Optical fiber sensor used for measuring the rotation angle of the disconnecting switch | Scientific Reports

Scientific Reports volume 15, Article number: 5477 (2025) Cite this article

321 Accesses

Metrics details

In order to solve the problem that the traditional electronic angle sensor cannot be directly installed on the high voltage isolation switch for Angle measurement, a system for high voltage isolating le measurement is proposed. The switch rotation angle is converted to a straight line for measurement, and the rotation angle of the switch is calculated by photoelectric conversion technology. The light source fluctuation of the laser receiver causes the nonlinear output result of the laser receiver, which compensates for the light intensity received by the receiver in both hardware and software respectively. In the hardware aspect, a two-optical path receiving compensation method is proposed for compensation. In terms of software, the compensation method for sparrow search algorithm (SSA) (SSA-BP) is proposed to improve the detection accuracy. The experimental results show that the average absolute error decreased to 0.18 and 0.3 after the dual-path receiving compensation method and SSA-BP compensation.

High-voltage disconnect switches are important components in power systems, and their performance directly affects the stable operation of the power system. Therefore, real-time monitoring and assessment of their condition are of great significance1. Among these, the blade angle is one of the key parameters for monitoring the condition of disconnect switches and plays a crucial role in the safe operation of the power grid.

Many scholars both domestically and internationally have conducted research on the condition monitoring of high-voltage disconnect switches. For instance, Reference2 proposed a fault diagnosis method for high-voltage disconnect switches based on vibration signals and motor current signals. The method first extracts the RMS values of segmented envelopes and processes the vibration signals of the operating mechanism using Variational Mode Decomposition(VMD). It then constructs fault identification feature parameters by fusing the effective values of the current segment envelopes with the energy entropy values of each IMF component, and classifies the faults using Support Vector Machine(SVM). Reference3 proposed a mechanical fault diagnosis scheme based on torque and rotation angle detection. It draws a torque–angle curve based on the fault characteristic signals and compares this curve with the normal one to determine whether the operation of the disconnect switch is in place. However, neither of these methods can directly reflect the state of the blade, and misjudgment can easily occur if other faults lead to incomplete opening or closing. Reference4 proposed an online monitoring method for the operating condition of disconnect switches based on dual analysis of current and angular displacement. Using current transformers and rotary displacement sensors, it obtains the current value, current–time curve, and travel-time curve of the drive motor during the operation of the GIS disconnect switch. It then indirectly judges whether the opening and closing positions of the disconnect switch are in place based on the patterns of change in the current value, current–time curve, + 5%envelope, and the travel-time curve of the disconnect switch contacts. Reference5 proposed using attitude sensors to identify the fault status of disconnect switches, preprocessing the collected data with the GA-SMOTE-ET algorithm, and recognizing the condition of the disconnect switch using an improved Stacking model algorithm. Reference6 monitored and analyzed disconnect switches by collecting their vibration and acoustic signals. It used empirical mode decomposition of random errors to analyze the distribution of vibration signals in disconnect switches, extracted vibration signals within characteristic frequency bands, and applied them to the mechanical fault diagnosis of disconnect switches. These methods require the installation of electronic sensors on high-voltage disconnect switches for measurement. However, these electronic angle sensors face the following main issues in high-voltage environments:

Electromagnetic interference:The strong electric and magnetic fields surrounding high-voltage disconnect switches can severely interfere with the signal transmission and measurement accuracy of electronic sensors, leading to increased measurement errors.

Insulation issues:Electronic sensors require complex insulation measures to operate safely in high-voltage environments, which complicates installation and maintenance.

Power supply issues:Installing electronic sensors on high-voltage equipment requires additional power supply, which is inconvenient in practical applications.

