Non-intrusive equipment operation state monitoring method and system based on electromagnetic spectrum

Abstract

The invention relates to the technical field of state monitoring, in particular to a non-intrusive equipment operation state monitoring method and system based on an electromagnetic spectrum. The method comprises the following steps: dynamic multi-dimensional spectrum fingerprint extraction; generating a prototype according to the extracted multi-dimensional spectrum fingerprints; respectively performing intra-prototype contrast learning and inter-prototype contrast learning based on the generated prototypes; constructing a prototype contrast loss function based on contrast learning; image modeling is carried out according to the extracted multi-dimensional spectrum fingerprints, and an image comparison loss function is constructed based on image modeling; performing function mixing based on a loss function mixing strategy to obtain an overall loss function; performing equipment state judgment based on the overall loss function; and outputting a discrimination result. The problem of wide-frequency-band electromagnetic signal aliasing in a petroleum equipment dense scene is effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of condition monitoring, and in particular, to a non-invasive device operation condition monitoring method and system based on electromagnetic spectrum. Background Art

[0002] In the field of oil equipment, the health monitoring of electrical equipment and fracturing equipment is crucial for ensuring the safe and efficient production of oil and gas fields. However, existing monitoring technologies face multiple challenges in practical applications. Taking typical equipment such as the plunger pump, high-pressure pipe string, and explosion-proof motor of a fracturing truck as an example, their working environment is extremely harsh (high temperature, high pressure, strong vibration, corrosive medium). Traditional contact sensors (such as vibration and temperature sensors) need to be directly installed on the surface or inside of the equipment, resulting in complex installation processes and high maintenance costs. For example, the vibration sensor of the plunger rod of a fracturing pump needs to be calibrated frequently and is easily corroded and damaged by mud, resulting in a decrease in the reliability of the monitoring system; the installation of an embedded temperature sensor inside an explosion-proof motor requires shutdown and modification, seriously affecting production continuity. In addition, the explosion-proof requirements for sensors in the oil field's flammable and explosive environment are extremely high, further increasing the deployment cost and safety risk.

[0003] For non-invasive monitoring requirements, existing technologies attempt to analyze equipment status through electromagnetic radiation signals. However, the electromagnetic environment of oil equipment is particularly complex: equipment such as frequency converters and high-power motors of fracturing trucks are densely distributed, and their wide-band electromagnetic signals (30 MHz - 3 GHz) are superimposed on each other. Traditional spectrum sensing technologies rely on fixed-band filtering or single-energy threshold judgment, and it is difficult to separate the characteristics of target equipment from the mixed signals. For example, harmonic interference in similar frequency bands may be generated due to bearing wear of a fracturing pump and insulation aging of a motor. Existing methods lack multi-dimensional feature correlation analysis (such as time-domain transient response and modulation sideband characteristics), and misjudge environmental noise or signals of adjacent equipment as target faults, resulting in a false alarm rate of over 30%.

[0004] In addition, most oil fracturing operations are located in remote areas, with strict requirements for the real-time and offline processing capabilities of the monitoring system. Existing solutions mostly rely on cloud transmission of original spectrum data, but the network coverage at the fracturing site is unstable, and the transmission delay of high-bandwidth signals is serious, unable to meet the requirements of real-time fault warning. A few local embedded terminals are limited by the computing architecture and only support single-band FFT analysis or shallow machine learning models, and it is difficult to process the high-frequency and high-dimensional electromagnetic fingerprint data of oil equipment. For example, the early wear characteristics of the power-end bearing of a fracturing truck only exist in transient pulses of 2 - 5 ms. Due to insufficient computing power, traditional embedded systems cannot complete feature extraction and matching within a millisecond time window, resulting in missed detection of minor faults.

[0005] More critically, the failure modes of oil equipment are strongly coupled. For example, the failure of the plunger pump seal may cause the lubricating oil temperature to rise, which in turn leads to a decline in the insulation performance of the motor winding. Such cascading failures require the correlated analysis of multi-physical field data including electromagnetic, temperature, and vibration. Existing non-invasive technologies are limited to a single physical field (such as only analyzing electromagnetic radiation), lacking the ability to perform multi-dimensional modeling of the equipment operating state, making it difficult to trace the root cause of failures and predict the service life. They can only provide fragmented alarm information and cannot support preventive maintenance decisions.

