Near infrared spectrum real-time monitoring method and system

Abstract

The invention provides a near infrared spectrum real-time monitoring method and system, and relates to the technical field of spectral analysis. The method comprises the following steps: analyzing electromagnetic interference caused to near infrared spectrum equipment when high-voltage equipment works, and training to obtain a first electromagnetic interference prediction model according to a first power utilization characteristic data set and a first electromagnetic interference data set in a first spectrum acquisition environment; first high-risk interference information is obtained through prediction based on the power consumption characteristic data, second high-risk interference information is determined based on the electromagnetic interference data, and then target high-risk interference information is determined based on the connection relation between different electromagnetic interference types; and finally, adjusting acquisition parameters of the near infrared spectrum equipment and obtaining first aviation kerosene quality analysis information. According to the technical scheme of the invention, the interference of electromagnetic interference on near infrared spectrum recognition can be effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of spectral analysis technology, and in particular to a method and system for real-time monitoring of near-infrared spectroscopy. Background Technology

[0002] Near-infrared spectroscopy (NIRS) offers advantages such as speed and non-destructive testing in aviation kerosene quality inspection, and is widely used in petroleum refining and aviation fuel industries. However, in industrial environments with dense high-pressure equipment, such as refinery storage areas and airport oil depots, electromagnetic interference severely restricts the accuracy and reliability of testing. Electromagnetic interference generated during the operation of high-pressure equipment can intrude into the NIRS detection system through conduction and radiation, leading to baseline drift, blurred characteristic peaks, and reduced signal-to-noise ratio in spectral data. This significantly increases the prediction error of key quality indicators such as moisture content and sulfur content. Existing anti-interference technologies mostly employ static filtering or simple thresholding, which cannot effectively address the dynamic changes in interference types and their temporal continuity characteristics, resulting in unstable and poor repeatability of detection results in complex electromagnetic environments. Therefore, there is an urgent need for a NIRS detection method that can accurately identify interference types and adaptively suppress interference based on its continuity to improve the accuracy and industrial applicability of aviation kerosene quality inspection. Summary of the Invention

[0003] The purpose of this invention is to provide a near-infrared spectroscopy real-time monitoring method and system, which can improve the accuracy and industrial applicability of aviation kerosene quality detection.

[0004] This application proposes a real-time near-infrared spectroscopy monitoring method, which includes: S1: Determine the first set of power consumption characteristic data and the first set of electromagnetic interference data in the first spectrum acquisition environment, and train to obtain the first electromagnetic interference prediction model; Wherein, the first electricity consumption characteristic data set and the first electromagnetic interference data set are both data sets within a first preset time interval; S2: Based on the first power consumption characteristic data and the first electromagnetic interference data of the first spectral acquisition environment within the second preset time interval, the first high-risk interference information and the second high-risk interference information are determined respectively. S3: Based on the interference type succession relationship matrix and the second high-risk interference information, determine the target high-risk interference information from the first high-risk interference information; S4: Obtain the first near-infrared spectral acquisition parameters based on the target high-risk interference information, and determine the first aviation kerosene quality analysis information.

[0005] Preferably, S1 includes: S11: In the first spectral acquisition environment, acquire a first set of power consumption characteristic data of at least one first high-voltage device within a first preset time interval; S12: Analyze the electromagnetic interference data within the first preset time interval to determine the first electromagnetic interference data set; S13: A first electromagnetic interference prediction model is obtained by training based on the first set of electricity consumption characteristic data and the first set of electromagnetic interference data.

[0006] Preferably, the first electricity consumption characteristic data set includes: Device ID, timestamp, usage intensity, vibration characteristics, temperature characteristics, harmonic characteristics.

[0007] Preferably, S2 includes: S21: Obtain first power consumption characteristic data of at least one of the first high-voltage devices within a second preset time interval; S22: Input the first power consumption characteristic data into the first electromagnetic interference prediction model to determine the first high-risk interference information; S23: Based on the first electromagnetic interference data within the second preset time interval, determine the second high-risk interference information.

[0008] Preferably, the first high-risk interference information includes interference type, interference intensity, and interference duration.

