Fan rotating speed adjusting method and system based on dynamic pollution load

Abstract

The invention relates to the technical field of fan control, and discloses a fan rotating speed adjusting method and system based on a dynamic pollution load. The method comprises the following steps: collecting a target PM2.5 concentration, a target VOC concentration and a target microorganism concentration of each monitoring point, and calculating a comprehensive pollution load index of each monitoring point; traversing boundary critical points in all the monitoring points based on the comprehensive pollution load index and generating a boundary identification vector; the real-time pressure drop of the air inlet side and the air outlet side of the filter element in each fan is collected, and the pressure difference compensation gain is calculated; calculating a target reference rotating speed for all monitoring points in the responsible sector of each fan; and calculating the average pollution load index of the responsible sector of each fan, executing fuzzy reasoning to obtain a rotating speed increment, and generating a PWM duty ratio according to the target reference rotating speed and the rotating speed increment to drive each fan. The technical problems of space pollution response lag and energy consumption imbalance are solved, and the performance of the air purification system is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fan control, and particularly relates to a fan rotating speed regulation method and system based on dynamic pollution load. BACKGROUND

[0002] The existing air purification system generally adopts a unified global rotating speed control strategy, that is, all fans are synchronously regulated according to the global average pollution concentration. This method ignores the difference in pollution distribution in the space, resulting in that the local high-pollution area cannot be purified in time due to the lag of fan response, and the fan in the low-pollution area causes energy waste due to over operation.

[0003] The traditional control method only makes speed regulation decision based on a single pollution concentration or a simple arithmetic mean value, without considering the difference in health hazard degree of different pollutants and without introducing the influence weight of spatial distance on the purification effect, resulting in inaccurate pollution load evaluation. In addition, the existing technology ignores the problem of actual ventilation decay caused by the increase in resistance during the saturation process of the filter element. When the filter element is gradually blocked, the purification efficiency under the fixed rotating speed decreases obviously. SUMMARY

[0004] The main purpose of the present application is to provide a fan rotating speed regulation method and system based on dynamic pollution load. The present application solves the technical problems of space pollution response lag and energy imbalance, and improves the performance of the air purification system.

[0005] To achieve the above-mentioned purpose, the present application provides a fan rotating speed regulation method based on dynamic pollution load, comprising the following steps: Collecting target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculating comprehensive pollution load index of each monitoring point; Based on the comprehensive pollution load index, traversing all boundary critical points in the monitoring points and generating a boundary identification vector; Collecting real-time pressure drop of the inlet side and outlet side of the filter element in each fan, and calculating differential pressure compensation gain according to the real-time pressure drop; For all monitoring points in the responsibility sector of each fan, calculating target reference rotating speed according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain; Calculating the average pollution load index of the responsibility sector of each fan and performing fuzzy reasoning to obtain rotating speed increment, and generating PWM duty cycle according to the target reference rotating speed and the rotating speed increment to drive each fan.

[0006] Optionally, in the first implementation manner of the first aspect of the present application, the collecting target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculating comprehensive pollution load index of each monitoring point, comprises: Collecting original PM2.5 concentration, original VOC concentration and original microorganism concentration of each monitoring point; Summing the original PM2.5 concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain a target PM2.5 concentration; Summing the original VOC concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain a target VOC concentration; Summing the original microorganism concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain a target microorganism concentration; According to the target PM2.5 concentration, the target VOC concentration and the target microorganism concentration, calculating the comprehensive pollution load index of each monitoring point.

[0007] Optionally, in the second implementation manner of the first aspect of the present application, the comprehensive pollution load index of each monitoring point is calculated according to the target PM2.5 concentration, the target VOC concentration and the target microorganism concentration, including: Calculating the distance ratio of the distance from each monitoring point to the nearest fan to the maximum monitoring radius, multiplying the distance ratio by a first target value to obtain a spatial distance attenuation factor as an index term; Dividing the target PM2.5 concentration by a PM2.5 standard threshold value, multiplying the result by a second target value and the spatial distance attenuation factor to obtain a PM2.5 weighted component, dividing the target VOC concentration by a VOC standard threshold value, multiplying the result by a third target value and the spatial distance attenuation factor to obtain a VOC weighted component, and dividing the target microorganism concentration by a microorganism standard threshold value, multiplying the result by a fourth target value and the spatial distance attenuation factor to obtain a microorganism weighted component; Summing the PM2.5 weighted component, the VOC weighted component and the microorganism weighted component to obtain the comprehensive pollution load index of each monitoring point.

[0008] Optionally, in the third implementation manner of the first aspect of the present application, the boundary critical points in all monitoring points are traversed based on the comprehensive pollution load index, and a boundary identification vector is generated, including: Traversing all adjacent monitoring point pairs with a distance less than a preset distance threshold value, calculating the absolute value of the difference of the comprehensive pollution load index of the adjacent monitoring point pairs; When the absolute value of the difference is greater than a first boundary threshold value, the comprehensive pollution load index of the high pollution point is greater than a second boundary threshold value, and the comprehensive pollution load index of the low pollution point is less than the second boundary threshold value, marking the corresponding adjacent monitoring point pair as a boundary critical point; For the marked boundary critical point pair, dividing the difference of the comprehensive pollution load index by the distance between the two points to obtain a boundary gradient strength; The boundary gradient intensity at the current time is reduced by a first preset multiple of the boundary gradient intensity at the previous time, and then reduced by a second preset multiple of the boundary gradient intensity at the time two steps before, to obtain a boundary drift prediction value, and a boundary identification vector is generated based on the boundary drift prediction value and a boundary critical point label.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present application, the real-time pressure drops of the inlet side and the outlet side of the filter element in each fan are collected, and a differential pressure compensation gain is calculated according to the real-time pressure drops, comprising: The real-time pressure drops of the inlet side and the outlet side of the filter element in each fan are collected, and a resistance attenuation ratio is obtained by dividing the real-time pressure drops by the initial pressure drop calibrated when the filter element is used for the first time; An intermediate compensation value is obtained by multiplying the resistance attenuation ratio by a first preset coefficient after the resistance attenuation ratio is reduced by 1, and a nonlinear adjustment factor is calculated by inputting the resistance attenuation ratio multiplied by a second preset coefficient into a hyperbolic tangent function; The differential pressure compensation gain is obtained by multiplying the intermediate compensation value by the nonlinear adjustment factor and then adding 1.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present application, the target reference speed is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain for all monitoring points in the responsibility sector of each fan, comprising: The task priority weight is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain for all monitoring points in the responsibility sector of each fan; The total weight is obtained by summing up the task priority weights of each fan, and the relative allocation coefficient is obtained by normalizing the task priority weights of each fan by dividing the total weight; The target reference speed is obtained by adding the product of the lowest silent speed and the relative allocation coefficient and the speed adjustment range to the lowest silent speed.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present application, the task priority weight is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain for all monitoring points in the responsibility sector of each fan, comprising: All monitoring points in the responsibility sector of each fan are traversed, and it is judged according to the boundary identification vector whether each monitoring point in the responsibility sector is a boundary critical point; When the monitoring point is a boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1 plus twice the boundary indication value, and when the monitoring point is a non-boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1. The task priority weight is obtained by multiplying the comprehensive pollution load index of each monitoring point by the corresponding boundary weight factor, multiplying the differential pressure compensation gain of the fan, summing all monitoring points in the responsibility sector, and dividing by the total number of monitoring points in the responsibility sector.

