Building environment data acquisition method based on air quality dynamic perception

Abstract

The invention discloses a building environment data acquisition method based on air quality dynamic perception, and particularly relates to the field of building and industrial data acquisition, the method comprises the following steps: in a normal state, first-class data and second-class data of a building environment are acquired in real time, and the first-class data comprises real-time sensing data; the second type of data comprises air quality data missing factors and pollutant abnormal fluctuation factors; and before the first-class data is output, pre-constructing a pre-influence condition based on the second-class data, if the pre-influence condition is judged to be abnormal, reconstructing a first influence condition and a second influence condition based on the second-class data, and otherwise, directly outputting the first-class data. According to the scheme, real-time data collection and a chemical memory accumulation model are combined, and the limitation that a traditional sensor only monitors the air quality at the moment is broken through; by calculating the residual characteristics of the pollutants, the air composition state at a certain moment can be traced back, and the pollutant traceability and the environment change interpretation capacity are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of building and industrial data acquisition. More specifically, the present invention relates to a method for collecting building environment data based on dynamic air quality perception. Background Art

[0002] In the building and industrial environments, the perception of air quality is usually real-time, that is, only the concentration of relevant pollutants at a certain moment is recorded by sensors. However, the chemical components in the air do not change instantaneously. For example, NO2, VOCs, CO2, microbial spores, etc. actually show certain hysteresis and accumulation characteristics under the action of various factors such as aerodynamics, surface adsorption, temperature and humidity.

[0003] In other words, each building environment has a certain air "chemical memory", and this characteristic can be used to trace back the air conditions at a certain moment in the past and even infer the historical activity patterns inside the building.

[0004] In the prior art, traditional air quality monitoring methods focus on the air quality "at this moment" and lack a dynamic combination acquisition method, so it is not conducive to applications in the building and industrial environments. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for collecting building environment data based on dynamic air quality perception. By calculating the residual characteristics of pollutants, including influencing factors such as adsorption, release, and diffusion, the present solution can trace back the air component state at a certain moment and further infer the historical activity patterns inside the building to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solution: A method for collecting building environment data based on dynamic air quality perception, including:

[0007] Under normal conditions, collect a first type of data and a second type of data of the building environment in real time. The first type of data includes real-time sensing data; the second type of data includes factors missing air quality data and factors of abnormal fluctuations of pollutants.

[0008] Before the first type of data is output, pre-construct a pre-influence condition based on the second type of data. If the pre-influence condition is determined to be abnormal, re-construct a first influence condition and a second influence condition based on the second type of data, otherwise directly output the first type of data.

[0009] The pre-influence condition includes: if the data continuity interval value is greater than a preset data continuity interval threshold, or the instantaneous change rate of pollutants is greater than a preset instantaneous change rate threshold of pollutants.

[0010] If the first influencing condition or the second influencing condition is determined to be in a state to be inspected, the secondary data is defined as abnormal; otherwise, the primary data and the secondary data are collected again, and the determination of the pre-influencing conditions is performed again. If the number of times of re-execution is greater than two, an artificial verification warning signal is issued.

[0011] When the secondary data is determined to be abnormal, chemical memory information is obtained, an air perception model is established based on the chemical memory information and the secondary data, an air perception value is calculated based on the air perception model, and the output of data collection is realized by combining the air perception value with the primary data.

[0012] In a preferred embodiment, the factors for the lack of air quality data include the data continuity interval value and the sensor health score.

[0013] The factors for abnormal pollutant fluctuations include the instantaneous change rate of pollutants and the pollutant decay rate.

[0014] The first influencing condition includes: if the data continuity interval value is greater than the preset data continuity interval threshold and the sensor health score is less than the preset sensor health score threshold, it is determined to be in a state to be inspected.

[0015] The second influencing condition includes: if the instantaneous change rate of pollutants is greater than the preset instantaneous change rate threshold of pollutants and the pollutant decay rate is less than the preset pollutant decay rate threshold, it is determined to be in a state to be inspected.

[0016] In a preferred embodiment, a data continuity model S is constructed based on the factors for the lack of air quality data. T , the data continuity model S T evaluates the stability of data collection by describing the data interval, the sensor health status, and the environmental interference factors; the data continuity model S T is expressed as:

[0017] The time interval between two consecutive data points is described by the data interval value ΔT included in the factors for the lack of air quality data; the current health status of the sensor is described by the sensor health score H included in the factors for the lack of air quality data.