[0009] Preferably, S3 includes: S31: By performing a first transfer analysis on historical electromagnetic interference data, the interference type succession relationship matrix is ​​determined; S32: Based on the second high-risk interference type and the interference type succession relationship matrix, determine at least one target high-risk interference type from the first high-risk interference information; S33: Based on the second high-risk interference information, determine the target high-risk interference information corresponding to the target high-risk interference type.

[0010] Preferably, S32 includes: S321: Identify at least one first high-risk interference type from the second high-risk interference information; S322: Based on the interference type succession relationship matrix, obtain the second high-risk interference type corresponding to each of the first high-risk interference types; S323: Based on the second high-risk interference type and the first high-risk interference information, determine the target high-risk interference type.

[0011] Preferably, S4 includes: S41: Input the target high-risk interference information into the acquisition parameter determination model to determine the first near-infrared spectral acquisition parameters; S42: Obtain first near-infrared spectral data based on the first near-infrared spectral acquisition parameters, and determine first aviation kerosene quality analysis information based on the first near-infrared spectral data.

[0012] This application also proposes a near-infrared spectroscopy real-time monitoring system for implementing the aforementioned near-infrared spectroscopy real-time monitoring method.

[0013] This application proposes a real-time near-infrared spectroscopy monitoring method and system, relating to the field of spectral analysis technology. It analyzes the electromagnetic interference caused to near-infrared spectroscopy equipment during high-voltage operation. Based on a first set of power consumption characteristic data and a first set of electromagnetic interference data under a first spectral acquisition environment, a first electromagnetic interference prediction model is trained. Then, based on the power consumption characteristic data, a first high-risk interference information is predicted, and based on the electromagnetic interference data, a second high-risk interference information is determined. Next, based on the continuity relationship between different electromagnetic interference types, the target high-risk interference information is determined. Finally, the acquisition parameters of the near-infrared spectroscopy equipment are adjusted to obtain first aviation kerosene quality analysis information. The technical solution of this application can effectively reduce the interference caused by electromagnetic interference to near-infrared spectral identification. Attached Figure Description

[0014] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0015] Figure 1 This is an execution flowchart of a near-infrared spectroscopy real-time monitoring method in this invention.

[0016] Figure 2 This is the process for determining high-risk interference information in this application. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The illustrative embodiments and descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0019] The following is a detailed description of a near-infrared spectroscopy real-time monitoring method and system of the present invention.

[0020] This embodiment proposes a real-time near-infrared spectroscopy monitoring method, the specific process of which is as follows: Figure 1 As shown.

[0021] S1: Determine the first set of power consumption characteristic data and the first set of electromagnetic interference data in the first spectral acquisition environment, and train to obtain the first electromagnetic interference prediction model.

[0022] In the process of analyzing aviation kerosene quality based on near-infrared spectral data, electromagnetic interference caused by the use of various high-voltage equipment can greatly affect the accuracy of the identification results.

[0023] The usage and electromagnetic interference of high-voltage equipment are greatly affected by electricity consumption patterns. Therefore, in this step, it is necessary to analyze and sort out the correlation between electricity consumption patterns and electromagnetic interference characteristics based on historical electricity consumption data of a specific area, in order to prepare for subsequent analysis and processing.

[0024] S1 includes the following sub-steps: S11: In the first spectral acquisition environment, acquire a first set of power consumption characteristic data of at least one first high-voltage device within a first preset time interval.

[0025] Since the location for collecting near-infrared spectral data of aviation kerosene is usually relatively fixed, the first spectral acquisition environment in which the acquisition equipment is located is also relatively fixed. The first spectral acquisition environment refers to the operating pattern information of the surrounding high-voltage equipment, such as which times of the week or day the first high-voltage equipment will be in a busy operating state, and which times the first high-voltage equipment will be in an idle operating state, etc.

[0026] The electrical characteristics include: device ID, timestamp, usage intensity, vibration characteristics, temperature characteristics, and harmonic characteristics. Specifically, the electrical characteristics of all the first high-voltage devices in the first spectral acquisition environment need to be collected.

