A device and method for diagnosing the health status of a vibrating wire sensor

Abstract

This application discloses a device and method for diagnosing the health status of a vibrating wire sensor, relating to the field of sensor status diagnosis technology. The device includes a main processor, a vibration signal generation module, a linear amplifier circuit, and a saturation amplifier circuit. The vibration signal generation module generates an excitation modulation signal to cause the vibrating wire of the vibrating wire sensor to vibrate. The vibration signal of the vibrating wire sensor is acquired by a signal processing unit. The linear amplifier circuit and the saturation amplifier circuit obtain a linear amplified signal and multiple harmonic signals. The main processor performs time-domain and frequency-domain signal processing on the linear amplified signal and multiple harmonic signals to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal. Based on the time-domain and frequency-domain signals, SNR value, and duration of each harmonic signal in the linear amplified signal and multiple harmonic signals, the health status of the vibrating wire sensor is determined and diagnosed. This application realizes automatic diagnosis of the health status of vibrating wire sensors.

Description

Technical Field

[0001] This application relates to the field of sensor condition diagnosis technology, and in particular to a device and method for diagnosing the health status of a vibrating wire sensor. Background Technology

[0002] Vibrating wire sensors are a type of non-electrical quantity measurement sensor that is widely valued and used both domestically and internationally. Because vibrating wire sensors directly output the natural frequency signal of the vibrating wire, they possess characteristics such as strong anti-interference ability, low susceptibility to electrical parameters, small zero-point drift, low temperature susceptibility, stable and reliable performance, vibration resistance, and long lifespan. Vibrating wire sensors are widely used in long-term monitoring projects in engineering and scientific research.

[0003] The proper installation method and operational status of the vibrating wire sensor are crucial for its application. However, currently, there is a lack of devices capable of automatically diagnosing the health status of vibrating wire sensors. Summary of the Invention

[0004] The purpose of this application is to provide a device and method for diagnosing the health status of a vibrating wire sensor, which can realize automatic diagnosis of the health status of the vibrating wire sensor.

[0005] To achieve the above objectives, this application provides the following solution.

[0006] In a first aspect, this application provides a health status diagnostic device for a vibrating wire sensor, including a main processor, a vibration signal generation module, a linear amplifier circuit, and a saturation amplifier circuit.

[0007] The vibration signal generation module includes a signal processing unit and a vibrating wire sensor, and is used to: generate an excitation modulation signal through the signal processing unit to make the vibrating wire sensor vibrate, and acquire the vibrating wire vibration signal of the vibrating wire sensor through the signal processing unit, and preprocess the vibrating wire vibration signal to obtain the fundamental wave signal.

[0008] The linear amplifier circuit is used to linearly amplify the fundamental signal to obtain a linearly amplified signal.

[0009] The saturated amplifier circuit is used to amplify the linear amplified signal multiple times to obtain multiple harmonic signals.

[0010] The main processor is used to: perform time-domain and frequency-domain signal processing on the linear amplified signal and multiple harmonic signals to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal; and, based on the time-domain and frequency-domain signals, SNR value, and duration of each harmonic signal in the linear amplified signal and multiple harmonic signals, determine and diagnose the health status of the vibrating wire sensor to obtain the health status diagnosis result of the vibrating wire sensor.

[0011] Secondly, this application provides a method for diagnosing the health status of a vibrating wire sensor based on the vibrating wire sensor health status diagnosis device described in the first aspect, comprising the following steps.

[0012] An excitation modulation signal is generated by a signal processing unit to make the vibrating wire sensor vibrate. The vibrating wire vibration signal of the vibrating wire sensor is acquired by the signal processing unit, and the vibrating wire vibration signal is preprocessed to obtain the fundamental signal.

[0013] The fundamental signal is linearly amplified to obtain a linearly amplified signal.

[0014] By amplifying the linear amplified signal multiple times, multiple harmonic signals are obtained.