[0018] In the data continuity model S T a time decay factor η is introduced to describe the degree of influence of the control time interval on data continuity; and a sensor influence coefficient ξ is introduced to adjust the influence strength of the health status on the score in the data continuity model S T ; finally, the value output by S T is regarded as the data continuity score.

[0019] In a preferred embodiment, during the calculation of the sensor health score H: the remaining battery percentage of the sensor is described by the ratio of the current sensor battery level to the peak sensor battery level, to measure the continuous operation ability of the sensor; in addition, the error rate in the most recent N data packets is combined for calculation to form the calculation of the sensor health score H; the sensor health score H is expressed as:

[0020] where P power is the current sensor battery level; P max is the peak sensor battery level; E error is the number of errors in the most recent N data packets; E total is the total number of the most recent N data packets.

[0021] In a preferred embodiment, a pollutant fluctuation detection model S is constructed based on the pollutant abnormal fluctuation factors included in the secondary data P , and the pollutant fluctuation detection model S P is used to identify pollutant abnormal fluctuations. The construction of the pollutant fluctuation detection model S P is described as a non-linear detection model for identifying abnormalities in both sudden and slow-changing environments; the pollutant fluctuation detection model S P is described as: In the pollutant fluctuation detection model S P , the instantaneous change rate of the pollutant is non-linearly normalized by using the arctangent function tan -1 , and an exponential adjustment term e -βR decay is used to identify the long-term retention situation of the pollutant;

[0022] In the pollutant fluctuation detection model S P , the change rate of the pollutant concentration within a short period of time is described by the instantaneous change rate R inst of the pollutant, and the instantaneous change rate R inst of the pollutant is defined as: In the pollutant fluctuation detection model S P , the first-order decay model of the pollutant decay rate R decay [[ID=!47]]is used to show how the pollutant decays over time; the definition of the pollutant decay rate R decay is:

[0023] where C t and C t-Δt represent the pollutant concentrations at the current moment and the previous moment respectively; Δt is the time interval; is the time normalization factor.

[0024] In a preferred embodiment, a chemical memory accumulation model M is constructed based on chemical memory information to quantify the environmental residues of pollutants, calculate the impacts of pollutants remaining on different surfaces, and when current data is abnormal, trace back the past pollutant concentrations;

[0025]

[0026] where the historical pollutant concentration C i represents the historical pollution data stored in the chemical memory accumulation model M; t i is the pollutant deposition time; W i is the air flow rate; λ i is the pollutant adsorption coefficient; γ i is the pollutant degradation coefficient; δ i is the air flow impact factor; n is the total number of pollutant types, i is the index, and each i represents a specific pollutant; k a is the adsorption rate constant; k d is the desorption rate constant; P is the pollutant concentration.

[0027] In a preferred embodiment, the free diffusion of pollutants in the air is described by establishing a convective-diffusion equation; the convective-diffusion equation is expressed as:

[0028]

[0029] where C is the pollutant concentration; is the time change rate of the pollutant concentration; represents the air velocity vector; D is the turbulent diffusion coefficient in the air; the diffusion term of the pollutant in the above equation represents how the pollutant moves from a high-concentration region to a low-concentration region through diffusion; S is the source term of the pollutant; the convective term represents the drift of the pollutant along with the air flow;

[0030] A boundary transport model is constructed through the pollutant transport flux per unit area to describe the interaction of pollutants in the building surface and ventilation system;

[0031]

[0032] where J boundary is the pollutant transport flux per unit area; k s is the deposition rate of the pollutant on the surface; C wall is the pollutant concentration in the wall or other boundary materials; represents the flow exchange of pollutants in areas such as building ventilation openings, windows, etc.; V represents the air control volume in the building environment;

[0033] The final air flow and diffusion model DF Constructed jointly by the convection-diffusion equation combined with the boundary transport model:

[0034] In a preferred embodiment, the final air perception value APV is expressed as:

[0035]

[0036] where the air perception value APV is used for final data output; μ, ν, ω, ρ, θ, κ, σ, λ are adjustment parameters, and the adjustment parameters are used to control the influence weights of different factors;

[0037] In the above formula, the exponential term e ωM is used to highlight the influence of chemical memory; the power term is used to ensure the importance of data quality and pollution detection; the non-linear expansion term cosh(σM) introduces non-linear adjustment by applying the hyperbolic cosine function cosh(x), and is used to smooth the long-term influence of pollutants; the normalization term is used to balance the aerodynamic influence.

