A device state acquisition method and system based on a fiber-optic power supply sensor

Abstract

This invention relates to the field of industrial automation and equipment monitoring technology, and discloses a method and system for acquiring equipment status based on fiber optic powered sensors. The method includes acquiring multi-parameter data and energy supply levels and processing them to obtain an initial transmission signal; determining a stable energy supply based on the initial transmission signal; acquiring data under a stable energy supply and processing it synchronously to obtain a synchronous data sequence; extracting features from the synchronous data sequence and fusing them to determine a comprehensive status index; if the index deviates, performing dynamic weight optimization to determine an optimized weight set; verifying the location correlation strength based on the optimized weight set to obtain a verification correlation index; and generating a final status assessment report based on the optimized weight set and the verification correlation index, and verifying the sensor positioning reliability. This method improves the accuracy and reliability of status acquisition through energy compensation and dynamic weight optimization.

Description

Technical Field

[0001] This invention relates to the field of industrial automation and equipment monitoring technology, and in particular to a method and system for acquiring equipment status based on fiber optic powered sensors. Background Technology

[0002] Currently, equipment condition monitoring is a crucial link in ensuring system safety and efficiency in industries such as industry and energy. In complex industrial scenarios with long distances and high interference, real-time and accurate status data acquisition of multiple devices is a core requirement for ensuring reliable system operation.

[0003] In existing technologies, mainstream monitoring solutions often rely on cable power supply or wireless communication. However, in long-distance or high electromagnetic interference environments, data acquisition efficiency and stability are difficult to guarantee. To address this, some solutions attempt to use optical fibers to simultaneously supply power and transmit signals. However, this approach is limited by optical power attenuation, often resulting in insufficient power supply to remote intelligent sensor clusters, affecting the stability of data acquisition. Furthermore, when it is necessary to fuse multiple parameter data such as vibration and temperature, existing technologies struggle to optimize weight allocation in real time according to the dynamic environment, leading to a decrease in the accuracy of the acquisition results.

[0004] Therefore, existing technologies suffer from insufficient accuracy and reliability in status acquisition. Summary of the Invention

[0005] This invention provides a device status acquisition method and system based on fiber optic power supply sensors to solve the problems of insufficient accuracy and reliability of status acquisition in the prior art.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a device status acquisition method based on an optical fiber powered sensor, comprising: The system acquires and processes multi-parameter data and energy supply level fed back by the fiber-optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal. Based on the power attenuation analysis of the initial transmission signal, the amount of optical power injected into the optical fiber power sensor is dynamically adjusted to determine a stable energy supply. Acquire multi-parameter data under the stable energy supply and perform time-series synchronization processing to obtain a synchronization data sequence; Features are extracted from the synchronized data sequence and multi-parameter fusion processing is performed to determine the comprehensive status index; If the comprehensive status index deviates from the preset normal range, dynamic weight optimization processing is performed to determine the optimization weight set; The location association strength is verified based on the optimized weight set to obtain the verification association index; A final state evaluation report is generated based on the optimized weight set and the verification correlation index, and the reliability of sensor positioning is verified.

[0007] Secondly, the present invention provides a device status acquisition system based on an optical fiber powered sensor, comprising: The signal acquisition module is used to acquire and process multi-parameter data and energy supply level fed back by the fiber optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal. An energy stabilization module is used to analyze power attenuation based on the initial transmission signal, dynamically adjust the amount of optical power injected into the fiber optic power sensor, and determine a stable energy supply. The data synchronization module is used to acquire multi-parameter data under the stable energy supply and perform time-series synchronization processing to obtain a synchronized data sequence. The status assessment module is used to extract features from the synchronized data sequence and perform multi-parameter fusion processing to determine a comprehensive status index. The weight optimization module is used to perform dynamic weight optimization processing and determine the optimized weight set if the comprehensive status index deviates from the preset normal range. The association verification module is used to verify the location association strength based on the optimized weight set and obtain the verification association index. The reporting and verification module is used to generate a final state evaluation report based on the optimized weight set and the verification correlation index, and to perform sensor positioning reliability verification.

[0008] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the device status acquisition method based on the fiber optic power supply sensor described in any one of the above.

[0009] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the device status acquisition method based on the fiber optic power supply sensor described above.

[0010] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention determines a stable energy supply by analyzing the power attenuation of the initial transmission signal and activating the compensation mechanism, and constructs a closed loop of energy adaptive compensation, which ensures the stability of data acquisition and solves the problem of insufficient accuracy of state acquisition in the prior art.

[0011] (2) This invention establishes a dynamic weighting mechanism based on real-time status feedback by calculating the evaluation deviation value between the comprehensive status index and the normal range, and dynamically optimizing the weight set according to the deviation value, thereby improving the accuracy of data fusion and solving the problem of insufficient reliability of status acquisition caused by rigid fusion strategy and inability to adapt to changes in working conditions in the prior art.