Based on the above reasons, this paper proposes a disconnect switch blade angle measurement system based on a reflective fiber-optic angle sensor. This method converts the rotation angle of the moving contact to a straight line for measurement and calculates the rotation angle of the moving contact through photoelectric conversion technology. By leveraging the strong insulation and high interference resistance of fiber-optic sensors, they can be directly in contact with high-voltage equipment without the need for complex insulation treatment, effectively solving the problems of inconvenient installation and susceptibility to electromagnetic interference faced by traditional sensors in high-voltage environments. Moreover, the corrosion resistance and high-temperature tolerance of fiber-optic sensors make them more suitable for long-term operation. In practical applications, the fluctuation of the light source of the laser transmitter can cause nonlinearity in the output of the laser receiver, thereby affecting measurement accuracy. This paper employs dual-optical-path reception compensation and a compensation method optimized by the Sparrow Search Algorithm(SSA)for BP neural networks(SSA-BP)to compensate for the light intensity received by the receiver. By utilizing the powerful nonlinear mapping capability of neural networks, further optimization of the remaining nonlinearity after hardware compensation is achieved, significantly enhancing the system’s accuracy in measuring the blade angle. This effectively addresses the shortcomings in measurement accuracy improvement found in previous studies and provides more precise data support for the condition monitoring of high-voltage disconnect switches.

The closing state of the disconnecting switch can be confirmed by measuring the rotation angle of the switch. The traditional electronic Angle sensor is difficult in collecting power and easy to be electromagnetic interference in the high voltage isolation, so it cannot accurately detect the closing state of the switch. To solve this problem, we propose a switch angle measurement system based on reflective fiber sensor. The system emits a laser beam into a reflective fiber through a laser transmitter, which is then reflected back to the laser receiver through a reflector. The received optical signal is converted into electrical signals and transmitted to the control circuit board for processing. As shown in Fig. 1, when the knife pushes/pulls the pull rod, the slider will drive the reflector to move on the slide rail, thus changing the distance between the reflector and the reflector of the reflective optical fiber probe, and then changing the intensity of light reflected into the optical fiber. The change in light intensity will cause a change in the electrical signal output by the laser receiver. By establishing the relationship between the moving distance of the mirror and the output voltage of the laser receiver, the rotation Angle of the switch can be calculated. Because of its strong insulation and anti-interference ability, it can be directly installed in high voltage isolation for measurement.

Overall system design.

As shown in Fig. 2 for the reflective optical fiber sensor measurement model. After the laser beam in the transmitting fiber hits the reflector, there will be an emission Angle \(\theta\) between its light intensity and the horizontal line, which will be reflected back by the reflector into the receiving fiber, and its light intensity will change with the change of the distance \(d\) between the fiber probe and the reflector. Therefore, \(d\) can be obtained by detecting the size of the light intensity. Then the rotation Angle of the brake can be obtained through straight-curve transformation. In the figure, the diameter of the transmitting fiber and the receiving fiber is expressed as \(2r\), and the distance between the two is \(a\). The receiving fiber must be inside the cone of light formed by the image of the transmitting fiber in order to capture the light reflected back by the reflecting surface. According to the principles of geometric optics, this interaction between the transmitting fiber and the receiving fiber can be equivalent to the interaction between the mirror image of the transmitting fiber and the receiving fiber7.

Measurement model of the reflective optical fiber sensor.

Assuming that the receiving fiber is exactly within the light cone, there is:

The larger the numerical aperture of the fiber, the larger the emission Angle \(\theta\), expressed as:

\(NA\) is the spot aperture. In the numerical aperture of the fiber, combined formula (1) and (2) can obtain:

If \({\text{d}} < \frac{a}{2\tan \arcsin NA}\), the end surface of the receiving fiber was not within the coverage of the reflected laser beam. If \(d > \frac{a + 2r}{{2\tan \arcsin NA}}\), the receiving fiber could not receive the reflected laser beam, and there was no optical coupling rate between the receiving fiber and the transmitting fiber. At that time, the end surface of the receiving optical fiber can be fully covered by the reflected laser beam, that is to say, the reflected light intensity is the largest, the optical coupling rate is the highest, and the coupling coefficient is expressed as:

If \(\frac{a}{2\tan \arcsin NA} < d < \frac{a + 2r}{{2\tan \arcsin NA}}\), only some of the receiving fiber ends could receive the reflected laser signal.

