Air conditioner monitoring system based on intelligent olfaction chip

Abstract

The invention discloses an air conditioner monitoring system based on an intelligent olfaction chip, and relates to the technical field of air conditioner monitoring. A sensor network is adopted to collect data in a plurality of areas of an underground subway station to form a wide-coverage area data collection network; multi-source data from each sensor is mapped to a hidden space through an LSTM auto-encoder, original input is reconstructed by means of a decoder, and a reconstruction error is obtained through calculation; historical data statistics is utilized to construct a dynamic baseline reflecting the normal operation state of the equipment, so that a time-varying threshold judgment mechanism based on historical statistical characteristics is realized; by comparing a reconstruction error with a dynamic threshold value, a simple judgment function is adopted to identify local subtle anomalies, and the defects of a traditional fixed threshold value method in monitoring local chemical component fluctuation in a closed environment are overcome; the anomaly detection module can reflect real-time data changes on a plurality of sampling points at the same time, and a local anomaly state is clearly expressed as an anomaly indication signal.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioner monitoring, and particularly to an air conditioner monitoring system based on an intelligent olfactory chip. Background Art

[0002] Traditional air conditioner monitoring solutions rely on sensors such as temperature and humidity sensors, CO 2 and VOC sensors to collect data centrally and judge abnormal states based on fixed thresholds.

[0003] However, in practical applications, it is found that this mode is difficult to meet the monitoring requirements of areas with complex environments. Firstly, the device layout points are limited and cannot cover every corner of the space comprehensively, resulting in the omission of abnormal air changes in some areas. At the same time, there is a certain time delay in centralized data processing, and the early warning response is not fast enough in case of emergencies.

[0004] In the subway, the air circulation conditions are poor, and local harmful substances are more likely to accumulate. Conventional sensors are not sensitive enough to trace gases such as low-concentration ozone and NOx, and it is difficult to capture their subtle fluctuations in time. To make up for this defect, some solutions add sensors at key positions and adopt multi-source data fusion and dynamic threshold strategies, but the effect is limited, and it is still difficult to completely solve the problems of insufficient local monitoring and response lag. The existing methods lack flexibility in fixed-parameter early warning and cannot adjust the monitoring standards in real time according to different working conditions, resulting in difficulty in timely intervention in some emergencies and potential safety hazards. Generally speaking, traditional monitoring technologies have blind spots and delays in data collection, processing, and feedback. Therefore, the present invention provides an air conditioner monitoring system based on an intelligent olfactory chip to solve such problems. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides an air conditioner monitoring system based on an intelligent olfactory chip to solve the problems of blind spots and delays in data collection, processing, and feedback of traditional monitoring technologies. In the subway station, multiple plasma air conditioners work together, the air circulation is restricted, the local ozone or reactive oxygen concentration is high, but the distribution of monitoring points is uneven, resulting in delayed emergency early warning.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] An embodiment of the present invention provides an air conditioner monitoring system based on an intelligent olfactory chip, which includes

[0009] a data collection unit, the data collection unit is provided with a temperature and humidity sensor, a CO 2 sensor, a VOC sensor, and an intelligent olfactory chip, and the intelligent olfactory chip collects trace chemical components released during the operation of the plasma air conditioner equipment;

[0010] A data fusion module, electrically connected to the data acquisition unit. The data fusion module uses a deep time series model to fuse data signals from various sensors and establish a normal operation baseline;

[0011] An anomaly detection module, electrically connected to the data fusion module. The anomaly detection module performs dynamic threshold determination on the real-time collected data signals based on the normal operation baseline;

[0012] An edge computing platform, electrically connected to the anomaly detection module and the air conditioner control unit. The edge computing platform performs low-latency processing on real-time data signals;

[0013] An air conditioner control unit, electrically connected to the edge computing platform. The air conditioner control unit adjusts the operating parameters of the plasma air conditioner device according to the output of the anomaly detection module.

[0014] As a preferred solution of the air conditioner monitoring system based on an intelligent olfactory chip according to the present invention, wherein: the data acquisition unit is arranged in multiple areas of an underground subway station, and the distribution points of each sensor and the intelligent olfactory chip are determined according to the air sampling requirements in the area, forming a regional data acquisition network.