[0027] The specific electricity data collection process includes: The first high-voltage equipment's electricity consumption data is collected through a smart meter system within a first preset time interval. To ensure high coverage of the collected data, the first preset time interval is preferably 36 months, during which electricity consumption is recorded every 15 minutes.

[0028] Time series analysis is used to preprocess the data. The preferred processing model for preprocessing is the ARIMA model. Outliers are eliminated using the 3σ principle, and missing values ​​are filled using linear interpolation. For specific processing methods, please refer to existing technologies, which will not be elaborated here.

[0029] Extracting electricity consumption characteristics: daily cycle intensity index, such as morning peak 8:00-10:00 and evening peak 17:00-19:00; weekly cycle intensity index, such as weekdays vs. weekends; monthly cycle intensity index, such as high at the beginning of the month and low at the end of the month. The specific electricity intensity index can be determined by the ratio of real-time electricity consumption to the historical average electricity consumption for the same period.

[0030] In the output data, each of the first high-voltage devices corresponds to a multi-dimensional data vector. For example, for a high-voltage transformer, its multi-dimensional data vector includes multi-dimensional data vectors corresponding to multiple time points. For example, for time point 1, its multi-dimensional time vector includes: device ID, timestamp, electricity intensity index, vibration characteristics, temperature characteristics, harmonic characteristics, etc. Therefore, for each high-voltage device, at each time point within its first preset time interval, there is a corresponding multi-dimensional data vector.

[0031] The first electricity consumption characteristic data set can then be formed by combining the multiple multidimensional data vectors corresponding to all the first high-voltage equipment.

[0032] S12: Analyze the electromagnetic interference data within the first preset time interval to determine the first electromagnetic interference data set.

[0033] The first set of electricity consumption characteristic data determined in S11 can be used to predict the types of electromagnetic interference with a high probability of occurrence within a specific time interval in the future. To further improve the accuracy of electromagnetic interference type prediction, it is also necessary to determine time-sensitive electromagnetic interference types using real-time measured electromagnetic interference data. The final electromagnetic interference type is then determined by comparing the correlation between the measured electromagnetic interference type and the predicted electromagnetic interference type.

[0034] For the above purposes, this step requires determining the historical electromagnetic interference data within the first preset time interval from historical data.

[0035] The specific method for determining the historical electromagnetic interference data is as follows: Electromagnetic interference data is collected by an electromagnetic interference sensor system within the first preset time interval. The interference intensity and type are recorded every 15 minutes. Preferably, an electromagnetic interference sensor of model HMC5883L can be used.

[0036] Wavelet transform is used to extract interference features. For specific methods of interference feature analysis and extraction, please refer to existing technologies, which will not be elaborated here.

[0037] Electromagnetic interference types can include: power frequency interference, transient interference, high-frequency noise interference, and drift interference. Electromagnetic interference parameters can include: interference intensity and interference duration.

[0038] For each type of electromagnetic interference, its data type has corresponding characteristics, thus allowing the extraction of the corresponding electromagnetic interference data.

[0039] For example, regarding power frequency interference: 1. Frequency characteristics Fixed frequency: The power frequency interference frequency is the same as the power grid supply frequency, which is 50Hz in China and 60Hz in the United States; Periodicity: The period T = 20 ms (50 Hz) is a regular sine wave.

[0040] 2. Waveform characteristics Typical waveform: It appears as a 50Hz / 60Hz sine wave superimposed on the signal; Waveform distortion: The actual waveform may be distorted due to the influence of nonlinear loads (such as switching power supplies and fluorescent lamps).

[0041] 3. Signal characteristics Periodicity: The interference signal exhibits regular changes; Stability: Power frequency interference signals typically do not disappear suddenly or fluctuate significantly.

[0042] After determining the characteristics of electromagnetic interference data for a specific type of electromagnetic interference, relevant parameters for characterizing the interference type, intensity, and duration can be extracted. The electromagnetic interference data can then be correlated with time points to form a first electromagnetic interference data set. This first electromagnetic interference data set includes multiple sets of first electromagnetic interference data at multiple time points.

[0043] S13: A first electromagnetic interference prediction model is obtained by training based on the first set of electricity consumption characteristic data and the first set of electromagnetic interference data.