In the process of ranging, the reflective fiber sensor mainly judges the moving distance by receiving the optical fiber intensity, and the light intensity modulation function determines how the signal output of the sensor changes with the change of the measured physical quantity8, which is expressed as:

where \(M\) represents the optical stress function of the fiber sensor, \(P_{s}\) and \(P_{0}\) represent the optical power received by the receiving fiber and the optical power sent by the transmitting fiber, respectively. Combined with the analysis in 2.1, it can be seen that the light intensity emitted by the emitting fiber is:

where, \(r_{2}\) represents the radius of the receiving fiber, and \(\omega (2d)\) represents the diameter of the laser spot at a distance of \(2d\).

Figure 3 shows the distribution of the coupled light spots emitted by the transmitting fiber. The intensity modulation characteristic function can be obtained by calculating the ratio of the light area projected at the image of the transmitting fiber to the light area received by the receiving fiber.

Coupling spot distribution of transmitting fiber.

It can be seen from Fig. 3 that If \(\omega (2d) < a - r_{2}\), there is no overlap between the receiving spot and the reflected spot, indicating that the receiving fiber cannot receive the reflected laser beam. If \(a - r_{2} < \omega (2d) < a + r_{2}\), the receiving fiber can receive the laser signal reflected back, and the received light intensity can be calculated by the micro-element method, and \(P_{s}\) can be calculated by combining the formula (6) :

where, \(C\) is the Fresnel reflection interface, and 0.04 is taken at the reflector interface.

If \(\omega (2d) > a + r_{2}\), the distance between the fiber probe and the reflector increases, and the light intensity reflected to the receiving fiber decreases, then PS is:

Combined with the above analysis, \(M\) can be obtained, expressed as:

Light intensity modulation function is usually nonlinear, and by the external light, its mechanical structure, high voltage isolation switch vibration and other factors, the influence of the receiving fiber output light intensity is nonlinear change, namely after the laser receiver photoelectric conversion electrical signal is nonlinear change, so need to nonlinear compensation, in order to improve the system detection accuracy. In terms of light intensity compensation, hardware and software technology can be adopted. Among them, hardware compensation can be realized by adding hardware devices. This method is very responsive and has strong implementation, but it will increase the cost. Software compensation does not require additional hardware investment, and implementation costs are lower for existing systems. But the software processing usually has a certain delay. Therefore, combining the advantages of hardware and software compensation, using the method of changing the reflective fiber sensor probe and increasing the neural network.

From the analysis in 2.4, it can be seen that the design structure and parameters of the reflective fiber sensor probe affect its performance. In order to increase the reception range of the reflected light, the probe is designed into a two-ring coaxial end surface structure, and the compensation of the receiving light intensity is realized by calculating the ratio of the receiving light power of the inner and outer circles.

Let the optical flux obtained by the receiving fiber be:

where, \(\rho\) is the reflectivity of light,\(K\) is the loss of light, \(\exp \left( { - \sum\limits_{i} {\eta_{i} r_{i} } } \right)\) is the loss caused by external bending of the receiving fiber, and \(S\) is the area of the receiving fiber detected by the reflected laser。

The light flux obtained in the inner and outer circles are:

where \(d\) is the axial distance between the inner ring and the outer ring, \(z\) is the distance between the optical fiber probe and the mirror, and the central coordinates of the inner and outer rings are \((2z,d)\) and \((2z,2d)\) respectively. According to Eq. (9), the optical emphasis function can be obtained as follows:

It can be seen from the formula that \(k_{0}\) and \(\varphi_{0}\) are eliminated, indicating that the influence of light loss and light intensity change is ideally eliminated. Since both the receiving fiber and the transmitting fiber are of one type, formula (13) can be converted to:

When \(d\), \(a_{0}\), and \(\theta_{c}\) are stabilized, the optical stress function depends only on the distance \(z\) between the probe and the reflector. This design method eliminates the influence of light source fluctuation and fiber loss on light intensity.

In order to further compensate the nonlinearity that still exists after hardware compensation, the nonlinear mapping capability of neural network is utilized to achieve hardware compensation to improve the measurement accuracy, as shown in Fig. 4. \(p_{1}\) and \(p_{2}\) respectively represent the photoelectric conversion signal received by the inner and outer ring receiving fibers, wherein the light intensity obtained by the inner ring receiving fibers is related to the distance \(d\) from the probe to the reflector and the fluctuation of the light source of the laser emitter, while the light source received by the outer ring fibers is used as a reference light path, which is only related to the light source \(I_{0}\) of the laser emitter, expressed as:

Schematic of light intensity compensation based on the neural network.

