CN-122015923-A - Synchronous sensing method and device for temperature and pressure of bridge ship collision prevention system parameter monitoring

CN122015923ACN 122015923 ACN122015923 ACN 122015923ACN-122015923-A

Abstract

The invention provides a synchronous sensing method and device for temperature and pressure of parameter monitoring of a bridge ship collision prevention system, and relates to the technical field of optical fiber sensing. The method comprises a sensor calibration module and a real-time demodulation module. The sensor calibration module comprises a sensor calibration unit, a spectral feature extraction unit, a database unit, a data preprocessing and standardization unit, a model construction and training unit and a super-parameter optimization unit. The model construction and training unit is used for constructing a multi-output support vector regression model, specifically, a regression function is respectively established for temperature and pressure and comprises a first SVR model and a second SVR model, and the method can realize synchronous and high-precision measurement of the temperature and the pressure by utilizing a single sensing unit and a single demodulation model, so that the reliability, the compactness and the long-term stability of the system are obviously improved while the measurement precision is ensured, and the harsh requirements of safety monitoring of important infrastructures such as ship crashing bridges are met.

Inventors

LIU YIN
YUAN TING
Eddie Bai Ainivar
LIU XINRAN
Feruza Tursun Hazi
CHEN HONGBING
LI ZEYU
XU GANG

Assignees

北京科技大学

Dates

Publication Date: 20260512
Application Date: 20260108

Claims (10)