[0015] As a preferred solution of the air conditioner monitoring system based on an intelligent olfactory chip according to the present invention, wherein: the deep time series model is a model based on an LSTM autoencoder, which is used to establish a normal operation baseline for historical data, and the normal operation baseline provides a reference for dynamic threshold determination.

[0016] As a preferred solution of the air conditioner monitoring system based on an intelligent olfactory chip according to the present invention, wherein: the way that the data fusion module uses a deep time series model to fuse data signals from various sensors and establish a normal operation baseline is as follows:

[0017] Let the multi-source data input sequence be x t ∈R n , t = 1, 2, …, T, where x t represents the sensor data vector at time t, R represents the set of real numbers, n represents the number of sensors, including temperature and humidity, CO 2 , VOC and intelligent olfactory chip data, and t represents the total number of time series moments;

[0018] In the encoder stage, calculate the hidden state, and the formula is:

[0019] h t = σ(W e x t + U e h t-1 + b e ),

[0020] Among them, h t ∈R m represents the hidden state at time t, and h t-1 represents the hidden state at the previous time. W e ∈R m×n represents the input weight matrix, and U e ∈R m×m represents the recurrent weight matrix, b e ∈R m represents the bias vector, σ(·) represents the activation function, and m represents the dimension of the hidden layer;

[0021] The decoder stage reconstructs the input, which is expressed as:

[0022]

[0023] Among them, represents the reconstructed output vector, and W d ∈R n×m represents the decoding weight matrix, and b d ∈R n represents the decoding bias vector, and φ(·) represents the decoder activation function;

[0024] The Euclidean norm is used to calculate the reconstruction error, and the formula is:

[0025]

[0026] Among them, e t represents the reconstruction error at time t, and |·| 2 represents the 2-norm.

[0027] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip described in the present invention, where: the data fusion module uses a deep time series model to fuse the data signals from each sensor, and the method for establishing the normal operation baseline further includes:

[0028] Establish a normal operation baseline based on historical data, and calculate the error mean and standard deviation. The formula is:

[0029]

[0030] Among them, μ e represents the reconstruction error mean, and σ e represents the reconstruction error standard deviation;

[0031] Construct a dynamic threshold θ t :

[0032] θ t = μ e + ασe ,

[0033] where θ t represents the dynamic threshold at time t, and α is a constant factor.

[0034] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip according to the present invention, wherein: the step of the anomaly detection module performing dynamic threshold determination on the real-time collected data signal based on the normal operation baseline is as follows:

[0035] Based on the fusion data and the normal baseline, an anomaly detection is performed using a decision function:

[0036] If e t > θ t , then a t = 1,

[0037] If e t ≤ θ t , then a t = 0,

[0038] where a t represents the anomaly indication at time t, with a value of 1 indicating a local anomaly and a value of 0 indicating normal, e t is the reconstruction error, and θ t is the dynamic threshold, which is calculated from the baseline.

[0039] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip according to the present invention, wherein: the edge computing platform connects each module using a low-latency data bus, and the low-latency data bus is used for the transmission of data signals between modules.

[0040] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip according to the present invention, wherein: the air conditioner control unit adjusts the operating parameters of the plasma air conditioner device including the discharge intensity and the startup period, and the adjustment of the discharge intensity and the startup period is based on the dynamic determination result output by the anomaly detection module.

[0041] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip according to the present invention, wherein: the step of the air conditioner control unit adjusting the operating parameters of the plasma air conditioner device including the discharge intensity and the startup period based on the dynamic determination result output by the anomaly detection module is as follows:

[0042] The feedback adjustment adopts a linear correction method to adjust the parameters of the plasma air conditioner device, and the adjustment method is:

[0043] d new = d old - βa t ,

[0044] T s,new = T s,old + γa t ,

[0045] Wherein, d new represents the updated discharge intensity, d old represents the current discharge intensity, β is the discharge intensity adjustment factor, T s,new represents the updated startup cycle, T s,old represents the current startup cycle, γ is the startup cycle adjustment factor, a t is an anomaly indication variable.