BP (Back Propagation) neural network9,10 is a commonly used feedforward neural network with strong nonlinear mapping ability. Therefore, it is used to fit the linear relationship between the optical power obtained by the receiving fiber and the distance between the probe and the reflector, and then adjust the parameters by calculating the reverse error between the actual output value \(v\) and the expected output value \(d\), and obtain:

\(x\) is a minimum integer.

Because the BP neural network is easy to fall into the local optimal solution during training, the generalization ability of the network is insufficient11. The sparrow search algorithm (Sparrow Search Algorithm, SSA) can avoid the premature convergence of12,13 in the search process by simulating the foraging and vigilance behavior of sparrows, so as to help the BP neural network to jump out of the local optimal solution and find a better parameter setting14. Therefore, this paper proposes the Sparrow search algorithm (SSA) optimization BP neural network (SSA-BP) model to realize the non-linear compensation of light intensity in the software, as shown in Fig. 5. The specific steps of the model are as follows:

Flow chart of the BP neural network optimized by SSA.

S1. Determine the input and output variables. The input variable is the photoelectric conversion signal received by the receiving fiber in the inner circle and the outer circle, and the output variable is the value after compensation.

S2. Data preparation: Collect sample data for training and verification, normalize the data to ensure that the data is representative of the actual situation.

S3. BP Neural Network modeling: Build a backpropagation (BP) neural network model, which includes the input layer, hidden layer and output layer. Ensure that the network structure is reasonable, and choose the appropriate activation function and loss function according to the actual situation.

S4. Initial parameter setting: randomly initialize the weights and threshold values of the BP neural network. These parameters will be optimized during the training process by using a sparrow search algorithm.

S5. Hyperparameter Determination: The method of combining grid search with cross-validation is employed to determine the optimal hyperparameter combination. The learning rate range for the BP neural network is set to [0.001, 0.1]. Considering the dimensionality of the input data and the complexity of the problem, the number of neurons in the hidden layer is set to range from [5, 50]. The proportion of discoverers for the SSA is set to range from [0.1, 0.5], and the alarm value range is [0.5, 0.9]. These hyperparameters are combined in a grid manner, with the learning rate taking values in increments of 0.005, the number of neurons in the hidden layer in increments of 5, and the parameters of the Sparrow Search Algorithm (SSA) also taking values in set increments within their respective ranges, forming numerous hyperparameter combinations. Then, for each combination of hyperparameters, a fivefold cross-validation method is used to train the SSA-BP model on the training set. The training set data is divided into 5 equal parts, with 4 parts selected as training data and 1 part as validation data each time, and the Root Mean Square Error (RMSE) of the model on the validation set is calculated. After cross-validation of all combinations, the hyperparameter combination with the smallest RMSE is selected as the final hyperparameters for the model.

S6. Sparrow search algorithm: The sparrow search algorithm is used as an optimization algorithm to adjust the parameters of BP neural network, which is optimized by iterative searching and updating the current optimal solution.

S7. Training neural network: train BP neural network with sample data, and optimize the weights and thresholds according to the sparrow search algorithm. The back-propagation algorithm was used to calculate the gradient, and the parameters were updated using the optimization algorithm. This step is repeated until the convergence conditions are reached.

S8. Model evaluation and validation: The trained BP neural network model was evaluated using the validation data set. Calculate the performance indicators of the model, such as root mean square error (RMSE) or determination coefficient (R2). If the model performs poorly, the network structure or parameters can be adjusted and re-trained and validated.

S9. Sensitivity Analysis: If it is found that the model converges too slowly, the learning rate can be appropriately increased. If the model exhibits overfitting, reducing the number of neurons in the hidden layer can be considered. Simultaneously, for the parameters of the Sparrow Search Algorithm (SSA), depending on the specific requirements of the problem and the characteristics of the data, the proportion of discoverers and the alarm value can be flexibly adjusted to balance the model’s exploration and exploitation capabilities, thereby improving the model’s generalization performance.