[0012] (3) This invention combines optimized weight set and verification correlation index to re-integrate data, generate final status assessment report and perform sensor positioning reliability verification. It combines dynamically optimized data with sensor position distribution verification to form a complete closed loop from data to decision, and realizes highly reliable equipment status assessment. Attached Figure Description

[0013] Figure 1 This is a schematic flowchart of a device status acquisition method based on an optical fiber power supply sensor provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of a device status acquisition system based on an optical fiber power supply sensor provided in the second embodiment of the present invention. Detailed Implementation

[0014] 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.

[0015] Reference Figure 1 The first embodiment of the present invention provides a device status acquisition method based on an optical fiber power supply sensor, comprising the following steps: S11: Acquire multi-parameter data and energy supply level fed back by the fiber optic powered sensor through the fiber optic transmission link, process them, and obtain the initial transmission signal. S12, based on the power attenuation analysis of the initial transmission signal, dynamically adjust the optical power injection amount to the optical fiber power sensor to determine a stable energy supply; S13, acquire multi-parameter data under the stable energy supply, and perform time-series synchronization processing to obtain a synchronization data sequence; S14, extract features from the synchronized data sequence and perform multi-parameter fusion processing to determine the comprehensive status index; S15, If the comprehensive status index deviates from the preset normal range, dynamic weight optimization processing is performed to determine the optimization weight set; S16, Verify the location association strength based on the optimized weight set to obtain the verification association index; S17. Generate a final state evaluation report based on the optimized weight set and the verification correlation index, and perform sensor positioning reliability verification.

[0016] In step S11, multi-parameter data and energy supply level fed back by the fiber-optic powered sensor through the fiber-optic transmission link are acquired and processed to obtain the initial transmission signal, including: The raw sensing data is obtained by acquiring the multi-parameter data and the energy supply level. Perform signal separation processing on the original sensing data to obtain separated data; The separated data are subjected to interference suppression processing to obtain suppressed data; The suppressed data is normalized to obtain the initial transmission signal.

[0017] In one implementation, this embodiment acquires the multi-parameter data and the energy supply level from a sensor cluster deployed at the monitoring site via an optical fiber medium. It should be noted that the multi-parameter data may include signals such as vibration, temperature, and pressure, and the energy supply level is the real-time feedback value of the optical power transmitted to the sensor via the optical fiber. This embodiment combines the acquired multi-parameter data and the energy supply level to obtain the raw sensor data.

[0018] In one implementation, the signal separation processing is a digital filtering operation based on preset frequency characteristics. This operation applies band-pass filtering and low-pass filtering to the original sensor data according to the preset frequency characteristics of each signal. It should be noted that the cutoff frequencies of each filter are determined during the system calibration phase. This embodiment acquires original sensor data samples under representative operating conditions, performs spectral analysis using Fast Fourier Transform (FFT) on the sample data, and identifies the characteristic frequency bands where the energy of different signals is mainly concentrated. This embodiment sets the cutoff frequencies of each filter according to the characteristic frequency bands, separating high-frequency vibration signals from low-frequency temperature and pressure signals to obtain the separated data. For example, if the spectral analysis identifies that the vibration signal is mainly concentrated in the [20Hz, 100Hz] frequency band, while the temperature and pressure signals are both in the [0Hz, 5Hz] frequency band, then the cutoff frequencies of the band-pass filtering operation can be set to 20Hz and 100Hz, and the cutoff frequency of the low-pass filtering operation can be set to 5Hz.

[0019] In another implementation, the interference suppression processing can be achieved using wavelet transform. This embodiment performs a three-level wavelet decomposition on the separated data. It should be noted that the decomposition level (three levels) was determined through comparative experiments during the system debugging phase. Different decomposition levels (e.g., levels 2, 3, 4, and 5) were set to process sample data containing typical interference, and the signal-to-noise ratio (SNR) of the processed signal and its cross-correlation coefficient with the original clean signal were calculated. This embodiment selects a level (three levels) that achieves a balance between SNR and cross-correlation coefficient as the final configuration. This decomposition process automatically decomposes the signal into multiple sets of wavelet coefficients corresponding to different scales (i.e., different frequency components). Subsequently, this embodiment performs threshold quantization processing on the coefficients that mainly correspond to high-frequency noise among the multiple sets of wavelet coefficients. It should be noted that this processing can specifically involve setting wavelet coefficients with amplitudes lower than a preset noise threshold to zero. The preset noise threshold was determined during the system debugging phase through statistical analysis of a large amount of sensor data under clean conditions without equipment failure. This embodiment analyzes the amplitude distribution of high-frequency wavelet coefficients under the clean state and sets a threshold that can represent the normal noise baseline based on statistical principles (e.g., selecting the amplitude corresponding to the 99th percentile). Finally, this embodiment obtains the suppressed data through wavelet reconstruction.