[0015] The linear amplified signal and multiple harmonic signals are processed to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal. Based on the time-domain and frequency-domain signals, SNR values, and durations of each harmonic signal in the linear amplified signal and multiple harmonic signals, the health status of the vibrating wire sensor is determined and diagnosed, resulting in a health status diagnosis result for the vibrating wire sensor.

[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a device and method for diagnosing the health status of a vibrating wire sensor. An excitation modulation signal is generated by a signal processing unit to cause the vibrating wire of the vibrating wire sensor to vibrate. The vibrating wire vibration signal of the vibrating wire sensor is collected by the signal processing unit, preprocessed to obtain a fundamental signal, and then a linearly amplified signal and multiple harmonic signals are obtained through a linear amplification circuit and a saturation amplification circuit. The main processor processes the linearly amplified signal and multiple harmonic signals to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal. The above-mentioned time-domain and frequency-domain signals, along with the aforementioned physically meaningful data such as the SNR and duration, are integrated and synchronously input into the main processor to determine and diagnose the health status of the vibrating wire sensor, obtaining the health status diagnosis result. This application obtains the health status diagnosis result from the vibration signal waveform of the vibrating wire sensor, realizing automatic diagnosis of the vibrating wire sensor's health status. This allows for the determination of the rationality of the sensor's installation method and information such as the sensor's operating status, providing support for improving the working performance and fault diagnosis of the vibrating wire sensor. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the structure of a vibrating wire sensor health status diagnosis device provided in an embodiment of this application.

[0019] Figure 2 This is a flowchart illustrating the vibration waveform diagnosis steps provided in an embodiment of this application.

[0020] Figure 3 This is a schematic diagram of multiple harmonic signals provided in an embodiment of this application.

[0021] Figure 4 This is a schematic diagram of a neural network model for diagnosing the health status of a vibrating wire sensor, provided as an embodiment of this application.

[0022] Figure 5 This is a flowchart illustrating a method for diagnosing the health status of a vibrating wire sensor, provided in one embodiment of this application.

[0023] Figure reference numerals: Main processor—10, Excitation circuit—11, Vibrating wire sensor interface circuit—12, Coil—13, Magnetic core—14, Vibrating wire sensor—15, Vibration pickup circuit—16, Linear amplifier circuit—171, Saturated amplifier circuit—172, Dual-channel waveform data acquisition module—18, Power supply module—19, Display module—20. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] In one exemplary embodiment, such as Figure 1 As shown, a vibrating wire sensor health status diagnosis device is provided, including a main processor (MCU) 10, a vibration signal generation module, a linear amplifier circuit 171, and a saturation amplifier circuit 172.

[0027] The vibration signal generation module includes a signal processing unit and a vibrating wire sensor 15, and is used to: generate an excitation modulation signal through the signal processing unit to make the vibrating wire sensor 15 vibrate, and collect the vibrating wire vibration signal of the vibrating wire sensor 15 through the signal processing unit, and preprocess the vibrating wire vibration signal to obtain the fundamental wave signal.

[0028] The linear amplifier circuit 171 is used to linearly amplify the fundamental signal to obtain a linearly amplified signal.

[0029] The saturated amplifier circuit 172 is used to amplify the linear amplified signal multiple times to obtain multiple harmonic signals.

[0030] The main processor 10 is used to: process the linearly amplified signal and multiple harmonic signals to obtain the signal-to-noise ratio value and the duration of each harmonic signal, and to determine and diagnose the health status of the vibrating wire sensor 15 based on the time-domain and frequency-domain signals, the signal-to-noise ratio value and the duration of each harmonic signal in the linearly amplified signal and multiple harmonic signals, and to obtain the health status diagnosis result of the vibrating wire sensor 15.

[0031] The above-mentioned time-domain and frequency-domain signals, as well as the aforementioned data such as signal-to-noise ratio and duration with clear physical meaning, are integrated and synchronously input into the main processor. The neural network deployed on the main processor and already trained is used for processing, thereby realizing the determination and diagnosis of the health status of the vibrating wire sensor 15 and obtaining the health status diagnosis result of the vibrating wire sensor 15.