S10. Nonlinear compensation: The optimized BP neural network model is used to predict the measured value of the new fiber sensors, and the nonlinear compensation value is obtained.

In order to verify the accuracy of the system and the performance of the nonlinear compensation method, an experimental platform based on the reflective optical fiber sensor is built, as shown in Fig. 6. The high precision hall angle sensor is installed together on the rotating axis of the switch, each angle sensor to each Angle value of the optical fiber probe to the mirror, and then output a voltage value on the laser receiver, by establishing a model of the relationship between the angle value and the output voltage value, then obtain the Angle-voltage characteristic curve, and the final relationship between the Angle and the voltage.

Experimental Platform.

First, the uncompensated system was used for the experiment, and the high voltage isolation switch switch was rotated from 0 to 90°, and measured back and forth for 100 times. Five sets of experimental data were selected as shown in Fig. 7.

Voltage-angle characteristic curve without compensation.

As can be seen from the figure, the output voltage decreases as the angle increases, and the nonlinearity of the angle-voltage characteristic curve of the system is poor. Then, the hardware compensation system is used for the experiment, and the data collected before and after the hardware compensation is fitted by using the least squares method. The fitted lines are respectively:

The difference between the original output value and the fitted value before and after the hardware compensation is calculated by Eq. (17) as shown in Table 1.

Based on the data in Table 1, the non-linear error of the system can be calculated as:

Similarly, the nonlinear error of the system after the hardware compensation is calculated as follows:

According to Eqs. (16) and (17), it is known that the nonlinear error of the system is reduced by 9.88 percentage points after hardware compensation, indicating that the hardware compensation structure designed in this paper has achieved remarkable results.

Based on hardware compensation, we further employed the SSA-BP algorithm for software nonlinear compensation. The relevant dataset was derived from real measurements collected on the experimental platform, covering data from various angles and multiple environmental conditions(including vibration and temperature variations). To enhance the model’s generalization ability and training efficiency, the data were normalized and divided into training and validation sets at a ratio of 8:2, with the training set used for model training and the validation set for assessing model performance. Subsequently, the least squares method was applied to fit the data collected before and after software compensation, yielding the following fitted linear equations is:

Get the compensation results as shown in Table 2.

Similarly, the nonlinear error of the system after performing SSA-BP compensation is calculated as:

From the calculation results, after SSA-BP compensation, the system nonlinear error decreased by 10.08 percentage points again on the basis of hardware compensation. Then it was compared with several common neural network-based nonlinear compensation algorithms (BP neural network, particle swarm optimized BP neural network (Particle Swarm Optimization-Back Propagation, PSO-BP)15,16, and artificial swarm optimized BP neural network (Artificial Bee Colony-Back Propagation, ABC-BP)17,18), and obtained the results shown in Table 3.

As can be seen from the table, using BP neural network nonlinear compensation algorithm alone, can only reduce 2.62% on the basis of hardware compensation, and after the BP neural network optimization, the nonlinear error, compared with PSO-BP and ABC-BP, the system with SSA-BP compensation nonlinear error effect is most obvious, 4.08% and 3.62% lower than the two optimization algorithms respectively.

The fiber-optic angle sensor designed in this study was installed together with the high-precision MK415B Hall angle sensor on the rotating shaft of the disconnect switch blade for angle measurement. In the experiment, a vibration device was attached to the rotating shaft of the blade to simulate the vibration during the operation of the disconnect switch. Tests were also conducted under different temperature conditions, including ambient, high, and low temperatures. The angle values collected by the MK415B were used as reference values. Partial experimental data are shown in Table 4, and the measurement error curve is illustrated in Fig. 8.

Measurement error curve.