[0020] In one implementation, the normalization process employs a minimum-maximum normalization method. In this embodiment, the parameter values ​​in the suppressed data are linearly mapped to a unified interval [0,1] based on their preset physical ranges (e.g., temperature 0-100°C or pressure 0-5MPa). It should be noted that the preset physical ranges are determined during the calibration phase according to the specifications of the sensor used. The data after this normalization process constitutes the initial transmission signal.

[0021] For example, the acquired raw sensing data includes high-frequency vibration signals and low-frequency temperature signals. In this embodiment, the data is first separated into the separated data using a filter bank; then, wavelet transform is used to perform three-level decomposition and threshold denoising on the separated vibration signals to obtain the suppressed data; finally, the denoised vibration data (e.g., 50Hz) and temperature data (e.g., 30°C) are normalized respectively, with the physical range set based on the sensor specifications: the temperature sensor range is 0-100°C, and the pressure sensor range is 0-5MPa. The resulting normalized values ​​are 0.5 and 0.3, which together constitute the initial transmission signal.

[0022] In step S12, based on the power attenuation analysis of the initial transmission signal, the optical power injection amount to the fiber optic power sensor is dynamically adjusted to determine a stable energy supply, including: The initial transmitted signal is subjected to frequency domain decomposition processing to obtain the frequency domain decomposed signal; The frequency domain decomposed signal is subjected to attenuation suppression processing to obtain an attenuation suppressed signal; The optical power injection amount is dynamically adjusted based on the attenuation suppression signal to determine the stable energy supply.

[0023] In one implementation, the frequency domain decomposition process can be implemented using a Fast Fourier Transform (FFT). This embodiment transforms the initial transmitted signal from the time domain to the frequency domain to obtain the frequency-decomposed signal. It should be noted that the frequency-decomposed signal reflects the distribution of energy at different frequency components in the initial transmitted signal, providing a basis for analyzing power attenuation in a specific frequency band.

[0024] In one implementation, the attenuation suppression processing is a digital gain compensation operation. This embodiment compares the amplitude of the frequency-domain decomposed signal with a preset reference amplitude level, calculates the attenuation of each frequency component, and applies a compensation gain of opposite magnitude to the attenuation to enhance the target frequency component, thus obtaining the attenuation-suppressed signal. It should be noted that the preset reference amplitude level is determined during the system calibration phase. This calibration process involves directly connecting the optical power injection source to the spectrum analyzer via a standard test fiber (e.g., 1 meter), sending a reference signal with known power, and measuring the received frequency-domain amplitude, which is then determined as the preset reference amplitude level.

[0025] In another implementation, the dynamic adjustment is a feedback-based power regulation operation. This embodiment compares the current actual power represented by the attenuation suppression signal with a preset target power level, and adjusts the amount of optical power injected into the fiber optic medium in real time based on the deviation between the two. It should be noted that the preset signal-to-noise ratio baseline (e.g., 30dB) is predetermined based on the sensor specifications and application scenario's data accuracy requirements, representing the minimum signal-to-noise ratio requirement for the sensor to output stable and reliable readings. The preset target power level is determined through offline experiments during the system calibration phase. For example, the offline experimental environment can be set in a standard environment with a temperature of 25°C ± 5°C and humidity of 50% ± 10%, using a 10m long single-mode optical fiber. Under this experimental environment, the amount of optical power injected into the sensor is gradually increased while simultaneously monitoring the signal-to-noise ratio of its output signal. The first time the signal-to-noise ratio reaches the injected power value of the preset signal-to-noise ratio baseline is determined as the preset target power level. The final output of this dynamic adjustment is the stable energy supply.

[0026] For example, assuming that the frequency domain decomposed signal, after the attenuation suppression processing, results in an attenuation suppression signal indicating that the current sensor cluster's received power is 8mW. If the preset target power level is 10mW, this embodiment will calculate a power difference of 2mW and correspondingly increase the optical power injection amount until the received power fed back by the attenuation suppression signal reaches 10mW. At this point, the energy supply in the optical fiber network reaches a stable level, and this 10mW supply level is the stable energy supply.

[0027] In step S13, multi-parameter data under the stable energy supply is acquired and time-series synchronization processing is performed to obtain a synchronization data sequence, including: Acquire multi-parameter data under the stable energy supply, and perform timestamp marking processing on the multi-parameter data to obtain a timestamped multi-parameter data sequence; Perform timestamp alignment processing on the timestamped multi-parameter data sequence to obtain a time-aligned data sequence; The time-aligned data sequence is subjected to cache adjustment and reordering processing to obtain the synchronized data sequence.