[0012] Optionally, in the seventh implementation form of the first aspect of the present application, the calculation of the average pollution load index of the responsibility sector of each fan and the fuzzy inference to obtain the speed increment, and the generation of the PWM duty cycle based on the target reference speed and the speed increment to drive each fan, comprises: The average pollution load index of all monitoring points in the responsibility sector of each fan is obtained by averaging the comprehensive pollution load index of each monitoring point, the deviation is obtained by subtracting the target pollution load index from the average pollution load index, and the deviation change rate is obtained by subtracting the deviation at the previous moment from the deviation at the current moment. The fuzzy inference is performed based on the deviation and the deviation change rate to obtain the speed increment. The target reference speed is superimposed with the speed increment after limiting to obtain the expected speed, the PWM duty cycle is obtained by dividing the expected speed by the rated maximum speed and multiplying by 100%, and each fan is driven according to the PWM duty cycle.

[0013] Optionally, in the eighth implementation form of the first aspect of the present application, the fuzzy inference based on the deviation and the deviation change rate to obtain the speed increment, comprises: The deviation is quantized into corresponding M first fuzzy levels according to a preset interval threshold, and the deviation change rate is quantized into corresponding M second fuzzy levels. The deviation is input into a triangular membership function to calculate a first membership value corresponding to the first fuzzy level, and the deviation change rate is input into a triangular membership function to calculate a second membership value corresponding to the second fuzzy level. The first membership value and the second membership value are input into a fuzzy inference rule base, the activation strength of each rule is calculated by the max-min inference method to obtain an output fuzzy set, and the speed increment is calculated based on the output fuzzy set.

[0014] The present application also provides a fan speed regulation system based on dynamic pollution load, comprising: A collection module is configured to collect target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculate the comprehensive pollution load index of each monitoring point. A traversal module is configured to traverse all boundary critical points in all monitoring points based on the comprehensive pollution load index and generate a boundary identification vector. A compensation module is configured to collect real-time pressure drops of the filter core at the air inlet side and the air outlet side of each fan, and calculate a differential pressure compensation gain according to the real-time pressure drops; A calculation module is configured to calculate a target reference speed for all monitoring points in the responsibility sector of each fan according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain; A driving module is configured to calculate an average pollution load index of the responsibility sector of each fan, perform fuzzy reasoning to obtain a speed increment, and generate a PWM duty cycle according to the target reference speed and the speed increment to drive each fan.

[0015] In summary, the technical scheme provided by the present application realizes precise and efficient control of the air purification system through the task self-adaptive allocation type fan speed fuzzy closed-loop regulation method based on dynamic pollution load weighted fusion and differential pressure compensation. By introducing a spatial distance attenuation factor and a weighted fusion calculation of differentiated health weights of multiple pollutants, the PM2.5, VOC and microbial load are converted into a comprehensive pollution load index, the spatialization and risk-based precise quantification of the pollution state are realized, and the problem of inaccurate pollution evaluation in the traditional method is solved. Through the boundary critical point identification and boundary drift prediction algorithm, the precise positioning of the local high-pollution area and the pollution diffusion path is realized, which provides a spatial target basis for differentiated speed regulation and solves the response lag problem caused by unified control. Through real-time pressure difference monitoring and nonlinear compensation gain function, the speed demand under the saturated state of the filter core is dynamically corrected to ensure that the purification effect does not decrease due to the increase of filter resistance, and the defect of ignoring the filter core state in the traditional method is solved. Through task priority weight calculation and relative allocation coefficient normalization processing, the spatial self-adaptive optimization configuration of the multi-fan speed resource is realized, so that the high-pollution area and the boundary diffusion area obtain more purification resources, and the low-pollution stable area reduces energy consumption. Combined with the precise regulation of the double-input single-output fuzzy closed-loop controller, the present application solves the technical problems of spatial pollution response lag and energy imbalance, and improves the performance of the air purification system. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a fan speed regulation method step schematic diagram based on dynamic pollution load in an embodiment of the present application; Figure 2 is a fan speed regulation system structure block diagram based on dynamic pollution load in an embodiment of the present application.

[0017] The implementation of the present application, functional characteristics and advantages will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0018] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0019] With reference to Figure 1 The embodiment provides a fan rotating speed adjusting method based on dynamic pollution load, which comprises the following steps: S1, collecting target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculating comprehensive pollution load index of each monitoring point; Wherein, the pollution perception subsystem is constructed, through the deployment of the composite environmental monitoring node in the air purification area, the PM2.5 particle concentration, VOC gas concentration and microorganism concentration original value of each monitoring point are collected in real time under the fixed sampling period, and the original concentration data of the current time t is recorded, at the same time, the same type of pollution data of each monitoring point at t-1 and t-2 two historical time is saved, and the three time sliding sampling sequence of each type of pollution is constituted. The original PM2.5 concentration of the current time and the previous two times is summed and the arithmetic mean is taken, the target PM2.5 concentration value after smoothing processing is obtained, the same processing logic is adopted for VOC concentration and microorganism concentration, and the target VOC concentration and target microorganism concentration are obtained respectively, which effectively suppresses the interference caused by instantaneous abnormal value or mutation value to the judgment. Based on the target concentration data of three types of pollutants, the three types of pollutants are subjected to differential standard normalization processing and health hazard weight correction, wherein each type of pollutant corresponds to the safety limit value in the national or industry standard as the standardization reference value, for example, PM2.5 concentration is 35 μg / m 3 , VOC is 0.6 mg / m 3 , and microorganism concentration is 500 CFU / m 3 , the reference value is subjected to ratio normalization, and then combined with the preset health risk weight coefficient (for example, PM2.5 weight is 0.4, VOC is 0.35, and microorganism is 0.25), the three types of standardized concentration values are subjected to weighted fusion operation, and the comprehensive pollution load index of each monitoring point at the current time is obtained.