[0044] In S11 and S12, the first set of electricity consumption characteristic data and the first set of electromagnetic interference data within the first preset time interval have been obtained based on historical data analysis. Since the electromagnetic interference type needs to be predicted based on the electricity consumption characteristic data at a specific time point in subsequent steps, the first electromagnetic interference prediction model needs to be trained based on the first set of electricity consumption characteristic data and the first set of electromagnetic interference data in this step to prepare for subsequent steps.

[0045] The specific training process of the first electromagnetic interference prediction model is as follows: Preferably, an association matrix between electricity consumption characteristics and interference characteristics is constructed, and the probability of different interference types under each combination of electricity consumption characteristics in the first electricity consumption characteristic data set is calculated. For example, when the electricity consumption intensity is ≥1.5 during the peak hours (8:00-10:00), the power frequency interference index = 0.82, and the transient interference index = 0.15. The electromagnetic interference index is obtained by performing specified operations on the first electromagnetic interference data and is positively correlated with the interference intensity and duration. A random forest algorithm is used to optimize the association matrix to ensure accurate association probabilities.

[0046] In another embodiment, the first electromagnetic interference prediction model can be obtained by training a convolutional neural network. The first electricity consumption feature data set and the first electromagnetic interference dataset are combined as training sample data, the first electricity consumption feature data set is used as input data, and the first electromagnetic interference dataset is used as output data. The input and output data in the same set of training sample data must have the same time points. Thus, the first electromagnetic interference prediction model obtained through training can predict interference parameters such as the type, intensity, and duration of electromagnetic interference based on the electricity consumption feature data.

[0047] S2: Based on the first power consumption characteristic data and the first electromagnetic interference data of the first spectral acquisition environment within the second preset time interval, the first high-risk interference information and the second high-risk interference information are determined respectively.

[0048] In step S1, the first electromagnetic interference prediction model is determined. This model enables electromagnetic interference analysis based on electricity consumption patterns, yielding the first high-risk interference information. Combined with the second high-risk interference information determined through real-time interference characteristics, the two types of high-risk interference information are integrated in step S3 to determine the final target high-risk interference information.

[0049] S2 specifically includes the following sub-steps: S21: Obtain first power consumption characteristic data of at least one of the first high-voltage devices within a second preset time interval.

[0050] In this step, it is necessary to extract the first power consumption characteristic data of all the first high-voltage devices within the second preset time interval in the first spectral acquisition environment based on the current time point. For example, if the current time point is a Sunday at 18:00 at the beginning of the month, and the second preset time interval is 5 minutes, then it is necessary to extract the power consumption data within five minutes before the current time point 18:00 as the first power consumption characteristic data.

[0051] Based on the first electricity consumption characteristic data, the current electricity intensity index, vibration characteristics, temperature characteristics, harmonic characteristics, etc. of all the first high-voltage equipment are calculated. Among them, in order to make the prediction results more accurate, the electricity intensity index includes multiple dimensions such as monthly electricity intensity index, weekly electricity intensity index, and daily electricity intensity index.

[0052] S22: Input the first power consumption characteristic data into the first electromagnetic interference prediction model to determine the first high-risk interference information.

[0053] Since the correlation between electricity consumption characteristic data and electromagnetic interference data has been established in the first electromagnetic interference prediction model, the first high-risk interference information can be predicted using the first electricity consumption characteristic data output in S21 in this step.

[0054] The first high-risk interference information refers to the types of electromagnetic interference that may occur within the third preset time interval and the corresponding interference information.

[0055] The third preset time interval refers to the next time interval following the second preset time interval. This is because the first high-risk interference information predicted using the first electricity consumption characteristic data has a certain degree of uncertainty; it can only predict electromagnetic interference that is more likely to occur in the following third preset time interval based on the electricity consumption data characteristics of the current second preset time interval. Furthermore, in subsequent steps, the first high-risk interference information needs to be matched with the measured second high-risk interference information to identify the final electromagnetic interference information.

[0056] The first high-risk interference information includes information such as interference type, interference intensity, and interference duration. During the prediction process of the first electromagnetic interference prediction model, the first high-risk interference information is determined based on the similarity between the first electricity consumption characteristic data and feature points in the model.