[0032] By integrating the linearly amplified signal and the multiple harmonic signals, as well as the aforementioned data such as signal-to-noise ratio and duration with clear physical meaning, and synchronously inputting them into the trained neural network, the health status of the vibrating wire sensor is determined and diagnosed. This process maximizes the utilization of the time-domain and frequency-domain signals while simultaneously integrating the index data with clear physical meaning, enabling the neural network model to determine and diagnose the health status of the vibrating wire sensor under physical constraints, resulting in higher accuracy and efficiency.

[0033] This application fully considers engineering practice needs and application scenarios, providing an efficient and accurate method for diagnosing the health status of vibrating wire sensors. Supported by neural network technologies, it includes steps such as data collection and preprocessing, feature engineering, neural network model construction, model training, and verification. The neural network training process is as follows.

[0034] Step 1: Data Collection and Preprocessing: Before constructing the health status diagnostic model for vibrating wire sensors, it is necessary to collect a large amount of vibration time history data of different models of vibrating wire sensors under different excitation conditions, as well as sensor attribute data, including manufacturer, production batch, sensor product model, vibrating wire cross-sectional diameter, vibrating wire length, electromagnetic parameters of the excitation coil, excitation signal transmission distance, and whether there is a cable connection. Specifically, data on various common working states and fault conditions of vibrating wire sensors are collected, including health status, number of jumps, axial tensile stress exceeding the limit, axial compressive stress exceeding the limit, weld point detachment, insecure installation of protective shell, sensor installation angle not parallel to the pipeline axis, insecure installation of protective shell, stray current interference, excessive fatigue of vibrating wire steel wire, and poor contact of signal cable.

[0035] This embodiment collects data on BGK4100 vibrating wire sensors under various conditions, including healthy status, high jump count, solder joint detachment, sensor installation angle not parallel to the pipeline axis, stray current interference, and poor signal cable contact. The production batch, vibrating wire cross-sectional diameter, vibrating wire length, electromagnetic parameters of the excitation coil, and excitation signal transmission distance of the collected sensors are recorded. All collected sensors are processed to the original factory cable length, and there are no cable splices.

[0036] Sensors were categorized according to their health status. Vibration data for sensor excitation was sampled with at least 400 samples for each operating state, resulting in a total of 1000 samples from healthy sensors. This data was compiled into dataset D-0. Each vibration time history data sample was labeled with the corresponding vibrating wire sensor state. The categorized data in dataset D-0 were then one-hot encoded and converted into pure numerical values ​​while retaining essential information, allowing the neural network model to read them successfully, thus obtaining dataset D-1.

[0037] In the dataset used for model training, the data labels can be set to healthy / unhealthy according to the actual situation, or they can be set to more specific fault states, such as number of jumps, range exceeding limits, insecure installation of protective housing, sensor installation angle not parallel to the pipeline axis, fatigue exceeding limits of vibrating wire steel wire, poor contact of signal cable, etc.

[0038] II. Feature Engineering: The model architecture adopts a modular construction approach. This application only provides a typical model architecture example. Different parts of the model should be replaceable according to requirements.

[0039] Frequency domain analysis was performed on each collected vibration time history data sample, simultaneously acquiring the corresponding signal-to-noise ratio and the duration of different harmonic signals. The newly acquired analysis results were then integrated and inserted into dataset D-1. This ensured that the acquired dataset was within the constraints of professional theoretical conditions, without losing highly important or potentially deep information that could negatively impact model training. For numerical data in different fields of dataset D-1, standardization was performed. Data of different scales and dimensions were scaled to the same data interval and range to reduce the impact of differences in scale, features, and distribution on model performance, accelerate model fitting, and avoid overfitting. Considering that data from unhealthy sensor states might exhibit abnormal data distributions, linear scaling, logarithmic functions, and exponential functions were used during data processing to adjust the data distribution specifically, ensuring effective model training.

[0040] Then, the dataset DB-1 is split into a training set DB-1a, a validation set DB-1b, and a test set DB-1c in a 7:2:1 ratio. The training set is used to train the model, the validation set is used to tune the model's hyperparameters, and the test set is used to evaluate the model's performance. To prevent data leakage from affecting the model's predictive performance, there should be no overlap between the training, validation, and test sets.