1. The method for synchronously sensing the temperature and the pressure of the parameter monitoring of the bridge ship collision prevention system is characterized by being applied to the system for synchronously sensing the temperature and the pressure of the parameter monitoring of the bridge ship collision prevention system, and the system for synchronously sensing the temperature and the pressure of the parameter monitoring of the bridge ship collision prevention system comprises a sensor calibration module and a real-time demodulation module, and is characterized by comprising the following steps: the sensor calibration module comprises a sensor calibration unit, a spectral feature extraction unit, a database unit, a data preprocessing and standardization unit, a model construction and training unit and a super-parameter optimization unit, wherein the sensor calibration unit is used for collecting original spectral data, the model construction and training unit is used for constructing a multi-output support vector regression model, the multi-output support vector regression model is constructed by respectively constructing regression functions for temperature and pressure, the model learns independent mapping relations between temperature and pressure from the same spectral feature set, and the multi-output support vector regression model comprises a first SVR model and a second SVR model; the real-time demodulation module comprises a data acquisition unit, a preprocessing unit, a standardization unit and a model reasoning unit.
2. The synchronous sensing method of temperature and pressure for monitoring parameters of a bridge anti-ship collision system according to claim 1, wherein the sensor calibration unit comprises: placing the prepared single-mode-hollow-coreless optical fiber sensor in a high-precision temperature and pressure control box; Generating a sensor calibration task instruction to perform a calibration process to obtain original spectrum data, wherein the generating of the sensor calibration task instruction comprises defining a measurement condition instruction based on an expected application range, setting calibration points, and for each combination of a temperature value and a pressure value, acquiring complete spectrum data multiple times by using a spectrometer with spectral resolution not less than 0.01nm, wherein the acquiring complete spectrum data multiple times comprises acquiring multiple samples at each calibration point to reduce the influence of random noise, and the calibration points correspond to a series of temperature value and pressure value combinations generated in a coverage expected application range.
3. The method for synchronously sensing the temperature and the pressure monitored by the bridge anti-ship collision system parameter according to claim 1, wherein the spectral feature extraction unit comprises: Extracting main interference valley wavelengths from each original spectrum based on the original spectrum data, wherein the main interference valley wavelengths are accurately positioned by a cubic spline interpolation method, and the main interference valley wavelengths correspond to wavelength values corresponding to the lowest transmissivity in the spectrum; Extracting secondary interference valley/peak wavelengths from each original spectrum based on the original spectrum data, wherein the secondary interference valley/peak wavelengths are obtained by searching local extreme points, the secondary interference valley/peak wavelengths are wavelength values of another significant interference extreme point, and the significant interference extreme points are valleys or peaks; Calculating a main interference valley depth from each original spectrum based on the original spectrum data, wherein the main interference valley depth is calculated based on the light intensity at the valley bottom and is a normalized light intensity value at the bottom of the main interference valley; Calculating a spectrum second-order central moment from each original spectrum based on the original spectrum data, wherein the calculation of the spectrum second-order central moment from each original spectrum comprises the steps of calculating based on spectrum center-of-gravity wavelength and light intensity of corresponding wavelength, and the spectrum second-order central moment represents the broadening degree of spectrum energy distribution; calculating a spectral asymmetry from each of the original spectra based on the original spectral data, the calculating of the spectral asymmetry from each of the original spectra including calculating based on a spectral centroid wavelength, a light intensity of a corresponding wavelength, and a second order central moment, the spectral asymmetry being used to quantify an asymmetry of a spectral distribution; Based on the primary interference valley wavelength, the secondary interference valley/peak wavelength, the primary interference valley depth, the second-order central moment of the spectrum and the spectrum asymmetry, a multidimensional feature vector is formed.
4. The method for synchronously sensing the temperature and the pressure monitored by the parameters of the bridge anti-ship collision system according to claim 1, wherein the database unit comprises: at each calibration point, extracting corresponding spectral feature vectors based on the spectral feature extraction unit; synchronously recording corresponding real physical quantity labels; each calibration point, the corresponding feature vector and the corresponding physical quantity label assembly sample group are added to a database, and the physical quantity labels comprise temperature labels and pressure labels; And performing quality inspection and labeling on the database to obtain a complete database.
5. The method for synchronously sensing the temperature and the pressure monitored by the parameters of the bridge anti-ship collision system according to claim 1, wherein the model construction and training unit comprises: mapping the input spectral feature vector to a high-dimensional feature space by adopting nonlinear mapping; Respectively constructing two optimal linear regression hyperplanes of temperature and pressure in the space; Using the same input feature vector, constructing two mapping functions, respectively training independent SVR models for temperature and pressure, respectively corresponding to a first SVR model and a second SVR model, and obtaining a multi-output support vector regression model, wherein the first SVR model comprises a mapping function for temperature prediction, the mapping function for temperature prediction comprises a weight vector and a bias term of the temperature model, the second regression model comprises a mapping function for pressure prediction, and the mapping function for pressure prediction comprises a weight vector and a bias term of the pressure model.
6. The synchronous sensing method of temperature and pressure for monitoring parameters of a bridge anti-ship collision system according to claim 1, wherein the super-parameter optimizing unit comprises: extracting a spectral feature vector based on the complete database; converting the spectral feature vector into a standardized input feature by adopting standardized processing; taking the standardized input characteristics as input, respectively taking a temperature label and a pressure label as targets, and training a first SVR model and a second SVR model; Optimizing the super parameters of the first SVR model and the second SVR model by adopting a strategy combining cross validation and a strategy of grid search, wherein the strategy of cross validation comprises the steps of randomly dividing a calibration data set into mutually exclusive subsets, sequentially taking the first subset of the mutually exclusive subsets as a verification set, taking the first subset as a training set, training the model on the training set and evaluating the prediction performance of the first subset on the verification set, and the strategy of grid search comprises traversing in a super parameter combination space defined by grids; after model training is completed, the performances of the first SVR model and the second SVR model are comprehensively evaluated by using an independent test set and a preset evaluation index.
7. The method for synchronously sensing the temperature and the pressure of the parameter monitoring of the bridge anti-ship collision system according to claim 1, wherein the real-time demodulation module comprises a data acquisition unit, a preprocessing unit, a standardization unit and a model reasoning unit, and comprises the following steps: after the sensor starts measurement, the data acquisition unit immediately acquires the original spectrum in real time; a preprocessing unit extracts a feature vector from the original spectrum; The normalization unit performs normalization processing on the extracted feature vector to obtain a normalized feature vector; The model reasoning unit inputs the normalized feature vector into a pre-trained multi-output support vector regression model to obtain a predicted value of temperature and a predicted value of pressure; and carrying out rationality check on the model output result, wherein the rationality check comprises abnormal value detection, and triggering a re-measurement mechanism if the abnormal value is detected.
8. A synchronous temperature and pressure sensing device for monitoring parameters of a bridge ship collision prevention system, which is used for realizing the synchronous temperature and pressure sensing method for monitoring parameters of the bridge ship collision prevention system according to any one of claims 1-7, and is characterized in that the device comprises: The sensor calibration module specifically comprises a sensor calibration unit, a spectral feature extraction unit, a database unit, a data preprocessing and standardization unit, a model construction and training unit and a super-parameter optimization unit, wherein the sensor calibration unit is used for collecting original spectral data, the model construction and training unit is used for constructing a multi-output support vector regression model, the construction of the multi-output support vector regression model comprises the steps of respectively establishing regression functions for temperature and pressure so that the model learns independent mapping relations between temperature and pressure from the same spectral feature set, and the multi-output support vector regression model comprises a first SVR model and a second SVR model; The real-time demodulation module specifically comprises a data acquisition unit, a preprocessing unit, a standardization unit and a model reasoning unit.
9. A synchronous sensing device for temperature and pressure monitored by bridge ship collision prevention system parameters, which is characterized by comprising a synchronous sensing processor for temperature and pressure monitored by the bridge ship collision prevention system parameters, and a memory, wherein the memory is stored with computer readable instructions, and the computer readable instructions realize the method according to any one of claims 1 to 7 when being executed by the processor.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a program code, which is callable by a processor for executing the method according to any one of claims 1 to 7.