[0020] S2, based on the comprehensive pollution load index, traversing all boundary critical points in the monitoring points and generating a boundary identification vector; Specifically, among all the monitoring points, the spatial Euclidean distances between any two points are sequentially traversed, and the point pairs with a distance less than a set threshold (e.g., 4.5 meters) are selected as "adjacent monitoring point pairs" for processing. Between each adjacent monitoring point pair, the absolute value of the difference between the corresponding comprehensive pollution load index values is calculated as a local gradient index of the pollution concentration. A boundary criticality judgment is performed on the local gradient value, and a double judgment mechanism is adopted, that is, only when the absolute value of the difference between the comprehensive pollution load index values exceeds a first boundary threshold (e.g., 0.4), and the point with a larger comprehensive pollution load index value in the point pair exceeds a second boundary threshold (e.g., 0.7) and the point with a smaller comprehensive pollution load index value is below the second boundary threshold, the current point pair is marked as a boundary critical point. For the monitoring point pairs marked as boundary critical, the boundary gradient strength is calculated, which is the difference between the comprehensive pollution load index values of the current point pair divided by the actual spatial distance between the two points, resulting in a pollution load jump strength per unit length, with a dimensionless pollution difference per meter. The larger the value, the more intense the spatial variation of the pollution concentration. A time series model based on inertia and damping characteristics is introduced to record the boundary gradient strength values at the current time, the previous time, and the time two times ago, respectively. The boundary drift prediction value is constructed by subtracting 0.7 times the previous time value from the current gradient and then subtracting 0.3 times the time two times ago value. When the boundary drift prediction value is greater than 0.2, it indicates that the pollution is diffusing to the low pollution area, and the fan speed of the boundary sector can be adjusted to suppress it; when the boundary drift prediction value is less than -0.2, it indicates that the pollution boundary is shrinking, and the fan power can be adjusted to save energy. Based on the above analysis results, a boundary identification vector is generated, where for the monitoring points in the boundary critical area, the corresponding boundary identification vector is assigned a value of 1, otherwise 0. The calculation of the boundary drift prediction value is based on a second-order linear prediction model. For a time series signal x(t), its prediction value at t+Δt can be linearly extrapolated from the current value and historical derivative terms, expressed as x(t+Δt)≈x(t)+x'(t)·Δt+x''(t)·Δt 2 / 2, the continuous time model is discretized and the derivative term is calculated using the backward difference method, and a prediction model based on three historical data at t+Δt is obtained x pred (t+Δt)=a0·x(t)+a1·x(t-Δt)+a2·x(t-2Δt), where the coefficients satisfy a0+a1+a2=1 to ensure unbiased estimation. For the drift prediction of the boundary gradient strength G(t), the present invention adopts a modified prediction-correction structure, and the prediction value is defined as the deviation of the current gradient from the historical trend, that is, ΔG pred=G(t)-[β1·G(t-Δt)+β2·G(t-2Δt)], the physical meaning of the formula is that if the boundary gradient is continued according to historical inertia, it should satisfy G(t)≈β1·G(t-Δt)+β2·G(t-2Δt), and the difference between the actual value and the expected value according to inertia is the drift. The determination of the first preset multiple β1 is based on the optimal weight distribution principle in the exponential smoothing theory. For a pollution monitoring system with a sampling period Δt=30 seconds, the characteristic time constant τ of the boundary diffusion is about 70-90 seconds, which is obtained by experimentally measuring the time for the pollution plume to reach the stable boundary from the initial release. According to the discretization formula of the first-order inertia link β1=exp(-Δt / τ)≈exp(-30 / 80)≈0.69, β1=0.7 is taken into account for numerical stability and calculation simplicity. The determination of the second preset multiple β2 needs to ensure the stability of the prediction model. For a second-order AR model z 2 -β1·z-β2=0, the stability conditions are |β2|<1, β1+β2<1 and β1-β2>-1. In the case of β1=0.7, the value range of β2 is (-0.3, 0.3). Through the root mean square error (RMSE) analysis of 20 groups of actual scene data, the prediction bias is smallest (RMSE=0.08) when β2=0.3. Therefore, the second preset multiple β2=0.3 is taken, which satisfies β1+β2=1.0, indicating that the sum of the historical term weights is equal to the current term weight, forming a balanced trend tracking characteristic. The boundary identification vector B is an N×1 vector, and N is the total number of monitoring points. The assignment logic of the element B[i] is as follows: for the ith monitoring point, the default B=0 indicates a non-boundary point. If the point is marked as a boundary critical point, further judgment is made. When ΔG pred is greater than 0.2 and G(t) is greater than the boundary gradient intensity judgment threshold G threshold , B[i]=1.0 is assigned, indicating a boundary diffusion indication. When ΔG pred is less than -0.2 and G(t) is less than G threshold , B[i]=-1.0 is assigned, indicating a boundary contraction indication. In other cases, B[i]=0.5 is assigned, indicating a boundary stable indication. G threshold =0.15 is the judgment threshold of the boundary gradient intensity, which is calculated according to the limit value of the pollution concentration gradient. The requirement is that the pollution load index difference within a distance of 1 meter should not exceed 0.15. The thresholds 0.2 and -0.2 are set based on the experimental statistical values of the turbulent diffusion rate in the Reynolds number Re=1500-2000 range in pollution diffusion dynamics. When the boundary gradient change rate exceeds the threshold, it indicates that there is significant convective transport action that needs to be regulated and responded.