[0006] The above problems indicate that existing monitoring technologies face four major contradictions in the scenario of oil equipment: the contradiction between the non-invasive requirement and signal aliasing interference, the contradiction between the high-dimensional feature requirement and insufficient edge computing power, the contradiction between the real-time requirement and transmission dependence, and the contradiction between the fault coupling and single analysis dimension. Therefore, there is an urgent need for a new monitoring technology for oil equipment that can achieve high-precision signal separation and multi-dimensional feature extraction in a complex electromagnetic environment and complete real-time diagnosis through an embedded terminal, thus breaking through the bottlenecks of reliability, real-time performance, and multi-dimensional analysis capabilities. Summary of the Invention

[0007] To solve the problems mentioned above, the present invention provides a non-invasive equipment operating state monitoring method and system based on electromagnetic spectrum.

[0008] In the first aspect, a non-invasive equipment operating state monitoring method based on electromagnetic spectrum provided by the present invention adopts the following technical solution: A non-invasive equipment operating state monitoring method based on electromagnetic spectrum includes: Dynamic multi-dimensional spectrum fingerprint extraction; Prototype generation based on the extracted multi-dimensional spectrum fingerprints; High-dimensional feature generation through a time series encoder; Intra-prototype contrast learning and inter-prototype contrast learning are respectively performed based on the generated prototypes; Construct a prototype contrast loss function based on contrast learning; Perform image modeling based on the extracted multi-dimensional spectrum fingerprints, and construct an image contrast loss function based on the image modeling; Perform function mixing based on a loss function mixing strategy to obtain an overall loss function; Optimize the parameters of the time series encoder based on the overall loss function; Discriminate the equipment state based on the high-dimensional features generated by the time series encoder; Output the discrimination result.

[0009] Furthermore, the dynamic multi-dimensional spectrum fingerprint extraction includes synchronously collecting the original electromagnetic radiation signals in the frequency band of 0.1 MHz - 3 GHz through the multi-band parallel sampling circuit of the radio frequency sensor at a sampling rate of 200 MHz, covering the fundamental wave, harmonic, and modulation sideband characteristics of the device. Among them, based on the adaptive dynamic range compression method, the dynamic range of the signal amplitude is detected in real time at the embedded terminal, and the dynamic gain is calculated as: G Calculation is expressed as: where is the gain adjustment amount, is the peak value of the signal amplitude within the current sampling window, and the preset target amplitude range is , automatically adjusting the gain of the front-end amplifier to avoid data saturation caused by strong interference signals or loss of weak features.

[0010] Furthermore, aiming at the signal aliasing problem in the complex electromagnetic environment of the industrial site, the dynamic multi-dimensional spectrum fingerprint extraction system innovatively adopts a two-stage anti-aliasing filtering architecture: the first stage uses an analog-domain tunable band-pass filter, and the center frequency is dynamically adjusted by real-time monitoring of the fundamental frequency characteristics of the target device and the quality factor Q value is dynamically calculated: where is the preset bandwidth, defaulting to , and the transfer function of the second-order active band-pass filter can be abbreviated as: where s is the complex frequency variable, is the center angular frequency, K is the passband gain, and the output signal of the first-stage filter is: where are the Laplace transform and its inverse transform respectively.

[0011] The second stage performs time-frequency domain energy threshold segmentation on the aliasing residual components through wavelet packet transform in the digital domain to eliminate the vibration coupling noise from the original electromagnetic signal, expressed as: where, J is the number of wavelet packet decomposition layers, j is the current decomposition layer, k is the frequency band index, is the wavelet packet coefficient, is the time domain form of the bd6 wavelet basis function.

[0012] Furthermore, the dynamic multi-dimensional spectrum fingerprint extraction further includes spatio-temporal alignment of multi-modal signals. Among them, the rising edge of the electromagnetic signal detected by the radio frequency sensor is used as the global trigger source, and the acquisition windows of the infrared thermal imager and the microphone are synchronously started through the FPGA hardware trigger pin to achieve microsecond-level time alignment of cross-modal data. Taking the electromagnetic trigger timestamp as the reference, intercept the time window of the multi-modal signal: where is the temperature signal, is the acoustic vibration signal, and is the time window width.

[0013] Furthermore, the prototype generation based on the extracted multi-dimensional spectrum fingerprints includes randomly generating two different enhanced views for each spectrum fingerprint using each enhancement method in the data enhancement library containing G types of enhancement types to obtain two sets of enhanced view sets; then inputting the enhanced views into the temporal encoder to obtain its high-dimensional hidden representation , : where are the learnable parameters of the time encoder, and the prototype calculation formula is: where , G represents the number of enhancement types, represents the non-linear projection function.

[0014] Furthermore, the in-prototype contrast learning and inter-prototype contrast learning are respectively performed based on the generated prototypes, including achieving a uniform distribution of different spectrum fingerprint prototypes in the representation space through in-prototype contrast learning. Among them, the value controls the penalty intensity of negative samples, and the distance between the j th and the k th spectrum enhancement views and is calculated through the distance metric function , and then the distance is mapped to a value using the softmax function, and the calculation formula is: .