[0038] In a preferred embodiment, the final data acquisition value Q is jointly determined by a class of data D1 and the air perception value APV:

[0039]

[0040] where tanh(δM) is the hyperbolic tangent function, and in the above formula, the hyperbolic tangent function is used to smooth the influence of chemical memory; α, β, γ, δ, ζ are model adjustment parameters;

[0041] Introduce an abnormal pollution correction to Q, and output the corrected air quality acquisition data Q * ;

[0042]

[0043] where S P -S T is the difference between the pollutant fluctuation score and the data continuity score; the exponential term is used to reduce the fluctuation influence under unstable data conditions; ξ is the pollutant abnormal correction coefficient.

[0044] Technical effects and advantages of the present invention:

[0045] 1. By combining real-time data collection with a chemical memory accumulation model, this solution breaks through the limitation of traditional sensors that only monitor the air quality "at this moment". By calculating the residual characteristics of pollutants, including influencing factors such as adsorption, release, and diffusion, this solution can trace back the air composition state at a certain moment and further infer the historical activity patterns inside the building, improving the ability to trace pollutants and the explanatory power of environmental changes;

[0046] 2. A data continuity model is used to evaluate the health status of sensors, data loss, and transmission anomalies. An abnormal fluctuation detection model is constructed by combining the instantaneous change rate and attenuation rate of pollutants. Through non-linear normalization and time decay correction, false alarms caused by data anomalies are reduced, and at the same time, the real-time capture ability of pollution events is improved;

[0047] 3. The air perception model proposed in this solution integrates data continuity scores, pollutant fluctuation scores, aerodynamic influence factors, and chemical memory information. Through mathematical models of power functions, exponential functions, and hyperbolic cosine functions, accurate judgment of pollutant anomalies is achieved, and false alarms caused by short-term fluctuations are reduced, improving the credibility of air quality data;

[0048] 4. Traditional monitoring methods are difficult to accurately evaluate the flow characteristics of pollutants in the building environment. This solution uses the convective diffusion equation combined with the boundary transport model to simulate the spatial diffusion behavior of air pollutants, and corrects the data in combination with environmental factors, improving the applicability of data collection in complex building environments;

[0049] 5. By introducing an abnormal pollution correction mechanism, calculating the deviation between the pollutant fluctuation score and the data continuity score, and using an exponential decay function for adjustment, false alarms caused by short-term fluctuations of sensors or abnormal external inputs are reduced, and the response accuracy of air quality monitoring and collection in emergencies is improved;

[0050] 6. By integrating multi-source information such as aerodynamic factors and pollutant diffusion models, an intelligent data fusion architecture is formed, improving the applicability of air quality data collection, enabling it to adapt to different building environments, and optimizing the accuracy of pollution monitoring and tracing. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is the execution flow of the method of the present invention Figure 1 .

[0052] Figure 2 is the execution flow of the method of the present invention Figure 1 .

[0053] Figure 3 is the overall flowchart of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0055] Referring to the attached Figures 1 - 3 drawings, a method for collecting building environment data based on dynamic air quality perception according to an embodiment of the present invention includes:

[0056] Under normal conditions, collect the first type of data and the second type of data of the building environment in real time. The first type of data includes real-time sensing data; the second type of data includes air quality data missing factors and pollutant abnormal fluctuation factors;

[0057] Before the first type of data is output, pre-construct a pre-influence condition based on the second type of data. If the pre-influence condition is determined to be abnormal, re-construct the first influence condition and the second influence condition based on the second type of data, otherwise directly output the first type of data;

[0058] The pre-influence condition includes: if the data continuity interval value is greater than the preset data continuity interval threshold, or the pollutant instantaneous change rate is greater than the preset pollutant instantaneous change rate threshold;

[0059] If the first influence condition or the second influence condition is determined to be in a state to be inspected, define the second type of data as abnormal, otherwise re-collect the first type of data and the second type of data, and re-execute the determination of the pre-influence condition; if the number of re-executions is greater than two, send out an artificial verification warning signal;

[0060] When the second type of data is determined to be abnormal, obtain chemical memory information, establish an air perception model based on the chemical memory information and the second type of data, calculate the air perception value based on the air perception model, and realize the output of data collection by combining the air perception value with the first type of data.