[0021] S3, collecting the real-time pressure drops of the inlet and outlet sides of the filter core in each fan, and calculating the differential pressure compensation gain according to the real-time pressure drops; It should be noted that differential pressure sensors are arranged at both ends of the filter core assembly of each fan to continuously collect real-time pressure drop values between the inlet and outlet sides of the filter core, and the real-time pressure drop values are compared with the initial calibrated differential pressure recorded when the filter core is first used at a rated speed for 60 seconds. The current resistance decay ratio is calculated by ratio operation, which reflects the growth multiple of the current resistance of the filter core relative to the initial clean state. Subtract 1 from the resistance decay ratio to obtain the resistance growth amplitude relative to the initial state, and multiply the resistance growth amplitude by a first preset coefficient (such as 1.2) to obtain a linear intermediate compensation value to represent the speed compensation reference required due to the additional wind resistance caused by the aging of the filter core. In order to avoid the problem of compensation overshoot or instability caused by linear gain, a nonlinear adjustment module with resistance decay ratio as input is introduced. The nonlinear adjustment module inputs the resistance decay ratio multiplied by a second preset coefficient (such as 0.5) as an independent variable into a hyperbolic tangent function to obtain a smooth convergent nonlinear adjustment factor. The nonlinear adjustment factor slowly increases when the filter core resistance gradually rises, and tends to a saturation value when the resistance limit is close, forming an effective compensation growth inhibition effect. Multiply the linear intermediate compensation value by the nonlinear adjustment factor, and add 1 to the product to form a differential pressure compensation gain value, which represents the relative amplification coefficient that should be applied to the current fan speed reference value.

[0022] S4, for all monitoring points in the responsibility sector of each fan, calculating the target reference speed according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain; Specifically, according to the specific position of each fan in the spatial layout, each monitoring point is assigned to the responsibility sector of the fan closest to it according to the principle of minimum Euclidean distance, forming a sector attribution relationship. When calculating the task intensity of the fan, all monitoring points responsible for the fan are traversed, and the comprehensive pollution load index and boundary identification state of each point are read in turn. By combining the differential pressure compensation gain of the current fan, the pollution intensity, the boundary priority and the filter resistance are all included in the evaluation model to calculate the task priority weight representing the current purification task amount of each fan. After completing the task priority weight calculation of all fans, the weight values of all fans are summed to obtain the total weight, and the task weight of each fan is divided by the total weight to obtain the relative allocation coefficient of each fan, which reflects the proportion of purification resources that each fan should bear. Reasonable upper and lower limits are set in the control logic to avoid extreme allocation. The difference between the rated maximum speed and the minimum silent speed is taken as the effective range of speed regulation, and the minimum silent speed is taken as the starting point. The relative allocation coefficient is multiplied by the speed regulation range and added to determine the target reference speed of each fan.

[0023] S5, calculate the average pollution load index of the responsibility sector of each fan and perform fuzzy reasoning to obtain the speed increment, and generate the PWM duty cycle according to the target reference speed and the speed increment to drive each fan.

[0024] Among them, based on the responsibility sector division result, the pollution state of all monitoring points covered by a single fan is summarized in each control cycle, the comprehensive pollution load index of each point is averaged to obtain the average pollution load index representing the overall pollution degree of each sector, reflecting the overall pollution level of the current area. Set the target pollution load index as the control reference value, and calculate the difference between the current average pollution load index and the target value to obtain the deviation, which measures the deviation of the current actual state from the ideal state. At the same time, introduce a dynamic factor from the time dimension, calculate the deviation change rate by subtracting the deviation at the previous moment from the deviation at the current moment, to reflect whether the pollution state is deteriorating, tending to be stable or improving. Based on the deviation and the deviation change rate, fuzzy reasoning is performed to construct fuzzy language levels covering different deviation levels and change trends, and corresponding fuzzy control rules are set to respond to pollution deterioration, improvement, and stability in different situations. Through the parallel activation of membership functions and reasoning rules, the contribution results of each rule are integrated and de-fuzzification calculation is completed to output the speed increment for speed regulation. To prevent excessive speed fluctuation or abnormal equipment load, the speed increment is subjected to amplitude limiting processing to make it smooth within the allowed maximum adjustment range. Superimpose the speed increment after limiting the amplitude on the target reference speed to form the expected speed, which reflects the target speed of the fan that should be actually run under the driving of the pollution state. According to the proportional relationship between the expected speed and the rated maximum speed, the expected speed is converted into a PWM duty cycle signal in percentage form, and the PWM duty cycle signal is output to the speed regulation module of the fan motor to complete the real-time adjustment of the actual running speed of the fan.

[0025] In one example, the target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point are collected, and the comprehensive pollution load index of each monitoring point is calculated, including: Collect the original PM2.5 concentration, original VOC concentration and original microorganism concentration of each monitoring point; Sum the original PM2.5 concentrations of each monitoring point at the current time and the previous two times and divide by 3 to obtain the target PM2.5 concentration; Sum the original VOC concentrations of each monitoring point at the current time and the previous two times and divide by 3 to obtain the target VOC concentration; Sum the original microorganism concentrations of each monitoring point at the current time and the previous two times and divide by 3 to obtain the target microorganism concentration; According to the target PM2.5 concentration, the target VOC concentration, and the target microorganism concentration, a comprehensive pollution load index of each monitoring point is calculated.

[0026] In this example, multiple-point distributed monitoring nodes are arranged in a regular grid manner in the air purification area, each node synchronously collects PM2.5 particle concentration, VOC gas concentration, and air microorganism load of three types of pollution data, and forms a multi-channel original pollutant concentration sequence. In order to eliminate the sampling noise of the sensor itself and the transient peak interference caused by air disturbance, while maintaining sufficient response speed and enhancing data smoothness, a sliding window time series average processing operation is performed on each type of pollutant, and the original pollutant concentration values of the current time and the previous two sampling periods are taken to obtain the arithmetic average, respectively, to obtain the target PM2.5 concentration, the target VOC concentration, and the target microorganism concentration. Based on the target concentration values of the three pollutants, the pollution load modeling stage is entered, the health hazard degree and the treatment difficulty of the pollutants are distinguished, the normalized reference value is set for each type of pollutant according to the environmental standard or industry specification, so that the concentrations of various types are comparable, and the relative importance of the three types of pollutants in the air purification task is reflected through setting a differential weight factor. Among them, the particulate matter PM2.5 is given the highest weight because it can deeply enter the respiratory tract and has obvious health effects, VOC is given a medium weight because it has high volatility and sensitization characteristics, and the microorganism load is given a relatively low weight because of its settling speed and transmission path. Through the above normalization processing and weighting mechanism, the comprehensive pollution load index of each monitoring point is calculated.

[0027] In one example, according to the target PM2.5 concentration, the target VOC concentration, and the target microorganism concentration, a comprehensive pollution load index of each monitoring point is calculated, including: The distance ratio of the distance from each monitoring point to the nearest fan to the maximum monitoring radius is calculated, and the distance ratio is multiplied by the first target value to obtain a spatial distance attenuation factor as an index term; The target PM2.5 concentration is divided by the PM2.5 standard threshold value, multiplied by the second target value and the spatial distance attenuation factor to obtain a PM2.5 weighted component, the target VOC concentration is divided by the VOC standard threshold value, multiplied by the third target value and the spatial distance attenuation factor to obtain a VOC weighted component, and the target microorganism concentration is divided by the microorganism standard threshold value, multiplied by the fourth target value and the spatial distance attenuation factor to obtain a microorganism weighted component; The PM2.5 weighted component, the VOC weighted component, and the microorganism weighted component are summed to obtain the comprehensive pollution load index of each monitoring point.