[0015] Furthermore, the prototype contrast loss function constructed based on contrastive learning includes within-prototype contrastive learning and between-prototype contrastive learning, and uses various data augmentation methods to construct a highly generalizable representation and improve the classification performance of downstream tasks. The overall prototype-based contrast loss is defined as: where α is a hyperparameter.

[0016] Furthermore, the image contrast loss function constructed based on image modeling includes obtaining a representation through an image encoder: , and extracting the corresponding representation through a time series encoder ; by screening the information suitable for sequence-image contrastive learning, a more appropriate time series representation and image representation are obtained from and , and the spectral sequence-image contrast loss function is expressed as: where represents the contrast loss between the i th image representation and all spectral sequence representations in the batch, while represents the contrast between the i th spectral sequence and the image representation.

[0017] Furthermore, the overall loss function is obtained by mixing functions based on the loss function mixing strategy, including adopting a mixed loss strategy based on dynamic adaptive weights, adjusting the weight ratio according to the performance of different loss functions during the training process to obtain the overall loss function , and updating the parameters of the time series encoder through backpropagation. The parameter gradient for the update is: where represents the encoder parameters, and the symbol represents the partial derivative, which is used to describe the rate of change of a certain variable in a multivariable function.

[0018] In a second aspect, a non-intrusive device operating state monitoring system based on electromagnetic spectrum includes: A data acquisition module configured to extract dynamic multi-dimensional spectrum fingerprints; A prototype module configured to generate prototypes according to the extracted multi-dimensional spectrum fingerprints; perform within-prototype contrastive learning and between-prototype contrastive learning based on the generated prototypes; construct a prototype contrast loss function based on the contrastive learning; An image module, configured to perform image modeling according to the extracted multi-dimensional spectrum fingerprint, and construct an image contrast loss function based on the image modeling; A mixing module, configured to perform function mixing based on a loss function mixing strategy to obtain an overall loss function; A discrimination module, configured to perform device state discrimination based on the overall loss function; and output a discrimination result.

[0019] In a third aspect, the present invention provides a computer-readable storage medium, in which multiple instructions are stored, and the instructions are adapted to be loaded and executed by a processor of a terminal device to perform the non-intrusive device operating state monitoring method based on electromagnetic spectrum.

[0020] In a fourth aspect, the present invention provides a terminal device, including a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are adapted to be loaded and executed by the processor to perform the non-intrusive device operating state monitoring method based on electromagnetic spectrum.

[0021] In summary, the present invention has the following beneficial technical effects: The technical solution proposed by the present invention has achieved multiple breakthroughs in the field of non-intrusive monitoring, and its significant beneficial effects are reflected in multiple key technical dimensions. In terms of high-precision feature extraction in a complex electromagnetic environment, through the original dynamic multi-dimensional spectrum fingerprint technology, the time-domain transient waveform matching algorithm, the frequency-domain harmonic energy dynamic weighting strategy, and the modulation-domain sideband symmetry depth analysis are organically integrated, successfully overcoming the industry problem of wide-band electromagnetic signal aliasing in the dense scene of petroleum equipment. This core technology has reduced the misjudgment rate of target device feature extraction from the industry average of 30% to less than 5% in a cliff-like manner, and can be achieved only by making full use of the data collected by existing RF sensors, achieving significant improvements in both feature extraction accuracy and economy.

[0022] Aiming at the bottleneck problem of limited edge computing resources, the lightweight prototype enhancement architecture carefully designed in this patent has successfully realized the real-time parsing ability of high-dimensional spectrum fingerprints through feature view calculation optimization innovation on a general ARM / RISC-V processor platform. The end-to-end full-process processing delay is strictly controlled within the technical index of 50 ms, fully meeting the real-time processing performance requirements of high-frequency transient signals (such as 2 ms-level short pulses). Compared with traditional embedded terminal solutions, this technical solution has increased the computing efficiency by up to 8 times, and at the same time completely gets rid of the dependence on dedicated hardware acceleration modules, achieving the best balance between computing performance and deployment cost, providing stable and reliable edge computing technical support for the health monitoring of petroleum equipment.

[0023] In the field of real-time diagnosis under offline conditions, the end-side full-process diagnostic algorithm chain innovatively constructed in this patent perfectly realizes the local closed-loop processing mechanism of signal acquisition, spectral fingerprint extraction, and fault matching. The data volume of a single diagnostic result is optimized and compressed to below the technical standard of 1KB, and the communication bandwidth requirement is reduced by more than 99% year-on-year. This major technological breakthrough enables the system to operate completely autonomously at the fracturing operation site with poor network coverage, completely eliminating the dependence on cloud data transmission, and ensuring the absolute reliability and millisecond-level timeliness of the real-time warning function.