[0046] As a preferred solution of the air conditioner monitoring system based on the intelligent olfactory chip according to the present invention, wherein: in the air conditioner control unit, the determination method of the discharge intensity and the startup cycle adjustment factor is:

[0047] If e t > θ t , then

[0048] If e t ≤ θ t , then β = 0;

[0049] If e t > θ t , then

[0050] If e t ≤ θ t , then γ = 0,

[0051] Wherein, k d represents the discharge intensity sensitivity coefficient, k T represents the startup cycle sensitivity coefficient, e t represents the reconstruction error at time t, θ t represents the dynamic threshold at time t; when the reconstruction error exceeds the dynamic threshold, that is, there is an anomaly, the adjustment factor is calculated based on the ratio of the error to the threshold. If the error does not exceed the threshold, no adjustment is made.

[0052] The beneficial effects of the present invention are: The present invention uses a sensor network to collect data in multiple areas of an underground subway station, combines temperature and humidity, CO 2 , VOC sensors with an intelligent olfactory chip to form a wide-coverage area data collection network; through the LSTM autoencoder, maps multi-source data from each sensor to a hidden space, and reconstructs the original input with the help of a decoder to calculate the reconstruction error; uses historical data statistics to construct a dynamic baseline reflecting the normal operation state of the equipment, thus realizing a time-varying threshold determination mechanism based on historical statistical characteristics.

[0053] By comparing the reconstruction error with the dynamic threshold, the present invention uses a simple decision function to identify local subtle anomalies, making up for the deficiencies of the traditional fixed-threshold method in monitoring local chemical composition fluctuations in a closed environment; the anomaly detection module can simultaneously reflect real-time data changes at multiple sampling points, and clearly express the local abnormal state as an anomaly indication signal.

[0054] The present invention adjusts the operating parameters of the plasma air-conditioning equipment according to the anomaly indication signal. Specifically, when an abnormal state is detected, the discharge intensity of the equipment decreases according to a certain linear relationship, and the startup period is correspondingly extended, establishing a closed-loop control between the abnormal state and the equipment parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0056] Figure 1 It is a schematic framework diagram of the air-conditioning monitoring system based on the intelligent olfactory chip of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention in conjunction with the drawings of the specification.

[0058] Many specific details are set forth in the following description in order to fully understand the present invention, but the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0059] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that excludes other embodiments.

[0060] Embodiment 1, referring to Figure 1 , this embodiment provides an air-conditioning monitoring system based on an intelligent olfactory chip, including:

[0061] A data acquisition unit, which is provided with a temperature and humidity sensor, CO 2Sensors, VOC sensors, and intelligent olfactory chips. The intelligent olfactory chips collect trace chemical components released during the operation of the plasma air conditioner equipment;

[0062] The data acquisition unit is deployed in multiple areas of the underground subway station. The distribution points of each sensor and intelligent olfactory chip are determined according to the air sampling requirements in the area, forming a regional data acquisition network;

[0063] The data fusion module is electrically connected to the data acquisition unit. The data fusion module uses a deep time series model to fuse the data signals from each sensor and establish a normal operation baseline;

[0064] The deep time series model is a model based on the LSTM autoencoder, which is used to establish a normal operation baseline for historical data. The normal operation baseline provides a reference for dynamic threshold determination;

[0065] The way that the data fusion module uses a deep time series model to fuse the data signals from each sensor and establish a normal operation baseline is as follows:

[0066] Let the multi-source data input sequence be x t ∈R n , t = 1, 2, …, T, where x t represents the sensor data vector at time t, R represents the set of real numbers, n represents the number of sensors, including temperature and humidity, CO 2 , VOC, and intelligent olfactory chip data, and T represents the total number of time points in the sequence;

[0067] In the encoder stage, the hidden state is calculated using the formula:

[0068] h t =σ(W e x t +U e h t-1 +b e ),

[0069] where h t ∈R m represents the hidden state at time t, h t-1 represents the hidden state at the previous time point, W e ∈R m×n represents the input weight matrix, U e ∈R m×m represents the recurrent weight matrix, b e ∈R m represents the bias vector, σ(·) represents the activation function, and m represents the dimension of the hidden layer;

[0070] In the decoder stage, the input is reconstructed, expressed as:

[0071]