[0028] In one implementation, after acquiring the multi-parameter data under the stable energy supply, this embodiment adds a timestamp of a uniform format to each data point from different sensors to obtain the timestamped multi-parameter data sequence.

[0029] In one implementation, the timestamp alignment process first selects a data stream (e.g., a temperature data stream) as a reference time axis from the timestamped multi-parameter data sequence. Subsequently, in this embodiment, linear interpolation is used to adjust the remaining data streams to the reference time axis. It should be noted that the specific calculation for this interpolation is as follows: when a certain time point on the reference time axis is needed... Calculate an estimate at the location. In this embodiment, the distance is found from the data stream to be adjusted. The two most recent data points and ,in The estimated value It is calculated using the following formula:

[0030] All calculated estimates Together, they constitute the time-aligned data sequence.

[0031] It is worth noting that the cache adjustment and reordering process monitors the transmission delay of the time-aligned data sequence. When data packets arrive out of order (for example, a data frame with timestamp .125 arrives before a data frame with timestamp .123), this embodiment temporarily stores the earlier-arriving data frame through a caching mechanism and waits for the subsequent delayed data frames to arrive. After receiving the complete set of data packets, this embodiment reorders the time-aligned data sequence in ascending order of timestamps using a quicksort algorithm to obtain the synchronization data sequence.

[0032] For example, assuming the timestamp of the temperature data is 12:00:00.123 and the pressure data is 12:00:00.125, this embodiment uses linear interpolation to adjust the pressure data to the time point of 12:00:00.123, aligning it with the temperature data to obtain the time-aligned data sequence. If, due to transmission delay, the data frame with timestamp .125 arrives before the data frame with timestamp .123, this embodiment temporarily caches the data frame with timestamp .125. After the data frame with timestamp .123 arrives, it is reordered according to the timestamp order to .123 and .125, ultimately generating the synchronized data sequence.

[0033] In step S14, features are extracted from the synchronized data sequence and multi-parameter fusion processing is performed to determine the comprehensive state index, including: Vibration features, temperature features, and pressure features are extracted from the synchronous data sequence to obtain vibration feature sequences, temperature feature sequences, and pressure feature sequences, respectively. The vibration feature sequence, the temperature feature sequence, and the pressure feature sequence are fused to obtain a fused feature sequence. The fused feature sequence is classified to determine the comprehensive state index.

[0034] In one implementation, this embodiment segments the synchronous data sequence (which is a continuous data stream) into segments of a preset data length (e.g., 1024 data points), and performs targeted feature extraction operations on each data segment. It should be noted that the preset data length is determined during the system calibration phase. This calibration process analyzes the duration of the shortest known transient fault event in history (e.g., a pressure pulse or vibration shock), and, according to the sampling theorem, selects a data length that both fully covers the shortest duration and provides the required frequency resolution for the Fast Fourier Transform (FFT). For vibration data, this embodiment can use Fast Fourier Transform (FFT) processing to convert it from the time domain to the frequency domain, and find the frequency point with the highest amplitude in the frequency domain, using this frequency value as the vibration feature. For temperature data, this embodiment can use mean filtering processing to smooth instantaneous fluctuations by calculating the average temperature within the data segment, extracting stable temperature values ​​as the temperature feature. For pressure data, this embodiment can use a differential algorithm to process it. By calculating the pressure difference between all adjacent time points within the data segment and taking the average value, the average pressure change rate is extracted as the pressure feature.

[0035] In one implementation, the fusion process includes normalization and weighted averaging. This embodiment first performs min-max normalization on the vibration feature, temperature feature, and pressure feature (i.e., three single values ​​extracted from a data segment), scaling all feature values ​​to a uniform range of [0,1]. Then, this embodiment calculates a weighted average of the normalized feature values ​​based on the currently used weight set, resulting in a single value, which is the fused feature sequence. It should be noted that when this fusion process is executed for the first time, the currently used weight set is a set of preset weights (e.g., vibration 0.4, temperature 0.3, pressure 0.3). These preset weights are determined through offline historical data analysis, which includes calculating the correlation coefficient between each historical feature sequence and a binary sequence of known equipment fault states (e.g., 0 representing normal, 1 representing fault). Subsequently, the proportion normalization method is used, i.e., the formula is adopted:

[0036] The preset weights are obtained by normalizing the absolute values ​​of the correlation coefficients of all parameters. In subsequent processing, if S15 generates an optimized weight set based on the evaluation deviation value, the currently used weight set will be updated to this optimized weight set for processing subsequent data segments.