[0057] S23: Based on the first electromagnetic interference data within the second preset time interval, determine the second high-risk interference information.

[0058] In this step, information such as frequency, waveform, and signal needs to be collected based on the electromagnetic interference data from at least one of the first high-voltage devices. The electromagnetic interference data refers to data from at least one of the first high-voltage devices within a second preset time interval.

[0059] The second high-risk interference information refers to the type of electromagnetic interference and its corresponding interference information that occurs within the second preset time interval. This second high-risk interference information has a certain degree of certainty, as it is obtained through actual measurements.

[0060] In the specific implementation process, interference data is collected in real time through an electromagnetic interference sensor system, recorded every 50ms. The frequency, waveform, and signal are collected, and wavelet transform is used to extract real-time interference characteristics, thereby obtaining: interference type, interference intensity, and interference duration.

[0061] The identified interference type, interference intensity, and interference duration are used as the second high-risk interference information.

[0062] Both the first high-risk interference information and the second high-risk interference information include one or more types of electromagnetic interference.

[0063] S3: Based on the interference type succession relationship matrix and the second high-risk interference information, determine the target high-risk interference information from the first high-risk interference information.

[0064] In step S2, the electromagnetic interference that may occur within the third preset time interval has been estimated based on power consumption data, and the electromagnetic interference that has already occurred within the second preset time interval has been determined based on measured electromagnetic interference data. In this step, it is necessary to determine the logical relationship strength between the first high-risk interference information and the second high-risk interference information according to the continuity relationship between the second and third preset time intervals, thereby determining a relatively accurate target high-risk interference information.

[0065] S3 includes the following sub-steps: S31: By performing the first transfer analysis on the historical electromagnetic interference data, the interference type succession relationship matrix is ​​determined.

[0066] Because the usage patterns of high-voltage equipment are relatively predictable, different types of electromagnetic interference often occur sequentially. For example, there is a high probability of transient interference following power frequency interference, and a high probability of high-frequency noise interference following transient interference. Therefore, this step requires constructing an interference type succession matrix based on historical data to determine the succession relationships between various interference types.

[0067] Specifically, the first transfer analysis process performs statistical analysis based on historical data, constructs an interference type succession relationship matrix, and calculates the probability of other interference types appearing after each interference type. For specific statistical analysis methods, please refer to existing technologies, which will not be elaborated here.

[0068] For example, the probability of transient interference occurring within 10 minutes after a power frequency interference is 0.75, and the probability of high-frequency noise occurring within 10 minutes after a transient interference is 0.6. Preferably, a Markov chain model can be used to optimize the succession relationship matrix to ensure accurate transition probabilities.

[0069] The interference type succession relationship matrix output in this step contains a 4×4 dimension of transition probabilities, which includes the probability of mutual transition between the four types of electromagnetic interference.

[0070] S32: Based on the second high-risk interference type and the interference type succession matrix, determine at least one target high-risk interference type from the first high-risk interference information.

[0071] In this step, it is necessary to perform a continuity relationship strength analysis on the first high-risk interference information and the second high-risk interference information determined in S2, so as to determine at least one target interference type from the first high-risk interference information whose continuity relationship strength with the second high-risk interference information meets a preset value.

[0072] S32 may specifically include the following sub-steps: S321: Determine at least one first high-risk interference type from the second high-risk interference information.

[0073] Since the second high-risk interference information contains at least one first high-risk interference type, this step only requires information extraction from it.

[0074] S322: Based on the interference type succession relationship matrix, obtain the second high-risk interference type corresponding to each of the first high-risk interference types.

[0075] The interference type succession relationship matrix records the transition probability values ​​from each electromagnetic interference type to the other three electromagnetic interference types.

[0076] Therefore, in this step, the transfer probability values ​​from one of the first high-risk interference types to the other three types of electromagnetic interference will be obtained, and the electromagnetic interference type with the highest transfer probability value will be determined as the second high-risk interference type.

[0077] S323: Based on the second high-risk interference type and the first high-risk interference information, determine the target high-risk interference type.