[0061] The air quality data missing factors include the data continuity interval value and the sensor health score;

[0062] The pollutant abnormal fluctuation factors include the pollutant instantaneous change rate and the pollutant attenuation rate;

[0063] The first influence condition includes: if the data continuity interval value is greater than the preset data continuity interval threshold and the sensor health score is less than the preset sensor health score threshold, it is determined to be in a state to be inspected;

[0064] The second influencing condition includes: if the instantaneous change rate of the pollutant is greater than the preset instantaneous change rate threshold of the pollutant and the pollutant decay rate is less than the preset pollutant decay rate threshold, it is determined to be in a state to be inspected.

[0065] Construct a data continuity model S based on the air quality data missing factors T , the data continuity model S T evaluates the data acquisition stability by describing the data interval, the sensor health status, and the environmental interference factors; the data continuity model S T is expressed as:

[0066] Describe the time interval between two consecutive data points through the data interval value ΔT included in the air quality data missing factors; describe the current health status of the sensor through the sensor health score H included in the air quality data missing factors. The value range of H includes but is not limited to 0 to 1. In the above formula, if the sensor health is lower, the greater the possibility of sensor failure;

[0067] In the data continuity model S T introduce a time decay factor η to describe the degree of influence of the control time interval on data continuity; and introduce a sensor influence coefficient ξ in the data continuity model S T to adjust the influence strength of the health on the score; finally, S T The output value is regarded as the data continuity score. The value range of the data continuity score is (0, 1]. In practical applications, the lower the data continuity score, the worse the data quality; in addition, e in the above formula is the base of the natural logarithm, which is used to describe the exponential decay effect.

[0068] In the calculation process of the sensor health score H: describe the remaining battery proportion of the sensor by the ratio of the current sensor battery to the peak sensor battery, and measure the continuous operation ability of the sensor; in addition, combine the calculation of the error rate in the recent N data packets to form the calculation of the sensor health score H. The sensor health score H is used to dynamically characterize the availability of the sensor. The larger the value of the sensor health score H, the better the health status, and the lower the possibility of data loss and error; the sensor health score H is expressed as:

[0069] where P power is the current sensor battery; P max is the peak sensor battery; E error is the number of errors in the recent N data packets; E total is the total number of the recent N data packets;

[0070] Through the calculation of the sensor health score H, S TThe score is more representative and can better evaluate data continuity issues.

[0071] Construct a pollutant fluctuation detection model S based on the abnormal fluctuation factors of pollutants included in the secondary data P , the pollutant fluctuation detection model S P is used to identify abnormal fluctuations of pollutants. Since the changes in air pollutants are not linear but may show sudden changes or long-term slow accumulations due to the coupled effects of multiple factors such as the release rate of pollution sources, aerodynamic diffusion, and building ventilation, the construction of the pollutant fluctuation detection model S P is described as a non-linear detection model, which can accurately identify abnormalities in both sudden change and slow change environments; the pollutant fluctuation detection model S P is described as: In the pollutant fluctuation detection model S P , the non-linear normalization of the instantaneous change rate of pollutants is performed by using the arctangent function tan -1 to ensure numerical stability, and the exponential adjustment term e -βR decay is used to identify the long-term retention of pollutants;

[0072] In the pollutant fluctuation detection model S P , the change rate of pollutant concentration in a short period of time is described by the instantaneous change rate R of pollutants inst , and the instantaneous change rate R of pollutants inst is defined as: In the pollutant fluctuation detection model S P , the first-order decay model of the pollutant decay rate R decay is used to show how pollutants decay over time; the pollutant decay rate R decay is defined as: In the pollutant decay rate R decay , the negative sign in the formula indicates that the concentration decreases over time;

[0073] where C t and C t-Δt represent the pollutant concentrations at the current moment and the previous moment respectively; Δt is the time interval; is the time normalization factor, which is used to prevent the data fluctuation from being distorted by the time ratio; dC is the small change in pollutant concentration, representing the instantaneous increase or decrease in concentration; dt is the small change in time, representing the time interval corresponding to the concentration change; represents the derivative of pollutant concentration with respect to time, indicating the decay rate of pollutants over time.