[0037] In another implementation, the classification process can be implemented using a pre-trained Support Vector Machine (SVM) model. This embodiment processes the fused feature sequence (single numerical value) according to pre-established classification rules. It should be noted that the pre-established classification rules are generated by training the SVM model offline. This training process includes: acquiring a labeled historical dataset containing a large number of historical fused feature sequence samples (single numerical values) and corresponding verified device states (e.g., "normal operation" or "abnormal stress"). The SVM model can employ a radial basis function (RBF) kernel. This embodiment uses this dataset to train the model, and during training, the Adam optimizer iteratively minimizes the Hinge loss function. The key hyperparameters of the model, such as the penalty coefficient C and the kernel function parameter gamma, are determined by performing a grid search combined with k-fold cross-validation during the training phase. The search space of the grid search can be set such that the candidate value set for C is {0.1, 1, 10, 100}, and the candidate value set for gamma is {0.001, 0.01, 0.1, 1}. In this embodiment, the hyperparameter combination with the highest classification accuracy on the validation set is selected as the final configuration. Training terminates when the model's validation set loss value falls below a preset convergence threshold. This preset convergence threshold (e.g., 0.01) is determined by analyzing the model's learning curve, which is obtained by plotting the model's loss value on the validation set as a function of the number of training iterations. In this embodiment, the threshold is chosen when the loss value decline curve begins to flatten (i.e., the "inflection point"), achieving a balance between training efficiency and model performance. In this embodiment, the fused feature sequence acquired in real-time is input into the pre-trained SVM model, and the model outputs a numerical score representing the probability of anomalies (e.g., ranging from [0,1], with higher values ​​indicating a higher probability of anomalies). This numerical score sequence generated from continuous data segments constitutes the comprehensive state index.

[0038] For example, in this embodiment, from a data segment of 1024 points, the feature with a dominant frequency of 50Hz is extracted using FFT, the stable value of 40.0°C is extracted using mean filtering, and the average rate of change of 0.1MPa / s is extracted using a difference algorithm. Subsequently, this embodiment normalizes these three feature values ​​and performs a weighted average using preset weights (vibration 0.4, temperature 0.3, pressure 0.3) to obtain a fused feature sequence value (e.g., 0.65). Finally, this value of 0.65 is input into a pre-trained SVM model, and the model outputs an anomaly probability score (e.g., 0.15). This score serves as a data point in the comprehensive state index sequence.

[0039] In step S15, if the comprehensive status index deviates from the preset normal range, dynamic weight optimization processing is performed to determine the optimization weight set, including: If the comprehensive status index deviates from the preset normal range, the difference between the current value of the comprehensive status index and the preset normal range is calculated to obtain the evaluation deviation value; The preset weights used in the multi-parameter fusion processing are adjusted based on the evaluation deviation value to obtain the optimized weight set.

[0040] In one implementation, this embodiment compares the current value of the comprehensive status index (which is a numerical scoring sequence, such as a score representing the probability or severity of anomalies) output in S14 with a preset normal range.

[0041] It should be noted that the preset normal range (e.g., [0, 0.2]) is determined by collecting and statistically analyzing historical normal data. Specifically, at least 1000 sets of comprehensive status index data samples under normal equipment operation conditions can be collected, and the mean of the historical normal data can be calculated. with standard deviation And based on the 3-sigma principle, (And intersecting with the theoretical range of the indicator, such as [0,1]) is determined as the preset normal range. The mean of historical normal data; The standard deviation represents the historical normal data. If the current value of the comprehensive status index exceeds the normal range, this embodiment calculates the difference between the value and the nearest boundary of the normal range; this single difference is the evaluation deviation value. If the current value of the comprehensive status index does not exceed the normal range, the evaluation deviation value is not generated, and this embodiment directly proceeds to S16 without performing weight optimization processing.

[0042] For example, the preset normal range determined through historical data analysis is [0, 0.2]. If the current comprehensive status index value output by S14 is 0.35, since it exceeds the upper limit of 0.2, this embodiment calculates its deviation as 0.15 (i.e., 0.35 - 0.2), obtaining the evaluation deviation value of 0.15. If the current value is 0.18, it does not exceed the range, and no evaluation deviation value is generated.

[0043] In one implementation, the dynamic weight optimization process adjusts the preset weights used in the S14 fusion processing step when the evaluation deviation value exists (i.e., the indicator deviates from the normal range), based on the evaluation deviation value (which is a calculated single numerical deviation). This adjustment employs a contribution-based heuristic rule. First, this embodiment determines the contribution of each feature causing the deviation through sensitivity analysis. This sensitivity analysis can be specifically described as, for each feature, considering the vicinity of the feature values ​​in the current data segment. Apply a small perturbation to the value. And observe the comprehensive state index output by the SVM model in S14. Change ; among which disturbance It can be set to 1% of the characteristic value, such as when the characteristic value is 50Hz. =0.5Hz, will Min-max normalization is used to map to the [0,1] interval to ensure dimensionlessness. Sensitivity is calculated using the formula:

[0044] Subsequently, this embodiment calculates the normalized contribution of each feature. :

[0045] in, Representative characteristics Normalized contribution; Representative characteristics Sensitivity; Iterate through the indices of all features.