[0041] III. Neural Network Model Construction: Combining data characteristics with the health diagnosis objectives of vibrating wire sensors, the neural network model is constructed using a partial construction strategy for different data types and their characteristics.

[0042] For vibration time history data samples, RNN network components are preferred for modeling; for vibration time history data samples with one-hot encoding, CNN series network components are used for modeling; and for vibration time history data samples with remaining fields, FNN is used for modeling. The number of layers in the models built for different parts can be stacked according to characteristics such as data complexity and data volume. For example, if the FNN model has a deep number of layers, residual connection components may be added as appropriate. After the partial models are built, they are all aggregated into a single second-level data processing neural network. Considering the balance between data volume and model size, power supply and computing power conditions for field testing, and diagnostic efficiency, the Transformer architecture is excluded for this part of the network, and CNN or FNN architectures are preferred. The output settings of this part of the model should be related to the different state types of the vibrating wire sensor mentioned above. After the partial models are built, they are all aggregated into the same CNN second-level data processing neural network, and the output of the second-level network is 6-dimensional.

[0043] IV. Model Training and Validation: Initialize the aforementioned neural network model, select vibration time history data samples from the working state type field of the vibrating wire sensor in the DB-1a and DB-1b datasets as sample labels, and input the data of the remaining fields according to the data entry corresponding to the model.

[0044] The training set DB-1a is used for model training, and the validation set DB-1b is used for interim performance validation. During validation, parameters with better model performance are retained and replaced. Considering the application scenario, cross-entropy loss is used as the loss function in the final part of the model. Based on past experience, if certain types of faults in the vibrating wire sensor appear similar and difficult to distinguish in the model during training, parallel paths can be added for judgment to enhance the model's capabilities. During interim performance validation, parameters with better model performance are retained and replaced. During training, care must be taken to ensure that the model does not overfit.

[0045] After model training is complete, the generalization ability of the model is validated using the test set DB-1c. If there is a significant difference between the model's score on the test set DB-1c and the validation set DB-1b, it is necessary to determine whether the model is undertrained or overtrained, and take targeted measures to address this. As mentioned earlier, if certain types of faults in the vibrating wire sensor behave similarly in the model and are difficult to distinguish, it is advisable to lock the original model parameters, add parallel structural paths for judgment, and enhance the model's capabilities through incremental training.

[0046] Through the above process, this application utilizes the learning ability and powerful expressive power of neural network models to deeply and accurately mine and learn the correlations and influencing laws related to the health status of vibrating wire sensors, thereby achieving the goal of diagnosing the health status of vibrating wire sensors and obtaining a trained neural network model. The neural network model is an RNN model, a CNN model, or an FNN model.

[0047] The signal processing unit includes an excitation circuit 11, a vibrating wire sensor interface circuit 12, and a vibration pickup circuit 16.

[0048] The excitation circuit 11 is used to amplify and modulate the initial excitation signal to generate an excitation modulation signal; the initial excitation signal is generated by the main processor 10. The main processor 10 is used to control the excitation, that is, to generate the initial excitation signal and send it to the excitation circuit 11.

[0049] The vibrating wire sensor interface circuit 12 is used to connect the vibrating wire sensor 15, the excitation circuit 11, and the vibration pickup circuit 16; the vibrating wire sensor interface circuit 12 is a transistor. The transistor can be a MOSFET.

[0050] The vibration pickup circuit 16 is used to: pick up the vibrating wire vibration signal of the vibrating wire sensor 15, and filter the vibrating wire vibration signal to obtain the fundamental wave signal.