[0028] In this example, after the deployment of the pollution monitoring network is completed, the position of each monitoring point in space relative to all the fans is collected, the Euclidean distance from each monitoring point to each fan is calculated, and the minimum value is selected as the actual distance from the monitoring point to the nearest fan. The distance from the spatial center point to the farthest point in all monitoring points is counted as the global maximum monitoring radius, and the distance between each monitoring point and the nearest fan is divided by the maximum monitoring radius to obtain a normalized distance ratio to reflect the relative position of the monitoring point in the spatial range. Multiply the normalized distance ratio by the target coefficient as the input item of the exponential function, and generate a spatial distance attenuation factor using the principle of exponential decay. The value of the spatial distance attenuation factor tends to 1 near the fan, and decays exponentially away from the fan. Its physical meaning is that the closer the monitoring point to the fan, the higher the weight of the pollution data of the monitoring point to the fan control task. Process the concentration data of the three types of target pollutants respectively. The target PM2.5 concentration is normalized according to the PM2.5 daily average standard of the national ambient air quality standard, the VOC concentration is normalized according to the indoor air quality standard of civil buildings, and the microbial concentration is normalized according to the medical clean room standard. Multiply the normalized values by the target coefficient representing the pollution hazard weight and the spatial distance attenuation factor respectively to obtain the PM2.5 weighted component, the VOC weighted component, and the microbial weighted component. Sum the PM2.5 weighted component, the VOC weighted component, and the microbial weighted component to obtain the comprehensive pollution load index of each monitoring point, which reflects the pressure degree of the current purification task.

[0029] In one example, based on the comprehensive pollution load index, the boundary critical points in all monitoring points are traversed and a boundary identification vector is generated, including: Traverse all adjacent monitoring point pairs with a distance less than a preset distance threshold, and calculate the absolute value of the difference of the comprehensive pollution load indexes of the adjacent monitoring point pairs; When the absolute value of the difference is greater than a first boundary threshold and the comprehensive pollution load index of the high pollution point is greater than a second boundary threshold and the comprehensive pollution load index of the low pollution point is less than the second boundary threshold, mark the corresponding adjacent monitoring point pair as a boundary critical point; For the marked boundary critical point pair, divide the difference of the comprehensive pollution load indexes by the distance between the two points to obtain the boundary gradient strength; Subtract the boundary gradient strength at the previous time by a first preset multiple from the boundary gradient strength at the current time, and subtract the boundary gradient strength at the time two steps ago by a second preset multiple from the boundary gradient strength at the previous time to obtain a boundary drift prediction value. Based on the boundary drift prediction value and the boundary critical point marking, a boundary identification vector is generated.

[0030] In this example, in each data sampling period, based on the spatial layout information, all adjacent monitoring point pairs with a physical distance less than a set threshold (for example, 4.5 meters) are traversed, for each monitoring point pair satisfying the distance constraint condition, the respective comprehensive pollution load index value is read, and the absolute value of the numerical difference between the two points is calculated. The absolute value of the difference must meet two conditions to be judged as a boundary critical area, one is that the difference must be greater than the set first boundary judgment threshold, which is used to eliminate weak fluctuations and retain concentration jumps; the second is that the pollution load index of the higher one in the point pair must be higher than the set second boundary threshold, and the lower one must be lower than the second boundary threshold, forming a spatial jump of pollution concentration across the high risk threshold. For the monitoring point pairs marked as boundary critical, perform spatial gradient calculation, divide the comprehensive pollution load index difference by the actual Euclidean distance between the two points to get the unit distance strength of pollution jump, the greater the boundary gradient strength, the more obvious the pollution mutation in the region. In order to evaluate whether the pollution boundary is expanding or shrinking, a trend modeling mechanism in the time dimension is introduced, the boundary gradient intensity values of each boundary point pair at the current time, the previous time and the previous two times are recorded, and the boundary drift prediction value is calculated by weighted difference, which is composed of the current gradient minus the previous time gradient multiplied by the first preset multiple, and then minus the previous two time gradients multiplied by the second preset multiple. This structure embodies the combination logic of first-order inertia and second-order damping, which is used to judge whether the boundary has a continuous expansion or contraction trend. When the drift prediction value is greater than the positive threshold, it means that the pollution is expanding outward, and the corresponding region fan speed is raised to suppress it; when the prediction value is less than the negative threshold, it means that the pollution boundary tends to shrink, and the energy consumption response can be reduced. The boundary identification vector is generated based on the joint of the boundary critical point marking state and the drift trend.

[0031] In one example, the real-time pressure drop of the inlet side and the outlet side of the filter element in each fan is collected, and the differential pressure compensation gain is calculated according to the real-time pressure drop, including: The real-time pressure drop of the inlet side and the outlet side of the filter element in each fan is collected, and the resistance attenuation ratio is obtained by dividing the real-time pressure drop by the initial pressure drop calibrated when the filter element is first used; The intermediate compensation value is obtained by multiplying the resistance attenuation ratio minus 1 by a first preset coefficient, and the nonlinear adjustment factor is obtained by inputting the resistance attenuation ratio multiplied by a second preset coefficient into a hyperbolic tangent function; The differential pressure compensation gain is obtained by multiplying the intermediate compensation value and the nonlinear adjustment factor and adding 1.