[0046] This embodiment adjusts the weights according to the following rules regarding contribution. Biggest feature Its new weight In the original weight Based on this, add an adjustment amount proportional to the evaluation deviation value; for other features Its new weight In the original weight The weights are reduced accordingly to ensure that the sum of all new weights remains 1. The specific adjustment formula can be set as follows:

[0047]

[0048] in, The characteristic representing the greatest contribution The new weights; The characteristic representing the greatest contribution The original weights; The preset adjustment coefficient (determined through historical data grid search, with a search range of [0.1, 1.0], aiming to achieve the highest accuracy of the weighted state evaluation, for example, 0.5); This represents the evaluation deviation value; Representing other characteristics The new weights; Representing other characteristics The original weights; Traversal An index of all features except those mentioned above.

[0049] Finally, this embodiment applies to all new weights. Perform nonnegative clipping and normalization if If <0, then set = 0; if > 1, then set = 1. Subsequently, all weights are renormalized so that the sum is 1, ensuring that the weight set conforms to the probability distribution requirements, thus obtaining the optimized weight set.

[0050] In step S16, the location association strength is verified based on the optimized weight set to obtain the verification association index, including: The multi-parameter data is classified and processed according to the optimized weight set to generate a fused correlation sequence; The sensor distribution locations are obtained, and a correlation analysis is performed between the sensor distribution locations and the fused correlation sequence to obtain the connection strength value; The connection strength value is subjected to threshold verification processing to obtain the verification correlation index.

[0051] In one implementation, this embodiment performs weighted clustering on the multi-parameter data (output of S13, i.e., multiple parallel data streams after synchronization) based on the optimized weight set (output of S15). It should be noted that the clustering process can employ the weighted K-means method. The number of clusters K (e.g., K=3) is determined during the system calibration phase by applying the Elbow Method to historical multi-parameter data. This determination process is as follows: This embodiment runs K-means clustering for a preset range of K values ​​(e.g., K from 2 to 10) and calculates the Within-Cluster Sum of Squares (WCSS) for each K value; this embodiment plots a curve of WCSS changing with K values ​​and selects the K value corresponding to the "elbow" where the curve slope changes most significantly as the final configuration. This weighted K-means processing calculates the data points... With cluster center When calculating the distance between them, a weighted Euclidean distance was used, and its calculation formula is as follows:

[0052] in, Representative data points With cluster center Weighted distance between them; and Representing data points respectively and cluster center In the Values ​​in each feature dimension; Represents the first in the optimized weight set The weight values ​​corresponding to each feature.

[0053] This clustering process divides multi-parameter data points with similar characteristics into K clusters. Subsequently, this embodiment assigns a state label to each cluster by comparing the cluster center (i.e., the mean vector of all data points within the cluster) with predefined typical characteristic ranges of "normal," "slightly abnormal," and "severely abnormal" states. The fused association sequence records these data points and their corresponding state labels.

[0054] In one implementation, this embodiment first obtains the pre-configured sensor distribution locations. It should be noted that the sensor distribution locations can be a dataset containing each sensor number and its logical location in the industrial site, stored in a structured format such as JSON or XML, including sensor IDs, logical location descriptions, and coordinate information. For example, the file content could be {'S1': {'location': 'pump station inlet', 'coordinates': [x1, y1, z1]}, 'S2':{'location': 'valve downstream', 'coordinates': [x2, y2, z2]}}. Subsequently, this embodiment performs association analysis processing on the sensor distribution locations and the fused association sequence. This processing calculates specific logical locations. Sensor data is tagged with specific state labels. conditional probability The calculated conditional probability is then used as the connection strength value.

[0055] in, Represents a specific status label (e.g., "critical anomaly"); Represents a specific sensor logical location (e.g., "pump station inlet"); Represents a given logical position Under these conditions, sensor data is tagged as status labels. The conditional probability.

[0056] In another implementation, the threshold verification process compares the connection strength value with a preset connection threshold. It should be noted that the preset connection threshold is determined through Receiver Operating Characteristic (ROC) curve analysis. Specifically, the true positive rate (TPR) and false positive rate (FPR) at different thresholds are calculated using a historical dataset containing connection strength values ​​and corresponding real event labels. The threshold corresponding to the point closest to the top-left corner of the ROC curve is selected. For example, by analyzing 500 sets of historical data, a threshold of 0.7 is determined. In this embodiment, the point closest to the top-left corner of the curve (i.e., maximizing both TPR and minimizing FPR) is selected, and the connection strength value corresponding to that point is set as the preset connection threshold. If the connection strength value exceeds the preset connection threshold, this embodiment marks it as "highly relevant," and this mark is the verification correlation index.