[0024] In addition, the present invention also creatively breaks through the technical limitations of traditional single-field data analysis. Through the advanced algorithm of contrastive learning based on enhanced prototypes, the electromagnetic fingerprint features, infrared temperature field, and acoustic vibration signals are accurately aligned in space and time, and a leading multi-physical-field coupling analysis ability in the industry is constructed. This innovative technology significantly improves the accuracy of equipment operation state recognition by more than 25%. At the same time, it fully exploits the value of existing sensor data, completely avoids additional hardware investment, and provides unprecedented comprehensive technical guarantees for the comprehensive fault diagnosis of oil equipment.

[0025] Generally speaking, the technical solution of this patent not only achieves breakthroughs at the core algorithm level, but also significantly improves the monitoring efficiency, reduces the deployment cost, and enhances the adaptability and reliability of the system in the actual application scenario, providing an innovative solution for the field of oil equipment health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of a non-invasive device operation state monitoring method based on electromagnetic spectrum in Embodiment 1 of the present invention.

[0027] Figure 2 It is a schematic diagram of the method in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] The present invention will be further described in detail below with reference to the accompanying drawings.

[0029] Embodiment 1 Refer to Figure 1 , a non-invasive device operation state monitoring method based on electromagnetic spectrum in this embodiment includes: Dynamic multi-dimensional spectral fingerprint extraction; Prototype generation according to the extracted multi-dimensional spectral fingerprints; High-dimensional feature generation through a time series encoder; Prototype-in contrastive learning and prototype-inter contrastive learning are respectively performed based on the generated prototypes; Construct a prototype contrast loss function based on contrastive learning; Perform image modeling based on the extracted multi-dimensional spectral fingerprints, and construct an image contrast loss function based on the image modeling; Perform function mixing based on the loss function mixing strategy to obtain an overall loss function; Optimize the parameters of the time series encoder based on the overall loss function; Discriminate the device state based on the high-dimensional features generated by the time series encoder; Output the discrimination result.

[0030] Specifically: (1) Definition The spectral sequence data is defined as , where M is the number of data dimensions, T is the number of time steps. When M = 1, the sequence is called a one-dimensional time series. When M > 1, the sequence is called a multi-dimensional time series. In the dense scenario of oil equipment, the broadband electromagnetic signals generated by equipment such as fracturing trucks and frequency converters are superimposed on each other, and the spectral sequence data is M a multi-dimensional time series with X being a frequency band fingerprint sample.

[0031] The spectral fingerprint recognition task aims to assign predefined class labels to each spectral fingerprint, and each label represents a different operating state of the oil equipment. Given a dataset , where each spectral fingerprint , n is the number of samples. The goal of this work is to learn a mapping function that corresponds each spectral fingerprint to the operating state of the oil equipment, C being the total number of classes of the equipment operating state.

[0032] (2) Dynamic multi-dimensional spectral fingerprint extraction The data acquisition module of the present invention is based on existing sensors (including radio frequency sensors, infrared thermal imagers, industrial microphones, etc.), and adopts a multi-channel adaptive synchronous triggering mechanism and an anti-aliasing dynamic sampling strategy to achieve accurate capture of multi-physical field signals.

[0033] (2.1) Broadband electromagnetic signal acquisition Through the multi-band parallel sampling circuit of the radio frequency sensor, synchronously collect the original electromagnetic radiation signals in the frequency band of 0.1 MHz - 3 GHz at a sampling rate of 200 MHz , covering the fundamental wave, harmonic and modulation sideband characteristics of the equipment. Adopt the adaptive dynamic range compression (ADRC) technology to detect the dynamic range of the signal amplitude in real time at the embedded terminal, and the dynamic gain G The calculation formula is as follows: wherein is the gain adjustment amount, is the peak value of the signal amplitude within the current sampling window, and the preset target amplitude range is , automatically adjust the gain of the front-end amplifier (adjustment step ±6dB / μs) to avoid data saturation caused by strong interference signals or loss of weak features.

[0034] The normalized electromagnetic radiation signal is: And based on the operating conditions of the device (such as the starting stage of the fracturing truck motor), dynamically switch the sampling mode: continuous sampling with a 1MHz bandwidth is used in the steady state stage, and the instantaneous sampling rate is increased to 500MHz when triggered by a transient pulse (a current mutation at the 2ms level) to ensure complete capture of high-frequency transient features.