[0074] Construct a chemical memory accumulation model M based on chemical memory information to quantify the environmental residues of pollutants, calculate the effects of pollutants remaining on different surfaces, and retrieve the past pollutant concentrations when the current data is abnormal;

[0075]

[0076] where the historical pollutant concentration C i represents the historical pollution data stored in the chemical memory accumulation model M; t i is the pollutant deposition time; W i is the air velocity; λ i is the pollutant adsorption coefficient. In practical applications, the surface adsorption ability of different pollutants can be described by the pollutant adsorption coefficient; γ i is the pollutant degradation coefficient; δ i is the air flow influence factor. In the above formula, the air flow influence factor is used to affect the pollutant retention; n is the total number of pollutant types, i is the index, and each i represents a specific pollutant; k a is the adsorption rate constant; k d is the desorption rate constant; P is the pollutant concentration.

[0077] The free diffusion of pollutants in the air is described by establishing a convective diffusion equation; the convective diffusion equation is expressed as:

[0078]

[0079] where C is the pollutant concentration, and C varies with time and space; is the time change rate of the pollutant concentration; represents the air velocity vector, which includes the air flow velocities in the x, y, and z directions; D is the turbulent diffusion coefficient in the air, and the turbulent diffusion coefficient in the air is related to the air temperature and humidity; the diffusion term of the pollutant in the above formula represents how the pollutant moves from the high-concentration area to the low-concentration area through diffusion; S is the source term of the pollutant, such as pollutants generated by indoor release - combustion, outdoor input - vehicle exhaust, and biological metabolism - human respiration; the convective term represents the drift of the pollutant along with the air flow;

[0080] The boundary transfer model is constructed through the pollutant transfer flux per unit area to describe the interaction of pollutants on the building surface and in the ventilation system. Because in the building environment, pollutants not only move in the air but also exchange with surfaces such as walls, vents, floors, and ceilings;

[0081]

[0082] where J boundary is the pollutant transfer flux per unit area; k s is the deposition rate of the pollutant on the surface, and its deposition rate depends on the adsorption characteristics of the material; C wallis the pollutant concentration in wall or other boundary materials; represents the flow and exchange of pollutants in areas such as building ventilation openings, windows, etc.; V represents the air control volume in the building environment; the above formula consists of two parts: the first part is the pollutant deposition term -k s (C - C wall ), the pollutant deposition term represents the rate of pollutant deposition from the air to the wall; the second part is the air circulation term The air circulation term represents the transfer of pollutants by air flow through the boundary;

[0083] The final air flow and diffusion model D F is jointly constructed by combining the convection - diffusion equation with the boundary transport model:

[0084] The above formula combines spatial diffusion, air flow, and boundary interaction, ensuring that the system has a high - precision prediction ability for the spatial behavior of pollutants. The above formula can adjust k s and to adapt to different environmental requirements; in addition, dV is the volume element.

[0085] The final air perception value APV is expressed as:

[0086]

[0087] where the air perception value APV is used for the final data output; μ, ν, ω, ρ, θ, κ, σ, λ are adjustment parameters, and the adjustment parameters are used to control the influence weights of different factors;

[0088] In the above formula, the exponential term e ωM is used to highlight the influence of chemical memory. When the historical pollutant residue is high, its role is amplified; the power term is used to ensure the importance of data quality and pollution detection; the non - linear expansion term cosh(σM) introduces a soft non - linear adjustment by applying the hyperbolic cosine function cosh(x), which is used to smooth the long - term influence of pollutants; the normalization term is used to balance the aerodynamic influence and prevent over - amplifying the contribution of air flow.

[0089] The final data acquisition value Q is jointly determined by a class of data D1 and the air perception value APV:

[0090]

[0091] Where tanh(δM) is the hyperbolic tangent function, and in the above formula, the hyperbolic tangent function is used to smooth the influence of chemical memory to prevent its excessive influence; α, β, γ, δ, ζ are model adjustment parameters, where α, β, γ control the weights of real-time data, sensed values, and chemical memory respectively, and δ, ζ are used to affect the dynamic characteristics and normalization scale of non-linear fusion;

[0092] What needs to be further explained for the above formula in application is:

[0093] Exponentially weighted term APV β :

[0094] D1 of real-time data is used as the base value to ensure that the original information of the sensor is not lost. The air sensed value APV corrects the data to supplement environmental information and the influence of historical pollution. In addition, the influence ratio of the two is adjusted by α and β, and the parameters can be optimized in different environments;