[0072] Among them, represents the reconstructed output vector, and W d ∈R n×m represents the decoding weight matrix, and b d ∈R n represents the decoding bias vector, and φ(·) represents the decoder activation function;

[0073] The Euclidean norm is used to calculate the reconstruction error, and the formula is:

[0074]

[0075] Among them, e t represents the reconstruction error at time t, and |·| 2 represents the 2-norm;

[0076] The data fusion module uses a deep time series model to fuse the data signals from each sensor, and the methods for establishing the normal operation baseline also include:

[0077] Based on historical data, establish a normal operation baseline, and calculate the error mean and standard deviation. The formula is:

[0078]

[0079] Among them, μ e represents the reconstruction error mean, and σ e represents the reconstruction error standard deviation;

[0080] Construct a dynamic threshold θ t :

[0081] θ t = μ e + ασ e ,

[0082] Among them, θ t represents the dynamic threshold at time t, and α is a constant factor;

[0083] Specifically, here an LSTM autoencoder is used to extract the time series features of multi-source sensor data. The input data includes the data from the intelligent olfactory chip, which is mapped to the hidden state through the encoder, and then the original data is reconstructed by the decoder. The reconstruction error between the two is calculated. Using historical data, the mean and standard deviation of the reconstruction error are statistically calculated to form the normal operation baseline, and a dynamic threshold is generated using the constant factor to convert the original data into a unified index;

[0084] The anomaly detection module is electrically connected to the data fusion module. The anomaly detection module makes a dynamic threshold determination on the real-time collected data signal based on the normal operation baseline;

[0085] The steps for the anomaly detection module to perform dynamic threshold determination on the real-time collected data signals based on the normal operation baseline are as follows:

[0086] Based on the fused data and the normal baseline, an anomaly detection is performed using a decision function:

[0087] If e t > θ t , then a t = 1,

[0088] If e t ≤ θ t , then a t = 0,

[0089] where a t represents the anomaly indication at time t, with a value of 1 indicating a local anomaly and a value of 0 indicating normal. e t is the reconstruction error, and θ t is the dynamic threshold, which is calculated from the baseline;

[0090] Specifically, here, by comparing the real-time reconstruction error with the dynamic threshold, local subtle anomalies are identified. The anomaly determination formula is threshold comparison. If the current error is greater than the dynamic threshold, the current state is marked as abnormal; otherwise, it is marked as normal. This method relies on the previously constructed normal baseline. The dynamic threshold can reflect the statistical characteristics of historical data, and the determination process has time continuity and numerical responsiveness, making it suitable for real-time monitoring of multi-point data collection in a closed environment;

[0091] The edge computing platform is electrically connected to the anomaly detection module and the air-conditioning control unit, and the edge computing platform performs low-latency processing on the real-time data signals;

[0092] The edge computing platform connects each module using a low-latency data bus, and the low-latency data bus is used for the transmission of data signals between modules;

[0093] The air-conditioning control unit is electrically connected to the edge computing platform, and the air-conditioning control unit adjusts the operating parameters of the plasma air-conditioning equipment according to the output of the anomaly detection module;

[0094] The operating parameters of the plasma air-conditioning equipment adjusted by the air-conditioning control unit include the discharge intensity and the startup cycle, and the adjustment of the discharge intensity and the startup cycle is based on the dynamic determination result output by the anomaly detection module;

[0095] The steps for the air-conditioning control unit to adjust the operating parameters of the plasma air-conditioning equipment, including the discharge intensity and the startup cycle, based on the dynamic determination result output by the anomaly detection module are as follows:

[0096] The feedback adjustment adopts a linear correction method to adjust the parameters of the plasma air-conditioning equipment, and the adjustment method is as follows:

[0097] d new = d old - βa t ,

[0098] T s,new = T s,old + γa t ,

[0099] where d new represents the updated discharge intensity, d old represents the current discharge intensity, β is the discharge intensity adjustment factor, T s,new represents the updated startup cycle, T s,old represents the current startup cycle, γ is the startup cycle adjustment factor, and a t is the anomaly indication variable;