[0035] (2.2) Anti-aliasing and noise suppression, Regarding the problem of electromagnetic signal aliasing in the industrial field, a two-stage anti-aliasing filter is preset at the acquisition end. The first stage uses an analog-domain tunable band-pass filter ( Q The value dynamic range is 5 - 100), and according to the fundamental frequency of the target device (such as the fundamental frequency of the fracturing truck motor is 50Hz), dynamically calculate Q value: wherein is the preset bandwidth, with a default of 10% , and the transfer function of the second-order active band-pass filter can be abbreviated as: wherein s is the complex frequency variable, is the central angular frequency, K is the passband gain, and the output signal of the first-stage filter is: wherein are the Laplace transform and its inverse transform respectively.

[0036] The second stage performs time-frequency domain energy threshold segmentation on the aliasing residual components through wavelet packet transform (selecting the db6 wavelet basis) in the digital domain to eliminate vibration coupling noise from the original electromagnetic signal: wherein J is the number of wavelet packet decomposition layers, j is the current decomposition layer, k is the frequency band index, is the wavelet packet coefficient, It is the time-domain form of the bd6 wavelet basis function.

[0037] (2.3)Spatio-temporal alignment of multi-modal signals For the asynchronous data streams of the infrared thermal imager and the acoustic vibration sensor, the rising edge of the electromagnetic signal detected by the RF sensor (accuracy ±10 μs) is used as the global trigger source. The acquisition windows of the infrared thermal imager (temperature field) and the microphone (acoustic vibration signal) are synchronously started through the FPGA hardware trigger pin to achieve microsecond-level time alignment of cross-modal data. Taking the electromagnetic trigger timestamp as the reference, the time window of the multi-modal signal is intercepted: where is the temperature signal, is the acoustic vibration signal, is the time window width. Output of multi-dimensional spectral fingerprints, fusing the electromagnetic frequency band, temperature, and acoustic vibration characteristics into a unified matrix , providing a reference for subsequent calculations: where represents the m th energy sequence of the electromagnetic frequency band.

[0038] (3)Encoder design Views generated from different samples can capture the characteristics of spectral fingerprints under various transformations. In the present invention, enhanced views of spectral fingerprints are aggregated into prototypes to minimize the potential negative impacts brought by semantic changes that may be introduced by specific enhancements. At the same time, aggregating enhanced views can generate more stable sample representations, reduce the randomness of the representations, and highlight the essential features of spectral fingerprints. For each spectral fingerprint, two different enhanced views are randomly generated using each enhancement method in the data enhancement library containing G types of enhancements, obtaining two sets of enhanced view collections: where and respectively represent the two enhanced views generated by the k th enhancement method using different random parameters for the i th sample .

[0039] Subsequently, the enhanced views are input into the temporal encoder to obtain its high-dimensional hidden representation , : where are the learnable parameters of the time encoder, where is the specific content of model training. Through the loss function designed later, the parameters are corrected to better obtain the hidden representation of the input spectrum fingerprint for the downstream device operation status monitoring task. The encoder of the present invention is a hierarchical Transformer encoder. First, the input is sequence chunked. For convenience of representation, hereinafter is used to represent i.e., a certain enhanced view: where p is the chunk length, s is the sliding step ( s < p ), and after linear projection, is obtained: where is the learnable weight, is the positional encoding. The l th layer of the hierarchical Transformer encoding is: where , and the single-layer calculation details are as follows: After that, the hierarchical outputs are concatenated to obtain R : The hierarchical outputs are passed through max pooling to obtain the hidden representation r : (4) Intra-prototype contrastive learning, To achieve a uniform distribution of different spectrum fingerprint prototypes in the representation space, the present invention designs intra-prototype contrastive learning with adaptive parameters. Through the value controls the penalty intensity of negative samples. Through the distance metric function the distance j between the k th and the th spectrum enhanced views and is calculated, and then the softmax function is used to map the distance to a value . The calculation formula is: Performing contrastive learning on multiple augmented views within a prototype helps prevent views with significantly different amplitudes from dominating the prototype, thereby indirectly optimizing the aggregation process from views to the prototype. To output a low-dimensional representation suitable for contrastive learning, , generate corresponding representations through non-linear projection: For the contrastive loss within the prototype is defined as: where and . Before calculating , set to negative infinity to ensure that positive sample pairs are close to each other.

[0040] (5) Inter-prototype contrastive learning, The prototype calculation formula is: where , G represents the number of augmentation types, represents the non-linear projection function: where represents the input, represents the first-layer weight, represents the first-layer bias vector, represents the second-layer weight, represents the second-layer bias vector, represents the activation function.

[0041] For the i th sample, and are positive samples of each other, while they regard the prototypes of other samples in the batch as negative samples, the inter-prototype contrastive loss of is defined as: B where

[0042] represents the batch size. Through the above two training objectives, our prototype-based contrastive learning method effectively utilizes various data augmentation means to construct highly generalizable representations, thereby improving the classification performance of downstream tasks. The overall prototype-based contrastive loss is defined as: where α is a hyperparameter.