[0051] The excitation circuit 11, the vibrating wire sensor interface circuit 12, and the vibration pickup circuit 16, in conjunction with the coil 13 and the magnetic core 14, excite and pick up the vibrating wire vibration signal of the vibrating wire sensor 15. The vibrating wire sensor interface circuit 12 can switch the on / off state of the excitation circuit 11 and the vibration pickup circuit 16. During excitation, the vibrating wire sensor interface circuit 12 receives the excitation modulation signal sent by the excitation circuit 11. The vibrating wire sensor interface circuit 12 is connected to the coil 13. After the coil 13 is excited, the vibrating wire sensor interface circuit 12 is disconnected from the excitation circuit 11. At this time, the vibrating wire sensor interface circuit 12 is connected to the vibration pickup circuit 16, which picks up the vibrating wire vibration signal of the vibrating wire sensor 15 and further filters the vibrating wire vibration signal to obtain the fundamental signal.

[0052] After the excitation ends, the vibration signal of the vibrating string induced by the coil 13 and the magnetic core 14 is collected by the vibration pickup circuit 16. Then, the vibration pickup circuit 16 filters the vibrating string vibration signal to obtain the fundamental signal, which is then input into the linear amplifier circuit 171.

[0053] Linear amplifier circuit 171: linearly amplifies the fundamental signal output by the vibration pickup circuit 16 to obtain a linearly amplified signal.

[0054] Saturated amplifier circuit 172: After amplifying the linearly amplified signal multiple times, the amplifier enters a saturation distortion state, obtaining multiple harmonic signals of the saturated frequency harmonic waveform. Saturated amplifier circuit 172 is a power amplifier; it amplifies the acquired signal, and when the signal exceeds the linear operating range of the power amplifier, it acquires harmonic waveform data including those caused by saturation distortion.

[0055] like Figure 3 The diagram shows the fundamental frequency, second harmonic, third harmonic, and fourth harmonic obtained by the saturated amplifier circuit 172.

[0056] The vibrating wire sensor health status diagnostic device also includes a dual-channel waveform data acquisition module 18. This module 18 acquires linearly amplified signals and multiple harmonic signals, and transmits them to the main processor 10. The dual-channel waveform data acquisition module 18 has two channels that respectively input signals from two stages of amplifiers, including one linearly amplified signal and one signal with amplifier saturation distortion (i.e., multiple harmonic signals). The dual-channel waveform data acquisition module 18 is a waveform data acquisition sound card.

[0057] The vibrating wire sensor health status diagnosis device also includes a power supply module (v) 19, which provides electrical energy to power the entire device.

[0058] The main processor 10 includes a first signal processing unit, a second signal processing unit, a third signal processing unit, and a diagnostic unit. The main processor 10 utilizes a vibration spectrum diagnostic analysis algorithm to provide physical constraints with professional theoretical guidance for the neural network diagnostic model. The vibration spectrum diagnostic analysis algorithm includes a fundamental signal-to-noise ratio calculation method and a method for calculating the duration of each harmonic of the saturation harmonics. The specific calculation process is as follows.

[0059] The first signal processing unit is configured to: perform a Fourier transform on the linearly amplified signal to obtain a first time-domain signal and a first frequency-domain signal; and perform full-spectrum waveform analysis on the first frequency-domain signal to obtain the center frequency waveform power.

[0060] The second signal processing unit is configured to: perform Fourier transform on the multiple harmonic signals to obtain a second time-domain signal and a second frequency-domain signal; and perform full-spectrum waveform analysis on the second frequency-domain signal to obtain noise power and the duration of each harmonic signal.

[0061] The third signal processing unit is used to calculate the signal-to-noise ratio based on the center frequency waveform power and noise power.

[0062] In calculating the signal-to-noise ratio (SNR) value based on the center frequency waveform power and noise power, the third signal processing unit is further configured to: divide the center frequency waveform power and noise power to obtain a power ratio; perform logarithmic processing on the power ratio to obtain a logarithmic processing result; and multiply the logarithmic processing result by 10 to obtain the SNR value.

[0063] The diagnostic unit is used to: input the first time-domain signal, the first frequency-domain signal, the second time-domain signal, the second frequency-domain signal, the signal-to-noise ratio value, and the duration of each harmonic signal into the trained neural network model to determine and diagnose the health status of the vibrating wire sensor 15, and obtain the health status diagnosis result of the vibrating wire sensor 15.