[0057] For example, assuming that after weighted K-means clustering based on the optimized weight set, the vibration data of the "pump station inlet" sensor is marked as "severely abnormal," the fused correlation sequence is obtained. In this embodiment, the conditional probability is calculated to be 0.85 through correlation analysis, which is the connection strength value. If the preset connection threshold determined by ROC analysis is 0.7, since 0.85 exceeds 0.7, this embodiment determines the result as "high correlation," and "high correlation" is the verification correlation index.

[0058] In step S17, a final state evaluation report is generated based on the optimized weight set and the verification correlation index, and sensor positioning reliability verification is performed, including: The final state evaluation report is generated based on the comprehensive state index, the optimized weight set, and the verification correlation index. By analyzing the correlation strength between the location verification error and the location using the final status evaluation report and the verification correlation index, the reliability of sensor positioning is verified.

[0059] In one implementation, this embodiment combines the comprehensive status index (numerical score sequence) output in S14, the optimized weight set (adjusted weights) output in S15, and the verification correlation index (e.g., "high correlation") output in S16 to generate the final status assessment report. It should be noted that this report may include the device's current quantitative status score and potential risk warnings associated with specific sensor locations (if the verification correlation index is "high correlation"). For example, the report is output in JSON format and includes the fields: {'status_score': 0.85 (range 0-1), 'risk_locations': ['pump station inlet'], 'timestamp': '2023-10-23 12:00:00'}.

[0060] It is worth noting that in processing the verification of sensor positioning reliability, this embodiment first acquires the pre-configured position verification error data of the sensor. It should be noted that this position verification error data is measured during the sensor installation or calibration phase, for example, using a laser rangefinder during sensor installation, recording the deviation between the actual position and the designed position, with the unit set to mm. This data can be acquired along with the raw sensor data in S11. Subsequently, this embodiment combines the position verification error data, the risk locations indicated in the final state assessment report, and the position correlation strength in the verification correlation indicators to perform a final reliability verification analysis. This analysis checks whether the position verification error corresponding to sensor locations marked as "highly correlated" and indicating risk in the report is lower than a preset error tolerance threshold.

[0061] It should be noted that the preset error tolerance threshold (e.g., 5 mm) is determined based on the equipment monitoring accuracy requirements and the sensor installation process level. For example, for pipeline monitoring scenarios, the positioning error is required to be less than or equal to 5 mm to ensure data validity. If both "high correlation" and the position error are below the tolerance threshold, the risk positioning is confirmed to be reliable, and the positioning reliability verification passes; otherwise, it is marked as an abnormal positioning reliability, indicating that a misjudgment may be caused by sensor installation deviation.

[0062] In summary, this invention achieves high accuracy and high reliability in acquiring equipment status by constructing an energy adaptive compensation, dynamic weight optimization based on deviation feedback, and a verification and validation closed loop combined with position analysis.

[0063] Reference Figure 2 The second embodiment of the present invention provides a device status acquisition system based on an optical fiber powered sensor, comprising: The signal acquisition module is used to acquire and process multi-parameter data and energy supply level fed back by the fiber optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal. An energy stabilization module is used to analyze power attenuation based on the initial transmission signal, dynamically adjust the amount of optical power injected into the fiber optic power sensor, and determine a stable energy supply. The data synchronization module is used to acquire multi-parameter data under the stable energy supply and perform time-series synchronization processing to obtain a synchronized data sequence. The status assessment module is used to extract features from the synchronized data sequence and perform multi-parameter fusion processing to determine a comprehensive status index. The weight optimization module is used to perform dynamic weight optimization processing and determine the optimized weight set if the comprehensive status index deviates from the preset normal range. The association verification module is used to verify the location association strength based on the optimized weight set and obtain the verification association index. The reporting and verification module is used to generate a final state evaluation report based on the optimized weight set and the verification correlation index, and to perform sensor positioning reliability verification.

[0064] It should be noted that the device status acquisition system based on an optical fiber power supply sensor provided in this embodiment of the invention is used to execute all the process steps of the device status acquisition method based on an optical fiber power supply sensor in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0065] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a device status acquisition program based on a fiber optic powered sensor. When the processor executes the computer program, it implements the steps in the various embodiments of the device status acquisition method based on a fiber optic powered sensor described above, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above system embodiments, such as the signal acquisition module.