[0100] Specifically, the anomaly detection result is fed back to the adjustment of the plasma air conditioner device parameters here. Using the feedback adjustment formula, when an abnormal state is detected (a t = 1), the discharge intensity decreases by a fixed factor, and the startup cycle increases by a fixed factor. When no anomaly is detected (a t = 0), the device parameters remain unchanged. This linear feedback relationship directly maps the anomaly signal to a device control command to achieve closed-loop regulation;

[0101] In the air conditioner control unit, the determination methods of the discharge intensity and startup cycle adjustment factors are as follows:

[0102] If e t > θ t , then

[0103] If e t ≤ θ t , then β = 0;

[0104] If e t > θ t , then

[0105] If e t ≤ θ t , then γ = 0,

[0106] where k d represents the discharge intensity sensitivity coefficient, k T represents the startup cycle sensitivity coefficient, e t represents the reconstruction error at time t, and θ t represents the dynamic threshold at time t; when the reconstruction error exceeds the dynamic threshold, that is, when there is an anomaly, the adjustment factor is calculated based on the ratio of the error to the threshold. If the error does not exceed the threshold, no adjustment is made;

[0107] Specifically, the discharge intensity adjustment factor and the startup cycle adjustment factor respectively reflect the influence of the degree of abnormality on the operating parameters of the equipment. By using the factors calculated in the formula, the abnormal data is converted into specific values to achieve linear adjustment of the parameters of the plasma air-conditioning equipment.

[0108] In summary, the present invention uses a sensor network to collect data in multiple areas of an underground subway station, combines temperature and humidity, CO 2 , VOC sensors with an intelligent olfactory chip to form a wide-coverage area data collection network; through the LSTM autoencoder, multi-source data from each sensor is mapped to a hidden space, and the original input is reconstructed with the help of a decoder, and the reconstruction error is calculated; by using historical data statistics, a dynamic baseline reflecting the normal operating state of the equipment is constructed, thus realizing a time-varying threshold determination mechanism based on historical statistical characteristics.

[0109] By comparing the reconstruction error with the dynamic threshold, the present invention uses a simple decision function to identify local subtle abnormalities, making up for the deficiencies of the traditional fixed threshold method in monitoring local chemical composition fluctuations in a closed environment; the abnormal detection module can reflect real-time data changes at multiple sampling points and clearly express the local abnormal state as an abnormal indication signal.

[0110] According to the abnormal indication signal, the present invention adjusts the operating parameters of the plasma air-conditioning equipment. Specifically, when an abnormal state is detected, the discharge intensity of the equipment decreases according to a certain linear relationship, and the startup cycle is correspondingly extended, so as to establish a closed-loop control between the abnormal state and the equipment parameters.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An air conditioning monitoring system based on an intelligent olfactory chip, characterized in that: include, A data acquisition unit, wherein the data acquisition unit is provided with a temperature and humidity sensor, a CO2 sensor, a VOC sensor and an intelligent olfactory chip, wherein the intelligent olfactory chip collects trace chemical components released during the operation of the plasma air conditioning equipment; A data fusion module is electrically connected to the data acquisition unit, and the data fusion module uses a deep time series model to fuse the data signals from each sensor and establish a normal operation baseline; an abnormality detection module, electrically connected to the data fusion module, and wherein the abnormality detection module performs dynamic threshold determination on the real-time collected data signal according to the normal operation baseline; An edge computing platform, electrically connected to the anomaly detection module and the air conditioning control unit, the edge computing platform performs low-latency processing on real-time data signals; An air conditioning control unit is electrically connected to the edge computing platform, and the air conditioning control unit adjusts the operating parameters of the plasma air conditioning equipment according to the output of the abnormality detection module.

2. The air conditioning monitoring system based on the intelligent olfactory chip according to claim 1, characterized in that: The data acquisition units are deployed in multiple areas of underground subway stations, and the locations of each sensor and the intelligent olfactory chip are determined according to the air sampling requirements in the area to form a regional data acquisition network.

3. The air conditioning monitoring system based on the intelligent olfactory chip as claimed in claim 2, characterized in that: The deep time series model is a model based on the LSTM autoencoder, which is used to establish a normal operation baseline for historical data, and the normal operation baseline provides a reference for dynamic threshold determination.