Description

Synchronous sensing method and device for temperature and pressure of bridge ship collision prevention system parameter monitoring Technical Field The invention relates to the technical field of optical fiber sensing, in particular to a temperature and pressure synchronous sensing method and device for monitoring parameters of a bridge ship collision prevention system. Background Bridge ship collision prevention monitoring systems need to sense the impact force (pressure) of a ship and the change of the ambient temperature in real time, because both can cause stress strain of a bridge structure, but the physical mechanism and the influence on the structure of the bridge structure are quite different, and must be distinguished. The mixed interference structure formed by welding the single-mode fiber, the hollow fiber and the coreless fiber can generate an interference spectrum extremely sensitive to external physical quantity, has the potential of realizing ultra-compact and high-sensitivity sensing, is very suitable for being embedded into bridge protection facilities or key stress parts, and is used for monitoring local pressure distribution and long-term temperature field change at the moment of ship collision. However, this structure faces a core technical bottleneck in practical applications, namely that its interference spectrum is simultaneously affected by the cross sensitivity of temperature and pressure. The change in one physical quantity "masquerades" as a signal of another physical quantity, resulting in a system that cannot determine whether a pressure change due to a ship impact has occurred or that only normal ambient temperature fluctuations have occurred. This fundamental risk of misjudgment seriously hampers the practical application of such sensors in bridge safety monitoring where high reliability is required. To cope with the cross-sensitivity problem, the most typical scheme in the prior art is to adopt a strategy of 'dual-channel sensing unit+discrete algorithm demodulation'. In particular, it is common practice to deploy two sensors, one primary sensor that is sensitive to both temperature and pressure and the other temperature-compensated sensor that is sensitive only to temperature in a reference environment. The system needs to establish demodulation models for the two sensors respectively, firstly calculates a temperature value through a reference sensor, then substitutes the temperature value into the model of the main sensor to deduct the temperature influence, and finally calculates a pressure value. Another similar technique is to try to prepare two interference units with different sensitivity to temperature and pressure response in one sensor, but the back end still relies on two independently constructed demodulation models for step-by-step calculation. These existing solutions, while alleviating the problem to some extent, introduce the significant drawbacks of first of all, they rely on multiple sensors or complex sensing structures, increasing the cost, packaging difficulty and failure rate of the system, reducing the feasibility of large-scale deployment on bridges. Secondly, the calculation flow of step demodulation is lengthy, not only the burden of the signal processing unit is increased, but also errors are accumulated and transferred in two steps of calculation, and finally the absolute precision and reliability of pressure measurement are limited. For the application scene that the ship hits the bridge to monitor and needs to instantly capture weak collision signals and eliminate day-night temperature difference interference, the demodulation precision, the system complexity and the instantaneity are difficult to consider in the prior art, and a more direct, efficient and accurate technical scheme is urgently needed. The temperature compensation type pressure measurement system based on the reference sensor theoretically provides a path for solving the cross sensitivity, but in practical engineering application, particularly in the ship collision bridge monitoring scene with extremely high requirements on reliability and precision, the temperature compensation type pressure measurement system exposes a plurality of inherent defects which are difficult to overcome. First, the system relies on dual sensing units, which directly results in a multiplication of hardware costs, a significant increase in system packaging complexity, and an increase in overall failure rate. In a severe service environment of a bridge, the failure of any one sensor can lead to the loss of the whole monitoring function, and the reliability of the system faces serious challenges. Secondly, the algorithm architecture of the step demodulation introduces significant error accumulation and transfer effects in the signal processing chain. Any minor errors in the temperature calculation in the first step are amplified in the pressure calculation in the second step, so that the absolute ac