[0078] The second high-risk interference type indicates the type of electromagnetic interference that is likely to occur within the third preset time interval in the future, while the first high-risk interference type indicates the type of electromagnetic interference that is likely to occur within the third preset time interval, as predicted by the power consumption data of high-voltage equipment.

[0079] In this step, the target high-risk interference type can be determined based on the degree of overlap between the second high-risk interference type and the electromagnetic interference information contained in the first high-risk interference information.

[0080] S33: Based on the second high-risk interference information, determine the target high-risk interference information corresponding to the target high-risk interference type.

[0081] After determining the second high-risk interference type in S32, in this step, it is also necessary to determine the target high-risk interference information corresponding to the second high-risk interference type based on the interference intensity, interference duration, and other information contained in the second high-risk interference information.

[0082] The target high-risk interference information can be obtained based on expert experience assessment, or it can be obtained through big data prediction using a trained machine learning model. Specifically, the target high-risk interference information can be obtained by training a convolutional neural network, wherein the second high-risk interference information and the target high-risk interference type are used as input data for training samples, and the target high-risk interference information is used as output data for training samples.

[0083] S4: Obtain the first near-infrared spectral acquisition parameters based on the target high-risk interference information, and determine the first aviation kerosene quality analysis information.

[0084] In step S3, the high-risk interference types of the target that have a high probability of occurring within the third preset time interval have been identified. In this step, the corresponding first near-infrared spectral acquisition parameters can be determined for each of the high-risk interference types of the target, and the final aviation kerosene quality detection parameters can be obtained.

[0085] S4 includes the following sub-steps: S41: Input the target high-risk interference information into the acquisition parameter determination model to determine the first near-infrared spectral acquisition parameters.

[0086] Based on the high-risk interference information of the target determined by S33, the acquisition parameters of the near-infrared spectrometer are dynamically adjusted.

[0087] The first near-infrared spectral acquisition parameters are obtained through prediction using an acquisition parameter determination model, which is obtained by training a convolutional neural network model. Specifically, historical electromagnetic interference data is selected as samples, and data such as electromagnetic interference type, intensity, and duration are used as input to the model, while the near-infrared spectral acquisition parameters are used as output data.

[0088] For example, the equipment acquisition parameters for near-infrared spectroscopy are adjusted as follows: Power frequency interference: Increase the number of acquisitions to 25, adjust the acquisition interval to 50ms, and enable synchronous acquisition; Transient interference: Use high-speed acquisition mode and enable triggered acquisition; High-frequency noise: Add spectral smoothing processing and adopt adaptive bandpass filtering; Long-term drift: Enable real-time baseline correction and use adaptive windowing.

[0089] In summary, the adjusted data acquisition parameter configuration includes information such as the number of acquisitions, the interval, the preprocessing method, and the triggering conditions.

[0090] S42: Obtain first near-infrared spectral data based on the first near-infrared spectral acquisition parameters, and determine first aviation kerosene quality analysis information based on the first near-infrared spectral data.

[0091] Based on the first near-infrared spectral data, an interference-aware aviation kerosene quality prediction model is used to predict quality indicators. Model structure: Input = corrected spectral data + interference type features; Output = moisture content, sulfur content, aromatic hydrocarbon content, density. The model can use common machine analysis models in this field, such as convolutional neural network models; the specific training process of the model will not be detailed here.

[0092] The output of the first aviation kerosene quality analysis information includes aviation kerosene quality index prediction results, which may specifically include the predicted values ​​of the content of each component of aviation kerosene, prediction confidence level, interference type identifier, etc.

[0093] This application also proposes a near-infrared spectroscopy real-time monitoring system for performing the aforementioned near-infrared spectroscopy real-time monitoring method.

[0094] This application proposes a real-time near-infrared spectroscopy monitoring method and system, relating to the field of spectral analysis technology. It analyzes the electromagnetic interference (EMI) caused by high-voltage equipment to near-infrared spectroscopy equipment during operation. Based on a first set of power consumption characteristic data and a first set of EMI data under a first spectral acquisition environment, a first EMI prediction model is trained. Then, based on the power consumption characteristic data, a first high-risk EMI information is predicted, and based on the EMI data, a second high-risk EMI information is determined. Next, based on the continuity relationship between different EMI types, the target high-risk EMI information is determined. Finally, the acquisition parameters of the near-infrared spectroscopy equipment are adjusted to obtain first aviation kerosene quality analysis information. The technical solution of this application can effectively reduce the interference caused by EMI to near-infrared spectral identification. Furthermore, since the acquisition of EMI data and near-infrared spectral data has strong real-time requirements, this invention can achieve real-time dynamic monitoring.