[0043] (6) Spectrum fingerprint image modeling, Spectrum fingerprint mode enhancement based on numerical modeling is difficult to capture the structural information crucial for class recognition. Time series data in different fields is usually composed of line segments or curve segments. Therefore, compared with numerical values, capturing the structural information of spectrum fingerprints based on shape is more direct. To capture the structural features of spectrum fingerprints, the present invention introduces an image mode and proposes sequence-image contrast learning using a hybrid strategy to simultaneously obtain numerical and structural information.

[0044] First, each time series sample is converted into an image. Subsequently, the time series sample and its corresponding image form a positive sample pair, while the time series and images of other samples in the same batch are used as negative sample pairs. Although basic contrast learning can establish the correspondence between time series and images, since the negative samples do not consider the numerical characteristics of the time series, such differences are still insufficient to distinguish different time series samples. For this reason, we designed a geodesic hybrid method to generate hybrid representations located between the subspace representations of the two modes. Using these hybrid representations as negative samples for contrast learning can expand the effective subspace of the learned representations and make the samples easier to classify.

[0045] Visualizing spectrum fingerprint data through a line graph is a natural intuition for converting numerical data into the image mode. In the line graph, x the x-axis represents the timestamp, y and the y-axis represents the numerical value. The present invention uses the symbol "*" to mark the observed data points and connects them with straight lines. For different dimensions of the spectrum fingerprint sample , since each variable has a different dimension, a line graph is plotted for each variable separately, and the images of different variables are normalized to the same square size. In addition, different variables are identified with different colors, and the sub-images of each variable are finally stitched together into a complete image, denoted as , where H , W represents the normalized square size, C represents the number of channels. Subsequently, the representation is obtained through an image encoder: where the structure of the image encoder is: where is the convolutional kernel weight, is the convolutional operation, is the activation function, is the convolutional bias vector.

[0046] Global pooling is expressed as: (7) Image contrast loss function, Extract the corresponding representation through the time series encoder . Due to the characteristics of different modalities, there is incomparable information between the time series and the image modality. For example, the contrast of an image has nothing to do with the inherent attributes of time series data, and such information is not suitable for cross-modal contrast learning. Therefore, to filter out the incomparable information, the present invention performs non-linear projection on the representations of each modality so that they can be compared during the training process. By screening the information suitable for sequence-image contrast learning, more appropriate time series representations and are obtained from and image representations . The sequence-image contrast loss of the i th sample in the batch can be expressed as: where represents the contrast loss between the i th image representation and all spectral sequence representations in the batch, while represents the contrast between the i th spectral sequence and the image representation. The spectral sequence-image contrast loss function is defined as: (8) Loss function mixing strategy, During the model training process, the mixed loss function needs to comprehensively consider the balance relationship between different optimization objectives in the design of the mixed loss function. The present invention adopts a mixed loss strategy based on dynamic adaptive weights, which can automatically adjust the weight ratio according to the performance of different loss functions during the training process, thereby effectively improving the overall performance of the model: where is the overall loss function, is a dynamic weight coefficient that changes with the time step, and is defined where the relative convergence degree of the loss is : indicates loss stagnation, indicates loss decline, indicates loss oscillation.

[0047] Dynamically adjust the loss weight coefficient based on the convergence degree: where is used to dynamically adjust the sensitivity, indicates the number of loss terms, which is 2 here, is , For .

[0048] (9) Classification of equipment operation status Based on the overall loss function , the parameters of the time series encoder are updated through backpropagation , and the parameter gradient for update is where represents the encoder parameters, and the symbol represents the partial derivative, which is used to describe the rate of change of a certain variable in a multivariable function.

[0049] Chain rule decomposition where is the prototype representation is the total number of samples, and the parameter update rule is where is the learning rate t is the number of training steps

[0050] For the task of classifying the operation status of oil equipment, the present invention finally uses the time series encoder with optimized parameters to obtain a high-dimensional embedding as the input of the classifier, and finally obtains the equipment operation status where represents the input spectral fingerprint X the high-dimensional feature vector extracted by the feature encoding network r , C is the total number of status categories. Taking three key equipment, namely beam pumping unit, centrifugal pump and drilling motor, as examples, their common operation status classifications are as follows

[0051] Beam Pumping Unit, Normal status: The energy distribution in each frequency band is balanced, and the characteristic frequency amplitude is stable; Crankpin Wear: Sideband modulation phenomenon appears in the 630 - 850 Hz frequency band; Gear Tooth Breakage: High-order harmonics appear at the meshing frequency; Horsehead Misalignment: The energy at the 1 / 3 multiple frequency increases abnormally

[0052] Centrifugal Pump, Normal state: The amplitude of the Blade Passing Frequency (BPF) conforms to the design threshold; Cavitation: Continuous broadband noise appears in the high-frequency band (>5 kHz); Bearing Inner Race Defect: Impact peaks appear at the characteristic frequency (BPFI); Impeller Imbalance: The amplitude of 1X rotational frequency exceeds 200% of the baseline.

