selecting two individuals as parents, exchanging some of their genes according to the crossover probability to generate new offspring individuals; mutate the genes of the offspring individuals with a certain mutation probability; replace some individuals in the old population with the new individuals generated after crossover and mutation to form a new generation population; and determine whether the termination conditions of the genetic algorithm, such as the maximum number of iterations or the fitness convergence threshold, have been met. If the termination condition is met, proceed to the next step; otherwise, return to the step of substituting each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness, and continue iterating; select the individual with the best fitness from the final population, and the corresponding switching state combination is the heat exposure optimization objective value; according to the switching state combination corresponding to the heat exposure optimization objective value, actually adjust the switching state of the photovoltaic system's feeder segments to reduce the heat-electric exposure risk; under the adjusted photovoltaic system state, recalculate the heat exposure-related feature vector to obtain the optimized heat exposure feature vector.

[0246] In one embodiment, the generation unit 302 is used to take the optimized heat exposure feature vector as input to obtain other non-heat exposure related feature vectors; design a fully connected layer neural network, where the number of neurons in the input layer matches the total dimension of the feature vectors, the hidden layers use an appropriate activation function, and the number of neurons in the output layer matches the dimension of the respiratory stress feature vector; input the concatenated optimized heat exposure feature vector and other non-heat exposure feature vectors into the fully connected layer network; perform forward propagation calculation through the fully connected layer to achieve deep feature fusion and obtain the respiratory stress feature vector; construct a Softmax layer based on the output of the fully connected layer, where the input of the Softmax layer is the respiratory stress feature vector, and the output is a probability distribution representing the probability of different risk levels; use the Softmax layer to convert the respiratory stress feature vector into a continuous risk score, which is between 0 and 1, representing the probability of respiratory disease occurrence; discretize the continuous risk score into four risk levels to obtain a health risk score.

[0247] In one embodiment, the intervention unit is used to determine the current warning level based on the risk level range of the generated health risk score; if the warning level is orange or red, a signal is sent to the intelligent transportation system to dynamically adjust the traffic lights and reduce the vehicle speed limit to 20 km / h; when the warning level reaches orange or red, the energy management system is triggered to transfer the unstable feeder load to an adjacent transformer to reduce air conditioning downtime and reduce the risk of thermal-electrical exposure; if the warning level is orange or red, environmental intervention measures are initiated to control a drone to automatically spray 5% Na2S2O3 mist droplets to quickly reduce the local ozone concentration and alleviate the respiratory tract irritation caused by ozone pollution from cable arcing; according to the warning level, corresponding health tips are pushed to the wearable devices of high-risk groups.

[0248] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned disease detection and prevention system 300 based on dynamic monitoring data and its various units can be referred to the corresponding descriptions in the aforementioned method embodiments. For the sake of convenience and brevity, these details will not be repeated here.

[0249] The aforementioned disease detection and prevention system 300 based on dynamic monitoring data can be implemented as a computer program, which can, for example... Figure 3 It runs on the computer device shown.

[0250] Please see Figure 3 , Figure 3 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0251] See Figure 3 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0252] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a disease detection and prevention method based on dynamic monitoring data.

[0253] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0254] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a disease detection and prevention method based on dynamic monitoring data.

[0255] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0256] The processor 502 is used to run a computer program 5032 stored in the memory to implement all the steps of the disease detection and prevention method based on dynamic monitoring data.

[0257] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0258] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0259] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all the steps of the disease detection and prevention method based on dynamic monitoring data.

[0260] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0261] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0262] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0263] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0264] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0265] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A disease detection and prevention method based on dynamic monitoring data, characterized in that, include: Initial data were obtained by collecting traffic dust, heat-electricity, cable ozone, micro-meteorological and individual respiratory exposure data through roadside multi-sensors, photovoltaic inverter power monitoring modules, cable joint arc sensors, distributed micro-meteorological stations and wearable respiratory monitoring devices. The initial data is preprocessed and quality controlled. Respiratory stress features are extracted and fused using a self-attention mechanism, genetic algorithm, and health degree mapping function to generate a health risk score. Based on the aforementioned health risk score, a tiered early warning system is triggered to predict peak periods for medical visits.