[0095] The non-linear normalization term 1 + γ·tanh(δM) uses the hyperbolic tangent function tanh(x) to smooth the influence of chemical memory, avoiding the abnormal amplification of data caused by the long-term accumulation of pollutants. In addition, γ controls the influence amplitude of chemical memory on the final result, and δ adjusts its response sensitivity;

[0096] The overall power function is non-linearly adjusted by the exponent ζ to enhance the adaptability of the model under different pollution levels, so that the adjustment amplitudes in high-pollution states and low-pollution states are different;

[0097] An abnormal pollution correction is introduced to Q, and the corrected air quality acquisition data Q is output * ; because in some cases, if sudden abnormalities occur in air pollutants, such as chemical leaks, combustion events, etc., the real-time data D1 of the sensor may be disturbed, so further correction is required:

[0098]

[0099] Where S P -S T is the difference between the pollutant fluctuation score and the data continuity score, which is used to represent the abnormality degree of the data; the exponential term is used to reduce the fluctuation influence under unstable data conditions; ξ is the pollutant abnormal correction coefficient, and the pollutant abnormal correction coefficient is used to control the suppression degree of abnormal data.

[0100] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for collecting building environment data based on dynamic air quality perception, comprising: Under normal conditions, collecting type-one data and type-two data of the building environment in real time, where type-one data includes real-time sensing data; Type-two data includes factors for air quality data missing and factors for abnormal pollutant fluctuations; Before the type-one data is output, pre-constructing a pre-influence condition based on the type-two data. If the pre-influence condition is determined to be abnormal, re-constructing a first influence condition and a second influence condition based on the type-two data, otherwise directly outputting the type-one data; It is characterized in that: The pre-influence condition includes: if the data continuity interval value is greater than a preset data continuity interval threshold, or the instantaneous pollutant change rate is greater than a preset instantaneous pollutant change rate threshold; If the first influence condition or the second influence condition is determined to be in a to-be-inspected state, defining the type-two data as abnormal, otherwise re-collecting the type-one data and the type-two data, and re-executing the determination of the pre-influence condition; if the number of re-executions is greater than two, sending out an artificial verification warning signal; When the type-two data is determined to be abnormal, obtaining chemical memory information, establishing an air perception model based on the chemical memory information and the type-two data, calculating an air perception value based on the air perception model, and realizing the output of data collection by combining the air perception value with the type-one data.

2. The method for collecting building environment data based on dynamic air quality perception according to claim 1, characterized in that: The factors for air quality data missing include the data continuity interval value and the sensor health score; The factors for abnormal pollutant fluctuations include the instantaneous pollutant change rate and the pollutant decay rate; The first influence condition includes: if the data continuity interval value is greater than a preset data continuity interval threshold and the sensor health score is less than a preset sensor health score threshold, then it is determined to be in a to-be-inspected state; The second influence condition includes: if the instantaneous pollutant change rate is greater than a preset instantaneous pollutant change rate threshold and the pollutant decay rate is less than a preset pollutant decay rate threshold, then it is determined to be in a to-be-inspected state.

3. The method for collecting building environment data based on dynamic air quality perception according to claim 2, characterized in that: Constructing a data continuity model S based on factors of missing air quality data T , the data continuity model S T evaluates the stability of data collection by describing data intervals, sensor health conditions, and environmental interference factors; the data continuity model S T is expressed as: Describing the time interval between two consecutive data points through the data interval value ΔT included in the factors for air quality data missing; describing the current health state of the sensor through the sensor health score H included in the factors for air quality data missing; In the data continuity model S T A time decay factor η is introduced to describe the degree of influence of the control time interval on data continuity; and a sensor influence coefficient ξ is introduced in the data continuity model S T to adjust the influence of health on the score; finally, S T The output value is regarded as the data continuity score.

4. The method for collecting building environment data based on dynamic air quality perception according to claim 3, characterized in that: In the process of calculating the sensor health score H: the remaining battery percentage of the sensor is described by the ratio of the current sensor battery level to the peak sensor battery level, which measures the continuous operation ability of the sensor; in addition, the error rate in the most recent N data packets is combined to form the calculation of the sensor health score H; the sensor health score H is expressed as: Where P power is the current sensor power; P max is the peak sensor power; E error is the number of errors in the most recent N data packets; E total is the total number of the most recent N data packets.