[0053] Drilling Motor, Normal state: The amplitude of the electromagnetic characteristic frequency (such as slot harmonics) is stable; Broken Rotor Bar: Generates a characteristic frequency component of (1±2s)f_s; Stator Winding Short: Abnormal enhancement of the 3rd harmonic appears Shaft Misalignment: Generates 2X rotational frequency and combined frequency components. The cosine similarity calculation in the formula effectively eliminates the deviation caused by the difference in feature amplitudes, making the classification result only depend on the similarity of directions in the feature space.

[0054] Figure 2 In it is The prototype contrastive learning loss function, is the within-prototype contrastive loss function, is the between-prototype contrastive loss function, is the overall loss function, is the image sequence contrastive loss function, represents the contrastive loss between the image representation and all spectral sequence representations in the batch, while represents the contrastive loss between the spectral sequence and the image representation.

[0055] Experimental verification: Table 1 Comparison of prediction effects As shown in Table 1, based on the integration of dynamic spectrum fingerprint technology and lightweight architecture optimization, the proposed algorithm achieves a breakthrough in the accuracy of multi-device fault diagnosis while maintaining low memory occupancy. Compared with the RAM occupancy of up to 70.5 - 91.3 MB of the traditional Transformer model, this solution reduces the memory requirement to 18.7 - 38.5 MB (a decrease of 73.4% - 58.3%) through the feature view compression mechanism. At the same time, it achieves an accuracy of 0.9395 and an R2 determination coefficient of 0.9942 on the pumping unit equipment, which is a 13.5% improvement compared to the sub-optimal Transformer model. Especially in the scenario of drilling motors with limited computing resources, the proposed method realizes a significant 24.51% improvement in the fault diagnosis accuracy relative to the baseline model at the cost of only a 15% increase in RAM, verifying the efficiency and robustness advantages of the algorithm in complex industrial environments.

[0056] Embodiment 2 This embodiment provides a system, including: A data acquisition module, configured to: A computer-readable storage medium, in which multiple instructions are stored, and the instructions are adapted to be loaded and executed by a processor of a terminal device for the non-intrusive device operating state monitoring method based on electromagnetic spectrum.

[0057] A terminal device, including a processor and a computer-readable storage medium, the processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are adapted to be loaded and executed by the processor for the non-intrusive device operating state monitoring method based on electromagnetic spectrum.

[0058] The above are all the preferred embodiments of the present invention. Without limiting the protection scope of the present invention accordingly, therefore: All equivalent changes made according to the structure, shape, and principle of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A non-invasive method for monitoring the operating state of a device based on the electromagnetic spectrum, characterized in that, Including: Dynamic multi-dimensional spectrum fingerprint extraction; Prototype generation based on the extracted multi-dimensional spectrum fingerprints; High-dimensional feature generation through a temporal encoder; Intra-prototype contrast learning and inter-prototype contrast learning are respectively performed based on the generated prototypes; Construct a prototype contrast loss function based on contrast learning; Perform image modeling based on the extracted multi-dimensional spectrum fingerprints, and construct an image contrast loss function based on the image modeling; Perform function mixing based on a loss function mixing strategy to obtain an overall loss function; Optimize the parameters of the temporal encoder based on the overall loss function; Perform device status discrimination based on the high-dimensional features generated by the temporal encoder; Output the discrimination result.

2. The non-invasive device operating state monitoring method based on electromagnetic spectrum according to claim 1, wherein The dynamic multi-dimensional spectrum fingerprint extraction includes a multi-band parallel sampling circuit of a radio frequency sensor to synchronously collect the original electromagnetic radiation signals in the frequency band of 0.1 MHz - 3 GHz at a sampling rate of 200 MHz , covering the fundamental wave, harmonic wave and modulation sideband characteristics of the device; among them, based on the adaptive dynamic range compression method, the dynamic range of the signal amplitude is detected in real time at the embedded terminal, and the dynamic gain G is calculated as follows: wherein is the gain adjustment amount, is the peak value of the signal amplitude within the current sampling window, and the preset target amplitude range is to automatically adjust the gain of the front-end amplifier to avoid data saturation caused by strong interference signals or loss of weak features.