[0032] In this example, a differential pressure sensor is deployed between the inlet side and the outlet side of the filter element in the fan, and the difference between the inlet side pressure and the outlet side pressure is continuously monitored through the real-time data acquisition module to obtain the real-time pressure drop. The initial pressure drop parameter at the initial use of the filter element is called, and the initial pressure drop parameter is obtained through calibration experiments at the initial installation of the filter element and stored in the database as a reference benchmark for dynamic evaluation. The real-time pressure drop currently collected is calculated by ratio with the initial pressure drop to obtain the resistance attenuation ratio of the current state of the filter element. The larger the ratio, the higher the resistance of the filter element. Subtracting 1 from the resistance attenuation ratio represents the relative change value relative to the initial state. The relative change value multiplied by the first preset coefficient obtains the intermediate compensation value constructed linearly, and the setting of the preset coefficient needs to be determined in combination with the empirical weight of the actual wind resistance on the performance impact. Multiply the original resistance attenuation ratio by the second preset coefficient and input it into the hyperbolic tangent function. Through the nonlinear adjustment mechanism, the adjustment is gentle when the resistance changes slightly, and the control response is quickly enhanced when the resistance rises sharply. The output nonlinear adjustment factor. Multiply the intermediate compensation value and the nonlinear adjustment factor to fuse the linear response and nonlinear modulation effect, and add 1 on this basis to maintain a unit gain for the differential pressure compensation gain when the resistance of the filter element is still in the initial state, and dynamically enhance it as the resistance rises, to obtain the differential pressure compensation gain for regulating the air flow and maintaining the performance stability of the fan.

[0033] In one example, for all monitoring points in the responsibility sector of each fan, the target reference speed is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain, including: For all monitoring points in the responsibility sector of each fan, the task priority weight is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain; The task priority weights of each fan are summed to obtain the total weight, and the task priority weights of each fan are divided by the total weight to obtain the relative allocation coefficient; The rated maximum speed is subtracted from the lowest silent speed to obtain the speed adjustment range, and the lowest silent speed is added to the product of the relative allocation coefficient and the speed adjustment range to obtain the target reference speed.

[0034] In this example, the responsibility sector range of each fan is explicitly defined, and all configured environmental monitoring points within the responsibility sector range are traversed. For each monitoring point, the comprehensive pollution load index at the current time is extracted, which reflects the spatial distance weighted comprehensive pollution intensity of the multi-source pollutants (such as PM2.5, VOC, and microorganisms) monitored by the monitoring point. At the same time, the boundary identification vector is analyzed to indicate whether the current monitoring point is located in the pollution boundary area and its boundary drift degree, thereby embodying the strategic importance of pollution diffusion prevention and control. Combined with the pressure difference compensation gain calculated based on the state of the filter core corresponding to the current fan, the compensation adjustment requirement of the filter core resistance rise on the fan air supply capacity is reflected, and the influence of the filter core aging degree on the operation energy efficiency is indirectly reflected. The comprehensive pollution load index, boundary identification vector value, and pressure difference compensation gain are combined according to the set weighting factors to calculate the weight contribution value of the monitoring point in the current fan control task, and the weight values of all monitoring points in the responsibility sector are summed to obtain the task priority weight of the fan. After completing the task priority weight calculation of all fans, the total weight is obtained by adding the task priority weights of all fans again, and the task priority weight of each fan is normalized by dividing the total weight to obtain the relative allocation coefficient reflecting how the fan operation resources should be dynamically allocated under the current global pollution load spatial distribution. The difference between the rated maximum speed and the minimum silent speed of the fan is read to obtain the current adjustable speed adjustment range of the fan, and the minimum silent speed is added to the product of the relative allocation coefficient and the speed adjustment range to obtain the target reference speed of the fan under the current environmental state.

[0035] In one example, all monitoring points in the responsibility sector of each fan are traversed, and the task priority weight is calculated based on the comprehensive pollution load index, boundary identification vector, and pressure difference compensation gain, including: All monitoring points in the responsibility sector of each fan are traversed, and the boundary critical point is determined based on the boundary identification vector. When the monitoring point is a boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1 plus twice the boundary indication value, and when the monitoring point is a non-boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1. The comprehensive pollution load index of each monitoring point is multiplied by the corresponding boundary weight factor and then multiplied by the pressure difference compensation gain of the fan, and the sum of all monitoring points in the responsibility sector is divided by the total number of monitoring points in the responsibility sector to obtain the task priority weight.

[0036] In this example, all the monitoring points in each fan's responsibility sector are traversed one by one based on the responsibility sector covered by each fan, and the spatial characteristics of the monitoring points are classified in combination with the boundary identification vector. In the traversal process, it is judged whether the monitoring point is a boundary critical point according to the marking value of the corresponding monitoring point in the boundary identification vector, wherein the boundary identification vector has time-varying and region-sensitive characteristics. When the judgment result is a boundary critical point, it indicates that the monitoring point is in the edge area of the pollutant concentration mutation, which has a higher regulation value. At this time, the boundary weight factor of the monitoring point is set to 1 plus 2 times the boundary indication value, wherein the boundary indication value represents the boundary obviousness with a normalized real number from 0 to 1. When it is judged as a non-boundary critical point, i.e. the marking value is zero, the boundary weight factor is directly assigned as 1 to keep the basic weight of the monitoring point unchanged for the task priority evaluation. After completing the assignment of the boundary weight factor of all monitoring points, the comprehensive pollution load index of each monitoring point is multiplied by its boundary weight factor to amplify or maintain its contribution evaluation to the pollution control task. The product result is multiplied by the current differential pressure compensation gain of the fan, thereby introducing the influence factor of the filter pressure loss state on the actual air volume output capacity, reflecting the difference in regulation capacity of the fan under different filter attenuation levels. Through the weighted result, the contribution values of all monitoring points in the fan's responsibility sector are summed and divided by the total number of monitoring points in the sector, realizing the average evaluation of the comprehensive regulation load of the fan under the current pollution space state, and obtaining the task priority weight of the fan.

[0037] In one example, the average pollution load index of each fan's responsibility sector is calculated and fuzzy reasoning is performed to obtain a speed increment, and a PWM duty cycle is generated according to a target reference speed and the speed increment to drive each fan, including: The average pollution load index is obtained by averaging the comprehensive pollution load indexes of all monitoring points in each fan's responsibility sector, and the deviation amount is obtained by subtracting the target pollution load index from the average pollution load index. The deviation change rate is obtained by subtracting the deviation amount at the previous time from the deviation amount at the current time. Fuzzy reasoning is performed based on the deviation amount and the deviation change rate to obtain a speed increment. The target reference speed is superimposed with the limited speed increment to obtain an expected speed, the expected speed is divided by the rated maximum speed and multiplied by 100% to obtain a PWM duty cycle, and each fan is driven according to the PWM duty cycle.