[0064] Neural network model structure as follows Figure 4 As shown, its inputs are the signal-to-noise ratio, the duration of each harmonic signal, frequency domain data, and time domain data, and the output is the health status diagnosis result of the vibrating wire sensor 15.

[0065] By integrating the aforementioned time-domain and frequency-domain signals, as well as the aforementioned data such as signal-to-noise ratio and duration with clear physical meaning, a trained neural network is used for data processing and state determination, thereby realizing the determination and diagnosis of the health status of the vibrating wire sensor 15 and obtaining the health status diagnosis result of the vibrating wire sensor 15.

[0066] like Figure 2 As shown, the vibration fundamental wave (i.e., the linearly amplified signal) and noise spectrum data (i.e., multiple harmonic signals) are acquired by the dual-channel waveform data acquisition module 18. By performing Fourier transforms on the linearly amplified signal and the multiple harmonic signals respectively, the transformed waveforms, namely the first frequency domain signal and the second frequency domain signal, are obtained. Full-spectrum waveform analysis is performed on the first frequency domain signal and the second frequency domain signal to obtain the center frequency waveform power, noise power, and duration of each harmonic signal.

[0067] The ratio of the center frequency waveform power to the noise power is taken as the logarithm of 10 as the signal-to-noise ratio (SNR) value. The higher the value, the higher the signal quality output by the vibrating wire sensor 15. In this application, this index data serves as a component of the physical condition constraints of the neural network diagnostic model, carrying information about the sensor installation method or sensor operating status.

[0068] The amplified full-spectrum data is acquired by the dual-channel waveform data acquisition module 18 to obtain the duration of each harmonic. According to existing research, a longer harmonic duration generally indicates better sensor stability. In this application, this indicator data is also used as a component of the physical condition constraints of the neural network diagnostic model, carrying some sensor operating status information to support the neural network model in performing health status diagnosis of the vibrating wire sensor.

[0069] The vibrating wire sensor health status diagnosis device also includes a display module (LCD) 20, which is used to display the analysis and evaluation results and key indicator data of the main processor 10; the analysis and evaluation results and key indicator data include the signal-to-noise ratio value, the duration of each harmonic signal and the health status diagnosis results.

[0070] Based on the same inventive concept, this application also provides a method for diagnosing the health status of a vibrating wire sensor based on the aforementioned vibrating wire sensor health status diagnosis device. The solution provided by this method is similar to the solution described in the aforementioned device. Therefore, the specific limitations in one or more embodiments of the vibrating wire sensor health status diagnosis method provided below can be found in the limitations of the vibrating wire sensor health status diagnosis device described above, and will not be repeated here.

[0071] In one exemplary embodiment, such as Figure 5As shown, a method for diagnosing the health status of a vibrating wire sensor based on the above-mentioned vibrating wire sensor health status diagnosis device is provided, including the following steps 201 to 204.

[0072] Step 201: Generate an excitation modulation signal through the signal processing unit to make the vibrating wire sensor vibrate, and collect the vibrating wire vibration signal of the vibrating wire sensor through the signal processing unit, and preprocess the vibrating wire vibration signal to obtain the fundamental signal.

[0073] Step 202: Linearly amplify the fundamental signal to obtain a linearly amplified signal.

[0074] Step 203: Amplify the linearly amplified signal multiple times to obtain multiple harmonic signals.

[0075] Step 204: Perform time-domain and frequency-domain signal processing on the linear amplified signal and multiple harmonic signals to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal. Based on the time-domain and frequency-domain signals, SNR value, and duration of each harmonic signal in the linear amplified signal and multiple harmonic signals, determine and diagnose the health status of the vibrating wire sensor to obtain the health status diagnosis result of the vibrating wire sensor.

[0076] The time-domain and frequency-domain signals from the linear amplifier circuit and the saturated amplifier circuit, as well as the aforementioned data such as signal-to-noise ratio and duration with clear physical meaning, are integrated and synchronously input into the main processor. The neural network deployed on the main processor and already trained is then processed to determine and diagnose the health status of the vibrating wire sensor, thereby obtaining the health status diagnosis result of the vibrating wire sensor.