[0066] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0067] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0068] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0069] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0070] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0071] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0072] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for acquiring device status based on an optical fiber powered sensor, characterized in that, include: The system acquires and processes multi-parameter data and energy supply level fed back by the fiber-optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal. Based on the power attenuation analysis of the initial transmission signal, the amount of optical power injected into the optical fiber power sensor is dynamically adjusted to determine a stable energy supply. Acquire multi-parameter data under the stable energy supply and perform time-series synchronization processing to obtain a synchronization data sequence; Features are extracted from the synchronized data sequence and multi-parameter fusion processing is performed to determine the comprehensive status index; If the comprehensive status index deviates from the preset normal range, dynamic weight optimization processing is performed to determine the optimization weight set; The location association strength is verified based on the optimized weight set to obtain the verification association index; A final state evaluation report is generated based on the optimized weight set and the verification correlation index, and sensor positioning reliability is verified. The step of analyzing power attenuation based on the initial transmission signal, dynamically adjusting the optical power injection amount to the fiber optic power sensor, and determining a stable energy supply includes: performing frequency domain decomposition processing on the initial transmission signal to obtain a frequency domain decomposed signal; performing attenuation suppression processing on the frequency domain decomposed signal to obtain an attenuation suppression signal; and dynamically adjusting the optical power injection amount based on the attenuation suppression signal to determine the stable energy supply. The step of extracting features from the synchronized data sequence and performing multi-parameter fusion processing to determine the comprehensive state index includes: extracting vibration features, temperature features, and pressure features from the synchronized data sequence to obtain vibration feature sequences, temperature feature sequences, and pressure feature sequences; fusing the vibration feature sequences, temperature feature sequences, and pressure feature sequences to obtain fused feature sequences; and classifying the fused feature sequences to determine the comprehensive state index. The step of performing dynamic weight optimization processing to determine the optimized weight set if the comprehensive status index deviates from the preset normal range includes: if the comprehensive status index deviates from the preset normal range, calculating the difference between the current value of the comprehensive status index and the preset normal range to obtain an evaluation deviation value; and adjusting the preset weights used in the multi-parameter fusion processing according to the evaluation deviation value to obtain the optimized weight set. The step of verifying the location association strength based on the optimized weight set to obtain the verification association index includes: classifying the multi-parameter data according to the optimized weight set to generate a fused association sequence; obtaining the sensor distribution locations and performing association analysis between the sensor distribution locations and the fused association sequence to obtain a connection strength value; and performing threshold verification processing on the connection strength value to obtain the verification association index.

2. The device status acquisition method based on fiber optic power supply sensor according to claim 1, characterized in that, The process of acquiring and processing multi-parameter data and energy supply level fed back by the fiber-optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal includes: The raw sensing data is obtained by acquiring the multi-parameter data and the energy supply level. Perform signal separation processing on the original sensing data to obtain separated data; The separated data are subjected to interference suppression processing to obtain suppressed data; The suppressed data is normalized to obtain the initial transmission signal.

3. The device status acquisition method based on fiber optic power supply sensor according to claim 1, characterized in that, The process of acquiring multi-parameter data under stable energy supply and performing time-series synchronization processing to obtain a synchronization data sequence includes: Acquire multi-parameter data under the stable energy supply, and perform timestamp marking processing on the multi-parameter data to obtain a timestamped multi-parameter data sequence; Perform timestamp alignment processing on the timestamped multi-parameter data sequence to obtain a time-aligned data sequence; The time-aligned data sequence is subjected to cache adjustment and reordering processing to obtain the synchronized data sequence.

4. The device status acquisition method based on fiber optic power supply sensor according to claim 1, characterized in that, The step of generating a final state evaluation report based on the optimized weight set and the verification correlation index, and performing sensor positioning reliability verification, includes: The final state evaluation report is generated based on the comprehensive state index, the optimized weight set, and the verification correlation index. By analyzing the correlation strength between the location verification error and the location using the final status evaluation report and the verification correlation index, the reliability of sensor positioning is verified.

5. A device status acquisition system based on fiber optic powered sensors, characterized in that, For implementing the method as described in any one of claims 1-4, comprising: The signal acquisition module is used to acquire and process multi-parameter data and energy supply level fed back by the fiber optic powered sensor through the fiber optic transmission link to obtain the initial transmission signal. An energy stabilization module is used to analyze power attenuation based on the initial transmission signal, dynamically adjust the amount of optical power injected into the fiber optic power sensor, and determine a stable energy supply. The data synchronization module is used to acquire multi-parameter data under the stable energy supply and perform time-series synchronization processing to obtain a synchronized data sequence. The status assessment module is used to extract features from the synchronized data sequence and perform multi-parameter fusion processing to determine a comprehensive status index. The weight optimization module is used to perform dynamic weight optimization processing and determine the optimized weight set if the comprehensive status index deviates from the preset normal range. The association verification module is used to verify the location association strength based on the optimized weight set and obtain the verification association index. The reporting and verification module is used to generate a final state evaluation report based on the optimized weight set and the verification correlation index, and to perform sensor positioning reliability verification.

Images