3. The non-invasive device operation status monitoring method based on electromagnetic spectrum according to claim 2, wherein, The dynamic multi-dimensional spectrum fingerprint extraction further includes, aiming at the problem of electromagnetic signal aliasing in industrial sites, presetting a two-stage anti-aliasing filter at the acquisition end. The first stage uses an analog-domain tunable band-pass filter, and the center frequency is dynamically adjusted by real-time monitoring of the fundamental frequency characteristics of the target device. The quality factor Q value is dynamically calculated. The second stage performs time-frequency domain energy threshold segmentation on the aliasing residual components through wavelet packet transform in the digital domain to eliminate vibration coupling noise from the original electromagnetic signal, which is expressed as: Among them, J is the number of wavelet packet decomposition levels, j is the current decomposition level, and k is the frequency band index. is the wavelet packet coefficient. is the time-domain form of the bd6 wavelet basis function.

4. A method for monitoring the operating state of a non-invasive device based on electromagnetic spectrum according to claim 3, characterized in that, Including: The dynamic multi-dimensional spectrum fingerprint extraction further includes spatio-temporal alignment of multi-modal signals. Among them, the rising edge of the electromagnetic signal detected by the RF sensor is used as the global trigger source, and the acquisition windows of the infrared thermal imager and the microphone are synchronously started through the FPGA hardware trigger pin to achieve microsecond-level time alignment of cross-modal data. Taking the electromagnetic trigger timestamp as the reference, the time window of the multi-modal signal is intercepted: where is the temperature signal, is the acoustic vibration signal, is the time window width.

5. A method for monitoring the operating state of a non-invasive device based on electromagnetic spectrum according to claim 4, characterized in that, Including: The prototype generation based on the extracted multi-dimensional spectrum fingerprints includes randomly generating two different augmented views for each spectrum fingerprint using each augmentation method in a data augmentation library containing G types of augmentation types, obtaining two sets of augmented view sets; Then input the enhanced view into the temporal encoder to obtain its high-dimensional hidden representation , : where are the learnable parameters of the time encoder, The prototype calculation formula is: Among them , G represents the number of enhancement types, represents a non-linear projection function.

6. A method for monitoring the operating state of a non-invasive device based on electromagnetic spectrum according to claim 5, characterized in that, Including: The prototype-based in-prototype contrast learning and inter-prototype contrast learning are respectively performed, including achieving uniform distribution of different spectral fingerprint prototypes in the representation space through in-prototype contrast learning, where the value controls the penalty intensity of negative samples, and through the distance metric function calculate the distance between the j-th and k-th spectral enhancement views and and then use the softmax function to map the distance to a value, The calculation formula is: The calculation formula is: 。 7. A method for monitoring the operating state of a non-invasive device based on electromagnetic spectrum according to claim 6, characterized in that, Including: The construction of the prototype contrast loss function based on contrast learning includes intra-prototype contrast learning and inter-prototype contrast learning, using a variety of data augmentation means to construct a strongly generalized representation and improve the classification performance of downstream tasks. The overall prototype-based contrast loss is defined as: where α is a hyperparameter.

8. A non-invasive device operating state monitoring method based on electromagnetic spectrum according to claim 7, characterized in that, Including: The image contrast loss function constructed based on image modeling includes obtaining representations through an image encoder: , and extracting corresponding representations through a time series encoder ; By screening information applicable to sequence-image contrast learning, a more suitable time series representation and is obtained from and an image representation . The spectral sequence-image contrast loss function is expressed as: Among them, represents the contrast loss between the i-th image representation and all spectral sequence representations in the batch, while represents the contrast between the i-th spectral sequence and the image representation.

9. The non-invasive device operating state monitoring method based on electromagnetic spectrum according to claim 8, characterized in that, Including: The overall loss function is obtained by mixing functions based on a loss function mixing strategy, including adopting a hybrid loss strategy based on dynamic adaptive weights, and adjusting the weight ratio according to the performance of different loss functions during the training process to obtain the overall loss function , through the overall loss function time series encoder The update of the parameters is as follows: the parameter gradient for the update is: Among them represents the encoder parameter, and the symbol represents the partial derivative, which is used to describe the rate of change of a certain variable in a multivariable function.

10. A non-invasive device operating status monitoring system based on the electromagnetic spectrum, characterized in that, Including: A data acquisition module, configured to perform dynamic multi-dimensional spectrum fingerprint extraction; A prototype module, configured to perform prototype generation based on the extracted multi-dimensional spectrum fingerprints; perform intra-prototype contrast learning and inter-prototype contrast learning respectively based on the generated prototypes; Construct a prototype contrast loss function based on contrast learning; An image module, configured to perform image modeling based on the extracted multi-dimensional spectrum fingerprints, and construct an image contrast loss function based on the image modeling; A mixing module, configured to perform function mixing based on a loss function mixing strategy to obtain an overall loss function; A discrimination module, configured to perform device status discrimination based on the overall loss function; Output the discrimination result.

Images