[0095] The above description is only a preferred embodiment of the present invention. Therefore, all equivalent changes or modifications made to the structure, features and principles described in the claims of this patent application are included in the scope of this patent application.

Claims

1. A method for real-time monitoring of near-infrared spectroscopy, characterized in that, The method includes: S1: Determine the first set of power consumption characteristic data and the first set of electromagnetic interference data in the first spectrum acquisition environment, and train to obtain the first electromagnetic interference prediction model; Wherein, the first electricity consumption characteristic data set and the first electromagnetic interference data set are both data sets within a first preset time interval; S2: Based on the first power consumption characteristic data and the first electromagnetic interference data of the first spectral acquisition environment within the second preset time interval, the first high-risk interference information and the second high-risk interference information are determined respectively. S3: Based on the interference type succession relationship matrix and the second high-risk interference information, determine the target high-risk interference information from the first high-risk interference information; S4: Obtain the first near-infrared spectral acquisition parameters based on the target high-risk interference information, and determine the first aviation kerosene quality analysis information.

2. The near-infrared spectroscopy real-time monitoring method according to claim 1, characterized in that, S1 includes: S11: In the first spectral acquisition environment, acquire a first set of power consumption characteristic data of at least one first high-voltage device within a first preset time interval; S12: Analyze the electromagnetic interference data within the first preset time interval to determine the first electromagnetic interference data set; S13: A first electromagnetic interference prediction model is obtained by training based on the first set of electricity consumption characteristic data and the first set of electromagnetic interference data.

3. The near-infrared spectroscopy real-time monitoring method according to claim 2, characterized in that, The first set of electricity consumption characteristic data includes: Device ID, timestamp, usage intensity, vibration characteristics, temperature characteristics, harmonic characteristics.

4. The near-infrared spectroscopy real-time monitoring method according to claim 1, characterized in that, S2 includes: S21: Obtain first power consumption characteristic data of at least one of the first high-voltage devices within a second preset time interval; S22: Input the first power consumption characteristic data into the first electromagnetic interference prediction model to determine the first high-risk interference information; S23: Based on the first electromagnetic interference data within the second preset time interval, determine the second high-risk interference information.

5. The near-infrared spectroscopy real-time monitoring method according to claim 4, characterized in that, The first high-risk interference information includes the type of interference, the intensity of interference, and the duration of interference.

6. The near-infrared spectroscopy real-time monitoring method according to claim 1, characterized in that, S3 includes: S31: By performing a first transfer analysis on historical electromagnetic interference data, the interference type succession relationship matrix is ​​determined; S32: Based on the second high-risk interference type and the interference type succession relationship matrix, determine at least one target high-risk interference type from the first high-risk interference information; S33: Based on the second high-risk interference information, determine the target high-risk interference information corresponding to the target high-risk interference type.

7. The near-infrared spectroscopy real-time monitoring method according to claim 6, characterized in that, S32 includes: S321: Identify at least one first high-risk interference type from the second high-risk interference information; S322: Based on the interference type succession relationship matrix, obtain the second high-risk interference type corresponding to each of the first high-risk interference types; S323: Based on the second high-risk interference type and the first high-risk interference information, determine the target high-risk interference type.

8. The near-infrared spectroscopy real-time monitoring method according to claim 1, characterized in that, S4 includes: S41: Input the target high-risk interference information into the acquisition parameter determination model to determine the first near-infrared spectral acquisition parameters; S42: Obtain first near-infrared spectral data based on the first near-infrared spectral acquisition parameters, and determine first aviation kerosene quality analysis information based on the first near-infrared spectral data.

9. A near-infrared spectroscopy real-time monitoring system for implementing the near-infrared spectroscopy real-time monitoring method according to any one of claims 1-8.

Images