[0038] In this example, the comprehensive pollution load index of all monitoring points in the responsibility sector covered by each fan is read point by point and arithmetically averaged to obtain the average pollution load index of the responsibility sector at the current time, reflecting the overall level of the distribution of pollutants in the current fan control area. The average pollution load index is subtracted from the preset target pollution load index to obtain the pollution load deviation of the fan at the current time relative to the target control state, which is used to measure the distance between the current pollution control effectiveness and the set control target, and reflects the real-time deviation trend of the regulation effect. The deviation at the current time and the deviation at the previous time are differentiated to obtain the pollution load deviation change rate, which reflects whether the pollution load trend is rising, falling or stable. The pollution load deviation and the deviation change rate are input into the fuzzy reasoning system, which maps the deviation and the change rate to a fuzzy set based on a series of pollution control experience rules and a set of language variables, and then performs condition matching and weight fusion through a fuzzy rule base, and outputs a continuous variable as the fan speed increment for the current regulation period through the defuzzification process. The upper and lower limit amplitudes are applied to the speed increment to prevent sudden regulation from causing a sharp increase in energy consumption or instability. The limited amplitude speed increment output by the fuzzy reasoning system is added to the target reference speed to obtain the expected target fan speed for the current control period, which represents the optimal operating speed under the dynamic response of pollution load and the constraint of filter state. In order to adapt to the PWM control mode of the fan, the expected speed value is divided by the rated maximum speed of the fan to obtain a normalized percentage representation, and multiplied by 100% to convert to the actual PWM duty cycle. The actual PWM duty cycle parameter is used to drive the PWM module of the corresponding fan through the driver control interface.

[0039] In one example, fuzzy reasoning is performed based on the deviation and the deviation change rate to obtain the speed increment, including: quantizing the deviation into a corresponding M first fuzzy levels according to a preset interval threshold, and quantizing the deviation change rate into a corresponding M second fuzzy levels; inputting the deviation into a triangular membership function calculation to calculate a first membership value of the first fuzzy level, and inputting the deviation change rate into a triangular membership function calculation to calculate a second membership value of the second fuzzy level; inputting the first membership value and the second membership value into a fuzzy reasoning rule base, calculating the activation strength of each rule by the max-min reasoning method to obtain an output fuzzy set, and calculating the speed increment based on the output fuzzy set.

[0040] In this example, according to the set pollution load control strategy, the deviation and the deviation change rate are subjected to fuzzy quantization processing, wherein the deviation represents the difference between the current average pollution load and the target pollution load, and the deviation change rate represents the trend of the difference over time. A symmetric threshold interval covering negative deviation, zero deviation and positive deviation is preset, and the threshold interval is divided into M fuzzy levels. The fuzzy levels are described by language variables, such as "negative large", "negative medium", "negative small", "zero", "positive small", "positive medium" and "positive large", which are used to represent different degrees of deviation. By corresponding the deviation to the threshold interval, any actual deviation is quantized to a first fuzzy level or two adjacent first fuzzy levels, and the deviation change rate is quantized to a second fuzzy level in the same way. A triangular membership function is used to act on the deviation and the deviation change rate respectively to calculate their membership values in each fuzzy level, wherein the first membership value is used to describe the belonging degree of the current deviation in the first fuzzy level, and the second membership value is used to describe the belonging degree of the deviation change rate in the second fuzzy level. The first membership value and the second membership value are input into a predefined fuzzy rule base, the fuzzy rule base is composed of combinations of language variables as premises, and the output fuzzy level is combined with the corresponding control response to form the result. For example, "if the deviation is positive large and the deviation change rate is positive medium, then the output speed increment is large speed increment". The maximum and minimum inference method is used to perform traversal operation on all rules, the minimum value of the membership values of the premise part is taken as the activation strength of the corresponding rule, and the activation strength is mapped to the output fuzzy level to form an output fuzzy set. Multiple rules can be activated at the same time, so the normalization superposition fusion is performed on the output fuzzy sets of all activated rules. By performing the defuzzification process on the output fuzzy set, a deterministic speed increment value is extracted from the continuous output space by using the barycenter method or the maximum membership method.

[0041] Referring Figure 2 The embodiment provides a fan speed regulation system based on dynamic pollution load, which comprises: A collection module 1 is used for collecting target PM2.5 concentrations, target VOC concentrations and target microorganism concentrations of each monitoring point and calculating comprehensive pollution load indexes of the monitoring points; A traversal module 2 is used for traversing boundary critical points in all monitoring points based on the comprehensive pollution load indexes and generating a boundary identification vector; A compensation module 3 is used for collecting real-time pressure drops of the inlet side and the outlet side of the filter element in each fan and calculating a differential pressure compensation gain according to the real-time pressure drops; A calculation module is used for calculating target reference speeds of all monitoring points in the responsibility sector of each fan according to the comprehensive pollution load indexes, the boundary identification vector and the differential pressure compensation gain; The driving module 5 is configured to calculate the average pollution load index of the responsibility sector of each fan, perform fuzzy reasoning to obtain a speed increment, and generate a PWM duty cycle according to a target reference speed and the speed increment to drive each fan.

[0042] In the embodiments, the specific implementation of each unit in the system is described in the above method embodiments, and will not be described here.

[0043] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium provided by the present application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0044] It should be noted that in this document, the terms "comprise", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion, so that processes, systems, articles or methods that include a series of elements not only include those elements, but also include other elements not explicitly listed, or include elements inherent to such processes, systems, articles or methods. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, system, article or method that includes the element.

[0045] The above description is only the preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, as described in the specification and drawings of the present application, are also included in the patent protection scope of the present application.

Claims

1. A method of adjusting the rotational speed of a fan based on dynamic pollution load, characterized by, comprise: collecting target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculating comprehensive pollution load index of each monitoring point; traversing boundary critical points in all monitoring points based on the comprehensive pollution load index and generating a boundary identification vector; collecting real-time pressure drop of the inlet side and the outlet side of the filter element in each fan, and calculating differential pressure compensation gain according to the real-time pressure drop; for all monitoring points in the responsibility sector of each fan, calculating target reference speed according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain; calculating average pollution load index of the responsibility sector of each fan and performing fuzzy reasoning to obtain speed increment, and generating PWM duty cycle according to the target reference speed and the speed increment to drive each fan.

2. The method of claim 1, wherein, The collecting target PM2.5 concentration, target VOC concentration and target microorganism concentration of each monitoring point, and calculating comprehensive pollution load index of each monitoring point, comprises: collecting original PM2.5 concentration, original VOC concentration and original microorganism concentration of each monitoring point; summing the original PM2.5 concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain the target PM2.5 concentration; summing the original VOC concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain the target VOC concentration; summing the original microorganism concentration of each monitoring point at the current time and the previous two times and dividing by 3 to obtain the target microorganism concentration; calculating the comprehensive pollution load index of each monitoring point according to the target PM2.5 concentration, the target VOC concentration and the target microorganism concentration.