[0077] The linearly amplified signal and multiple harmonic signals are processed to obtain the signal-to-noise ratio (SNR) value and the duration of each harmonic signal. Based on the time-domain and frequency-domain signals, SNR value, and duration of each harmonic signal in the linearly amplified signal and multiple harmonic signals, the health status of the vibrating wire sensor is determined and diagnosed, resulting in a health status diagnosis result for the vibrating wire sensor. Specifically, this includes: performing a Fourier transform on the linearly amplified signal to obtain a first time-domain signal and a first frequency-domain signal; performing full-spectrum waveform analysis on the first frequency-domain signal to obtain the center frequency waveform power; performing a Fourier transform on the multiple harmonic signals to obtain a second time-domain signal and a second frequency-domain signal; performing full-spectrum waveform analysis on the second frequency-domain signal to obtain the noise power and duration of each harmonic signal; calculating the SNR value based on the center frequency waveform power and noise power; and inputting the first time-domain signal, the first frequency-domain signal, the second time-domain signal, the second frequency-domain signal, the SNR value, and the duration of each harmonic signal into a trained neural network model to determine and diagnose the health status of the vibrating wire sensor, thus obtaining a health status diagnosis result for the vibrating wire sensor.

[0078] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0079] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A device for diagnosing the health status of a vibrating wire sensor, characterized in that, The vibrating string sensor health state diagnosis device comprises a main processor, a vibration signal generation module, a linear amplification circuit and a saturation amplification circuit. The vibration signal generation module comprises a signal processing unit and a vibrating string sensor, and is configured to: generate an excitation modulation signal through the signal processing unit to make the vibrating string of the vibrating string sensor vibrate, and collect the vibrating string vibration signal of the vibrating string sensor through the signal processing unit, and pre-process the vibrating string vibration signal to obtain a fundamental wave signal; The linear amplification circuit is configured to linearly amplify the fundamental wave signal to obtain a linearly amplified signal; The saturation amplification circuit is configured to multiply amplify the linearly amplified signal to obtain a multiple harmonic signal; The main processor is configured to: perform time domain and frequency domain signal processing on the linearly amplified signal and the multiple harmonic signal to obtain a signal-to-noise ratio value and a duration of each harmonic signal, and perform judgment and diagnosis on the health state of the vibrating string sensor according to the time domain and frequency domain signals, the signal-to-noise ratio value and the duration of each harmonic signal in the linearly amplified signal and the multiple harmonic signal to obtain a health state diagnosis result of the vibrating string sensor.

2. The vibrating wire sensor health diagnostic apparatus according to claim 1, wherein The main processor comprises a first signal processing unit, a second signal processing unit, a third signal processing unit and a diagnosis unit; The first signal processing unit is configured to: perform Fourier transform on the linearly amplified signal to obtain a first time domain signal and a first frequency domain signal; and perform full spectrum waveform analysis on the first frequency domain signal to obtain a center frequency waveform power; The second signal processing unit is configured to: perform Fourier transform on the multiple harmonic signal to obtain a second time domain signal and a second frequency domain signal; and perform full spectrum waveform analysis on the second frequency domain signal to obtain a noise power and a duration of each harmonic signal; The third signal processing unit is configured to: calculate a signal-to-noise ratio value according to the center frequency waveform power and the noise power; The diagnosis unit is configured to: input the first time domain signal, the first frequency domain signal, the second time domain signal, the second frequency domain signal, the signal-to-noise ratio value and the duration of each harmonic signal into a trained neural network model to perform judgment and diagnosis on the health state of the vibrating string sensor to obtain the health state diagnosis result of the vibrating string sensor.

3. The vibrating wire sensor health diagnostic apparatus of claim 2, wherein The neural network model is an rnn model, a cnn model or a fnn model.