5. The method for collecting building environment data based on dynamic air quality perception according to claim 4, characterized in that: Construct a pollutant fluctuation detection model S based on the pollutant abnormal fluctuation factors included in the secondary data P , the pollutant fluctuation detection model S P is used to identify abnormal pollutant fluctuations. The construction of the pollutant fluctuation detection model S P is described as a non-linear detection model for identifying abnormalities in both sudden and gradual change environments; Pollutant Fluctuation Detection Model S P It is described as: In the pollutant fluctuation detection model S P the instantaneous change rate of pollutants is non-linearly normalized by using the arctangent function tan -1 and the exponential adjustment term e -βR decay is used to identify the long-term retention of pollutants; In the pollutant fluctuation detection model S P the change rate of pollutant concentration within a short period of time is described by the instantaneous change rate R of pollutants inst and the instantaneous change rate R of pollutants inst is defined as: In the pollutant fluctuation detection model S P the first-order decay model of the pollutant decay rate R decay is used to show how pollutants decay over time; the definition of the pollutant decay rate R decay is: where C t and C t-Δt represent the pollutant concentrations at the current moment and the previous moment respectively; Δt is the time interval; is the time normalization factor.

6. The method for collecting building environment data based on dynamic air quality perception according to claim 5, characterized in that: Constructing a chemical memory accumulation model M based on the chemical memory information to quantify the environmental residue of pollutants, calculating the influence of pollutants remaining on different surfaces, and retrieving the past pollutant concentration when the current data is abnormal; Among them, the historical pollutant concentration C i represents the historical pollution data stored in the chemical memory accumulation model M; t i is the pollutant deposition time; W i is the air flow velocity; λ i is the pollutant adsorption coefficient; γ i is the pollutant degradation coefficient; δ i is the air flow influence factor; n is the total number of pollutant types, i is the index, and each i represents a specific pollutant; k a is the adsorption rate constant; k d is the desorption rate constant; P is the pollutant concentration.

7. The method for collecting building environment data based on dynamic air quality perception according to claim 6, characterized in that: The free diffusion of pollutants in the air is described by establishing a convection-diffusion equation; the convection-diffusion equation is expressed as: where C is the pollutant concentration; is the time change rate of the pollutant concentration; represents the air velocity vector; D is the turbulent diffusion coefficient in the air; the diffusion term of the pollutant in the above equation represents how the pollutant moves from the high-concentration area to the low-concentration area through diffusion; S is the source term of the pollutant; the convection term represents the drift of the pollutant along with the air current; A boundary transport model is constructed through the pollutant transport flux per unit area to describe the interaction of pollutants on the building surface and in the ventilation system; Among which J boundary is the pollutant transport flux per unit area; k s is the deposition rate of the pollutant on the surface; C wall is the pollutant concentration in the wall or other boundary materials; represents the flow exchange of pollutants in areas such as building ventilation openings and windows; V represents the air control volume in the building environment; Final air flow and diffusion model D F It is jointly constructed by combining the convective-diffusion equation with the boundary transport model:

8. The method for collecting building environment data based on dynamic air quality perception according to claim 7, wherein: The final air perception value APV is expressed as: where the air perception value APV is used for final data output; μ, ν, ω, ρ, θ, κ, σ, λ are adjustment parameters, and the adjustment parameters are used to control the influence weights of different factors; In the above formula, the exponential term e ωM is used to highlight the influence of chemical memory; the power term is used to ensure the importance of data quality and pollution detection; the non-linear expansion term cosh(σM) introduces non-linear adjustment by applying the hyperbolic cosine function cosh(x), and is used to smooth the long-term impact of pollutants; Normalization term To balance the aerodynamic effects.

9. The method for collecting building environment data based on dynamic air quality perception according to claim 8, wherein: The final data collection value Q is jointly determined by a class of data D1 and the air perception value APV: where tanh(δM) is the hyperbolic tangent function, and in the above formula, the hyperbolic tangent function is used to smooth the influence of chemical memory; α, β, γ, δ, ζ are model adjustment parameters; Introduce an abnormal pollution correction to Q and output the corrected air quality acquisition data Q * ; where S P -S T is the difference between the pollutant fluctuation score and the data continuity score; the exponential term is used to reduce the fluctuation impact under unstable data conditions; ξ is the pollutant anomaly correction coefficient.

Images