2. The disease detection and prevention method based on dynamic monitoring data according to claim 1, characterized in that, The process of triggering a tiered early warning based on the health risk score and predicting peak medical visits also includes: Based on the level of the warning, implement intervention measures at the transportation, energy, environmental, and individual levels; Data collected after intervention is fed back, and incremental learning is used to optimize the model, evaluate the intervention effect, and dynamically adjust the strategy.

3. The disease detection and prevention method based on dynamic monitoring data according to claim 1, characterized in that, The initial data undergoes preprocessing and quality control, and respiratory stress features are extracted and fused using a self-attention mechanism, genetic algorithm, and health score mapping function to generate a health risk score, including: The initial data is Z-score standardized to eliminate dimensional differences, and outliers are removed based on physical boundaries to form a five-dimensional index. Based on the five-dimensional indicators, a current respiratory exposure vector is constructed, and a difference vector is calculated by combining historical data. The weight of each dimension is determined by using a self-attention mechanism to highlight key risk factors and determine the feature vector related to heat exposure. Based on the heat exposure-related feature vectors, an optimization objective function containing node heat exposure terms and power loss terms is constructed. A genetic algorithm is used to optimize and calculate the heat exposure target value, and the switching state of the photovoltaic feeder segments is adjusted to reduce the risk of thermal-electric exposure, so as to obtain the optimized heat exposure feature vector. The optimized heat exposure feature vector is input into a fully connected layer for deep fusion to generate a respiratory stress feature vector. Then, it is converted into a continuous risk score through a Softmax layer and discretized into four risk levels to obtain a health risk score.

4. The disease detection and prevention method based on dynamic monitoring data according to claim 3, characterized in that, The process involves constructing a current respiratory exposure vector based on the five-dimensional indicators, calculating a difference vector using historical data, determining the weights of each dimension using a self-attention mechanism, highlighting key risk factors, and identifying heat exposure-related feature vectors, including: The five-dimensional indicators are integrated into a vector to represent the respiratory exposure status at the current moment, so as to obtain the current respiratory exposure vector; Collect five-dimensional index data from the same period of the past 7 days, calculate their mean, and construct a historical exposure vector to reflect the level of respiratory exposure in the same period of history. The difference vector is obtained by comparing the current respiratory exposure vector with the historical exposure vector; The historical exposure vector is input into the feature projection layer to extract the severity features of each dimension, thus obtaining the historical feature vector. Specifically, the differences of each factor in the historical data are calculated to obtain the historical factor change vector. The historical factor change vector is then used to calculate the inner product to obtain the inner product result. Finally, the weight values ​​of each dimension are calculated through the fully connected layer and the Softmax layer. The difference vector is weighted based on the weight values ​​calculated by the self-attention mechanism. The dimensions related to heat exposure are then weighted to form a feature vector related to heat exposure.