3. The method of claim 2, wherein, The calculating the comprehensive pollution load index of each monitoring point according to the target PM2.5 concentration, the target VOC concentration and the target microorganism concentration, comprises: calculating the distance ratio of the distance from each monitoring point to the nearest fan to the maximum monitoring radius, and multiplying the distance ratio by a first target value to obtain a spatial distance attenuation factor as an exponential term; dividing the target PM2.5 concentration by a PM2.5 standard threshold value, multiplying a second target value and the spatial distance attenuation factor to obtain a PM2.5 weighted component, dividing the target VOC concentration by a VOC standard threshold value, multiplying a third target value and the spatial distance attenuation factor to obtain a VOC weighted component, and dividing the target microorganism concentration by a microorganism standard threshold value, multiplying a fourth target value and the spatial distance attenuation factor to obtain a microorganism weighted component; summing the PM2.5 weighted component, the VOC weighted component and the microorganism weighted component to obtain the comprehensive pollution load index of each monitoring point.

4. The method of claim 1, wherein, The traversing boundary critical points in all monitoring points based on the comprehensive pollution load index and generating a boundary identification vector, comprises: traversing all adjacent monitoring point pairs with a distance less than a preset distance threshold, and calculating the absolute value of the difference of the comprehensive pollution load index of the adjacent monitoring point pairs; mark a corresponding adjacent monitoring point pair as a boundary critical point when the difference absolute value is greater than a first boundary threshold value and the comprehensive pollution load index of the high pollution point is greater than a second boundary threshold value and the comprehensive pollution load index of the low pollution point is less than the second boundary threshold value; for the marked boundary critical point pair, the comprehensive pollution load index difference is divided by the distance between the two points to obtain a boundary gradient intensity; the boundary gradient intensity at the current time is reduced by a first preset multiple of the boundary gradient intensity at the previous time, and then reduced by a second preset multiple of the boundary gradient intensity at the previous two times to obtain a boundary drift prediction value, and a boundary identification vector is generated based on the boundary drift prediction value and the boundary critical point marking.

5. The method for adjusting the rotational speed of a fan based on dynamic pollution load according to claim 1, wherein, The real-time pressure drop of the inlet side and the outlet side of the filter element in each fan is collected, and the differential pressure compensation gain is calculated according to the real-time pressure drop, comprising: Collecting the real-time pressure drop of the inlet side and the outlet side of the filter element in each fan, dividing the real-time pressure drop by the initial pressure drop calibrated when the filter element is used for the first time to obtain a resistance attenuation ratio; The intermediate compensation value is obtained by multiplying the resistance attenuation ratio by a first preset coefficient after subtracting 1, and the nonlinear adjustment factor is obtained by inputting the resistance attenuation ratio multiplied by a second preset coefficient into a hyperbolic tangent function. The differential pressure compensation gain is obtained by multiplying the intermediate compensation value and the nonlinear adjustment factor and then adding 1.

6. The method of claim 1, wherein, The target reference speed is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain for all monitoring points in the responsibility sector of each fan, comprising: The task priority weight is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain by traversing all monitoring points in the responsibility sector of each fan; The total weight is obtained by summing the task priority weights of each fan, and the relative allocation coefficient is obtained by normalizing the task priority weights of each fan by dividing the total weight; The target reference speed is obtained by subtracting the lowest silent speed from the rated maximum speed, and adding the product of the lowest silent speed, the relative allocation coefficient and the speed adjustment range.

7. The method of claim 6, wherein, The task priority weight is calculated according to the comprehensive pollution load index, the boundary identification vector and the differential pressure compensation gain by traversing all monitoring points in the responsibility sector of each fan, comprising: According to the boundary identification vector, it is judged whether each monitoring point in the responsibility sector is a boundary critical point; When the monitoring point is a boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1 plus 2 times the boundary indication value, and when the monitoring point is a non-boundary critical point, the boundary weight factor of the corresponding monitoring point is set to 1; The comprehensive pollution load index of each monitoring point is multiplied by the corresponding boundary weight factor and the differential pressure compensation gain of the fan, and then the sum of all monitoring points in the responsibility sector is divided by the total number of monitoring points in the responsibility sector to obtain the task priority weight.

8. The method of claim 1, wherein, The method comprises the following steps: averaging the comprehensive pollution load indexes of all monitoring points in the responsibility sector of each fan to obtain an average pollution load index, subtracting a target pollution load index from the average pollution load index to obtain a deviation, subtracting a deviation at a previous moment from a deviation at a current moment to obtain a deviation change rate; performing fuzzy reasoning based on the deviation and the deviation change rate to obtain a speed increment; superimposing the target reference speed and the speed increment after limiting to obtain an expected speed, dividing the expected speed by a rated maximum speed and then multiplying by 100% to obtain a PWM duty cycle, and driving each fan according to the PWM duty cycle.

9. The method of claim 8, wherein, The method comprises the following steps: quantifying the deviation into corresponding M first fuzzy levels according to a preset interval threshold, and quantifying the deviation change rate into corresponding M second fuzzy levels; inputting the deviation into a triangular membership function to calculate a first membership value corresponding to the first fuzzy level, and inputting the deviation change rate into a triangular membership function to calculate a second membership value corresponding to the second fuzzy level; inputting the first membership value and the second membership value into a fuzzy reasoning rule base, calculating the activation strength of each rule by a max-min reasoning method to obtain an output fuzzy set, and calculating a speed increment based on the output fuzzy set.

10. A fan speed regulation system based on dynamic pollution load, characterized by, The method comprises the following steps: a collection module for collecting target PM2.5 concentrations, target VOC concentrations and target microorganism concentrations of each monitoring point and calculating comprehensive pollution load indexes of each monitoring point; a traversal module for traversing boundary critical points in all monitoring points based on the comprehensive pollution load indexes and generating a boundary identification vector; a compensation module for collecting real-time pressure drops of the inlet side and the outlet side of the filter element in each fan and calculating a differential pressure compensation gain according to the real-time pressure drops; a calculation module for calculating target reference speeds of all monitoring points in the responsibility sector of each fan according to the comprehensive pollution load indexes, the boundary identification vector and the differential pressure compensation gain; a driving module for calculating average pollution load indexes of the responsibility sector of each fan and performing fuzzy reasoning to obtain a speed increment, and generating a PWM duty cycle according to the target reference speed and the speed increment to drive each fan.

Images