4. The vibrating wire sensor health diagnostic apparatus of claim 2, wherein In terms of calculating the signal-to-noise ratio value according to the center frequency waveform power and the noise power, the third signal processing unit is further configured to: divide the center frequency waveform power by the noise power to obtain a power ratio value; perform logarithmic processing on the power ratio value to obtain a logarithmic processing result; and multiply the logarithmic processing result by 10 to obtain the signal-to-noise ratio value.

5. The vibrating wire sensor health diagnostic apparatus of claim 1, wherein, The signal processing unit comprises an excitation circuit, a vibrating string sensor interface circuit and a vibration pickup circuit; The excitation circuit is configured to: amplify and modulate an initial excitation signal to generate an excitation modulation signal; and the initial excitation signal is generated by the main processor; The vibrating string sensor interface circuit is configured to: connect the vibrating string sensor, the excitation circuit and the vibration pickup circuit; The vibration pickup circuit is configured to pick up the vibration signal of the vibrating string sensor and perform filtering processing on the vibration signal to obtain a fundamental wave signal.

6. The vibrating wire sensor health diagnostic apparatus of claim 1, wherein, The vibrating string sensor health state diagnosis device further comprises a double-channel waveform data acquisition module configured to acquire the linearly amplified signal and the multiple harmonic signals and transmit them to the main processor.

7. The string sensor health diagnosis device according to claim 1, characterized by, The vibrating string sensor health state diagnosis device further comprises a power supply module configured to provide electric energy.

8. The vibrating wire sensor health diagnostic apparatus of claim 1, wherein, The vibrating string sensor health state diagnosis device further comprises a display module configured to display the analysis and evaluation results and key index data of the main processor; the analysis and evaluation results and key index data include a signal-to-noise ratio value, a duration of each harmonic signal and a health state diagnosis result.

9. A vibrating wire sensor health condition diagnosis method based on the vibrating wire sensor health condition diagnosis apparatus according to any one of claims 1 to 8, characterized by, The vibrating string sensor health state diagnosis method comprises: generating an excitation modulation signal by the signal processing unit to make the vibrating string of the vibrating string sensor vibrate, and collecting the vibration signal of the vibrating string sensor by the signal processing unit, and performing preprocessing on the vibration signal to obtain a fundamental wave signal; linearly amplifying the fundamental wave signal to obtain a linearly amplified signal; multiple amplifying the linearly amplified signal to obtain multiple harmonic signals; processing the linearly amplified signal and the multiple harmonic signals to obtain a signal-to-noise ratio value and a duration of each harmonic signal, and determining and diagnosing the health state of the vibrating string sensor according to the time-domain and frequency-domain signals, the signal-to-noise ratio value and the duration of each harmonic signal in the linearly amplified signal and the multiple harmonic signals, to obtain a health state diagnosis result of the vibrating string sensor.

10. The vibrating wire sensor health diagnostic method of claim 9, wherein, processing the linearly amplified signal and the multiple harmonic signals to obtain a signal-to-noise ratio value and a duration of each harmonic signal, and determining and diagnosing the health state of the vibrating string sensor according to the time-domain and frequency-domain signals, the signal-to-noise ratio value and the duration of each harmonic signal in the linearly amplified signal and the multiple harmonic signals, to obtain a health state diagnosis result of the vibrating string sensor, specifically comprising: performing Fourier transform on the linearly amplified signal to obtain a first time-domain signal and a first frequency-domain signal; performing full-spectrum waveform analysis on the first frequency-domain signal to obtain a central frequency waveform power; performing Fourier transform on the multiple harmonic signals to obtain a second time-domain signal and a second frequency-domain signal; performing full-spectrum waveform analysis on the second frequency-domain signal to obtain a noise power and a duration of each harmonic signal; calculating a signal-to-noise ratio value according to the central frequency waveform power and the noise power; inputting the first time-domain signal, the first frequency-domain signal, the second time-domain signal, the second frequency-domain signal, the signal-to-noise ratio value and the duration of each harmonic signal into a trained neural network model to determine and diagnose the health state of the vibrating string sensor, to obtain the health state diagnosis result of the vibrating string sensor.

Images