5. The disease detection and prevention method based on dynamic monitoring data according to claim 3, characterized in that, The optimization objective function, which includes node heat exposure terms and power loss terms, is constructed based on the heat exposure-related feature vector. A genetic algorithm is used to calculate the optimized heat exposure objective value, and the switching states of the photovoltaic feeder segments are adjusted to reduce the risk of thermal-electrical exposure, resulting in an optimized heat exposure feature vector. This includes: Based on the heat exposure-related feature vectors, an optimization objective function containing node heat exposure terms and power loss terms is constructed. Set the initial parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, and mutation probability; An initial population is randomly generated, with each individual representing a possible state of a set of photovoltaic feeder segment switches; Substitute each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness. Based on the fitness of individuals, a selection strategy is used to select individuals with high fitness to enter the breeding pool. Individuals with higher fitness have a greater probability of being selected. Crossover is performed on individuals in the breeding pool. Two individuals are randomly selected as parents, and some of their genes are exchanged according to the crossover probability to generate new offspring individuals. The genes of offspring individuals are mutated with a certain probability of mutation. The new individuals produced through crossover and mutation replace some individuals in the old population to form a new generation of population; Determine if the termination condition of the genetic algorithm has been met, such as the maximum number of iterations or the fitness convergence threshold. If the termination condition is met, proceed to the next step; otherwise, return to the step of substituting each individual into the optimization objective function to calculate its corresponding heat exposure risk and power loss value to obtain the individual's fitness, and continue iterating. The individual with the best fitness is selected from the final population, and the corresponding combination of on / off states is the target value for heat exposure optimization. Based on the switching state combination corresponding to the thermal exposure optimization target value, the actual switching state of the feeder segments of the photovoltaic system is adjusted to reduce the thermal-electric exposure risk. Under the adjusted photovoltaic system conditions, the heat exposure-related feature vectors are recalculated to obtain the optimized heat exposure feature vectors.

6. The disease detection and prevention method based on dynamic monitoring data according to claim 4, characterized in that, The optimized heat exposure feature vector is input into a fully connected layer for deep fusion to generate a respiratory stress feature vector. This vector is then converted into a continuous risk score via a Softmax layer and discretized into four risk levels to obtain a health risk score, including: The optimized thermal exposure feature vector is used as input to obtain other non-thermal exposure related feature vectors. Design a fully connected neural network where the number of neurons in the input layer matches the total dimension of the feature vector, the hidden layer uses an appropriate activation function, and the number of neurons in the output layer matches the dimension of the respiratory stress feature vector. The spliced ​​and optimized thermal exposure feature vector and other non-thermal exposure feature vectors are input into the fully connected layer network; Forward propagation computation is performed through a fully connected layer to achieve deep feature fusion and obtain the respiratory stress feature vector; A Softmax layer is constructed based on the output of the fully connected layer. The input of the Softmax layer is the respiratory stress feature vector, and the output is a probability distribution representing the probability of different risk levels. The respiratory stress feature vector is transformed into a continuous risk score between 0 and 1 using a Softmax layer, which represents the probability of respiratory disease occurrence. Based on the magnitude of the continuous risk score, it is discretized into four risk levels to obtain the health risk score.

7. The disease detection and prevention method based on dynamic monitoring data according to claim 2, characterized in that, The aforementioned measures, based on the warning level, include interventions at the transportation, energy, environmental, and individual levels, including: The current warning level is determined based on the risk level range of the generated health risk score; If the warning level is orange or red, a signal is sent to the intelligent transportation system to dynamically adjust the traffic lights and reduce the speed limit to 20km / h. When the warning level reaches orange or red, the energy management system is triggered to transfer the unstable feeder load to the adjacent transformer, reducing air conditioning downtime and lowering the risk of thermal and electrical exposure. If the warning level is orange or red, environmental intervention measures will be initiated, and drones will be controlled to automatically spray 5% Na2S2O3 mist droplets to quickly reduce the local ozone concentration and reduce the respiratory irritation caused by ozone pollution from cable arcing. Based on the warning level, corresponding health alerts will be pushed to the wearable devices of high-risk groups.

8. A disease detection and prevention system based on dynamic monitoring data, characterized in that, include: The acquisition unit is used to collect traffic dust, heat-electricity, cable ozone, micro-meteorology and individual respiratory exposure data through roadside multi-sensors, photovoltaic inverter power monitoring module, cable joint arc sensor, distributed micro-weather station and wearable respiratory monitoring device to obtain initial data; The generation unit is used for preprocessing and quality control of the initial data, and extracts and fuses respiratory stress features using a self-attention mechanism, a genetic algorithm and a health degree mapping function to generate a health risk score; The early warning unit is used to trigger tiered early warnings based on the health risk score and predict peak periods for medical visits.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images