4. The air conditioning monitoring system based on the intelligent olfactory chip as claimed in claim 3, characterized in that: The data fusion module uses a deep time series model to fuse the data signals from each sensor and establish a normal operation baseline in the following manner: Suppose the multi-source data input sequence is x t ∈R n ,t=1,2,…,T, where x t represents the sensor data vector at time t, R represents a real number set, n represents the number of sensors, including temperature and humidity, CO2, VOC and smart olfactory chip data, and t represents the total number of sequence moments; The encoder stage calculates the hidden state, the formula is: h t =σ(W e x t +U e h t-1 +b e ), Among them, h t ∈R m represents the hidden state at time t, h t-1 represents the hidden state at the previous moment, W e ∈R m×n represents the input weight matrix, U e ∈R m×m represents the recursive weight matrix, b e ∈R m represents the bias vector, σ(·) represents the activation function, and m represents the hidden layer dimension; The decoder stage reconstructs the input, expressed as: in, Represents the reconstructed output vector, W d ∈R n×m represents the decoding weight matrix, b d ∈R n represents the decoding bias vector, φ(·) represents the decoder activation function; The Euclidean norm is used to calculate the reconstruction error, and the formula is: Among them, e t represents the reconstruction error at time t, and |·|2 represents the 2-norm.

5. The air conditioning monitoring system based on the intelligent olfactory chip as claimed in claim 4, characterized in that: The data fusion module uses a deep time series model to fuse the data signals from each sensor, and the method of establishing a normal operation baseline also includes: Establish a normal operation baseline based on historical data and calculate the error mean and standard deviation. The formula is: Among them, μ e represents the mean reconstruction error, σ e represents the standard deviation of the reconstruction error; Construct dynamic threshold θ t : i t =μ e +as e , Among them, θ t represents the dynamic threshold at time t, and α is a constant factor.

6. The air conditioning monitoring system based on the intelligent olfactory chip according to claim 5, characterized in that: The step of the abnormality detection module performing dynamic threshold determination on the real-time collected data signal according to the normal operation baseline is: Based on the fused data and the normal baseline, anomaly detection is performed using the decision function: If e t >θ t , then a t =1, If e t ≤θ t , then a t =0, Among them, a t Indicates the abnormal indication at time t. A value of 1 indicates a local abnormality, and a value of 0 indicates normality. t is the reconstruction error, θ t is the dynamic threshold, calculated from the baseline.

7. The air conditioning monitoring system based on the intelligent olfactory chip according to claim 6, characterized in that: The edge computing platform uses a low-latency data bus to connect the modules, and the low-latency data bus is used to transmit data signals between the modules.

8. The air conditioning monitoring system based on the intelligent olfactory chip as claimed in claim 7, characterized in that: The air conditioning control unit adjusts the operating parameters of the plasma air conditioning equipment including the discharge intensity and the start-up cycle, and the adjustment of the discharge intensity and the start-up cycle is based on the dynamic determination result output by the abnormality detection module.

9. The air conditioning monitoring system based on the intelligent olfactory chip according to claim 8, characterized in that: The air conditioning control unit adjusts the operating parameters of the plasma air conditioning device including the discharge intensity and the start-up cycle. The steps of adjusting the discharge intensity and the start-up cycle according to the dynamic determination result output by the abnormality detection module are as follows: Feedback regulation uses linear correction method to adjust the parameters of plasma air conditioning equipment. The adjustment method is: d new =d old -βa t , T s,new =T s,old +γa t , Among them, d new represents the updated discharge intensity, d old represents the current discharge intensity, β is the discharge intensity adjustment factor, T s,new Represents the startup period after the update, T s,old represents the current startup cycle, γ is the startup cycle adjustment factor, a t An exception indicator variable.

10. The air conditioning monitoring system based on the intelligent olfactory chip according to claim 9, characterized in that: In the air conditioning control unit, the discharge intensity and the start-up cycle adjustment factor are determined as follows: If e t >θ t ,but If e t ≤θ t , then β=0; If e t >θ t ,but If e t ≤θ t , then γ=0, Among them, k d Indicates the discharge intensity sensitivity coefficient, k T represents the startup cycle sensitivity coefficient, e t represents the reconstruction error at time t, θ t Represents the dynamic threshold at time t.

Images