From Table 4 and Fig. 8, it can be observed that the system’s mean absolute error(MAE)reached 0.48 without nonlinear compensation. After applying the dual-optical-path reception compensation method and the SSA-BP compensation, the MAE was reduced to 0.18, a decrease of 0.3, which is very close to the values measured by the high-precision Hall angle sensor. Additionally, the measurement errors of the fiber-optic angle sensor were relatively larger when the blade rotation angles were smaller or larger. This is mainly because vibrations occur during the opening and closing operations of the disconnect switch, and the mechanical structure of the fiber-optic angle sensor designed in this study is susceptible to these vibrations, leading to higher measurement errors.

To help us more intuitively understand the effectiveness of the compensation methods, confidence intervals were used to estimate the range of measurement errors of the fiber-optic sensor before and after compensation. For the pre-compensation error, the confidence level was set at 95%, with the corresponding Z-value being 1.96. The confidence interval is given by:

where \(m_{1}\) is the mean error before compensation,\(s_{1}\) is the standard deviation before compensation, and \(n\) is the sample size.

Similarly, the confidence interval after compensation is:

These confidence intervals demonstrate that the compensation method significantly reduced the range of measurement errors. The narrower post-compensation error interval indicates the effectiveness of the compensation method.

Based on the rotation characteristics, a reflective light angle sensor is designed. The optical fiber probe is designed into a two-ring coaxial end surface structure, and the compensation of the receiving light intensity is realized by calculating the ratio, the nonlinear compensation is realized, and the nonlinear error of the system is reduced by 9.88 percentage points. On the basis of hardware compensation, the compensation method of Sparrow search algorithm (SSA) (SSA-BP) is proposed, which makes the nonlinear error of the system decrease by 10.08 percentage points again. After the two-light path receiving compensation method and SSA-BP compensation, the average absolute error of the system measurement decreased to 0.18 and decreased by 0.3. It shows that the accuracy of system detection is greatly improved after software and hardware compensation.

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Zhou, T. et al. Mechanical defect detection of porcelain column high-voltage disconnector based on operating torque. Int. J. Adv. Robot. Syst. 17(1), 845 (2020).

Article MATH Google Scholar

Zhang, Z. et al. Mechanical fault diagnosis of a disconnector operating mechanism based on vibration and the motor current. Energies 15(14), 5194 (2022).

Article MATH Google Scholar

Jiang, J. et al. Mechanical fault diagnosis of high voltage isolation switch. J. Shenyang Univ. Technol. 45(2), 127–132 (2023).

MATH Google Scholar

Wang, Z., et al. Research and design of online monitoring device for operating state of GIS isolating switch operationg mechanism. In 2021 IEEE 5th Conference on Energy Internet and Energy System Integration (EI2). (IEEE, 2021).

Kemeng, Li. et al. Research on intelligent fault diagnosis of high voltage isolation switch based on attitude sensor. Power Grid Technol. 47(9), 3781–3790 (2023).

MATH Google Scholar

Zhang, Z., et al. Research on feature extraction and fault diagnosis of mechanical vibration and sound signal of Disconnector. In 2022 IEEE 5th International Electrical and Energy Conference (CIEEC). (IEEE, 2022).

Jiang, R. & Zheng, Y. Fiber-optic current sensing based on reflective polarization-bias-added structure. IEEE Sens. J. (2023).

Cao, J. Research on Tactile Sensor Based on Optical Emphasis (Beijing University of Posts and Telecommunications, 2022).

MATH Google Scholar

Zhu, F. et al. Research and design of hybrid optimized backpropagation (BP) neural network PID algorithm for integrated water and fertilizer precision fertilization control system for field crops. Agronomy 13(5), 1423 (2023).

Article MATH Google Scholar

Xu, L. et al. Grade evaluation of black-odorous urban rivers in the Greater Bay Area of China using an improved back propagation (BP) neural network. Environ. Sci. Pollut. Res. 30(19), 55171–55186 (2023).

Article Google Scholar

Kebin, Q. et al. Improved artificial potential field method for UAV escape from local minimum problem. Mod. Electron. Tech. 47(10), 107–110 (2019).

MATH Google Scholar

Gharehchopogh, F. S. et al. Advances in sparrow search algorithm: A comprehensive survey. Arch. Comput. Methods Eng. 30(1), 427–455 (2023).

Article PubMed MATH Google Scholar

Zhang, J. et al. A novel sparrow search algorithm with integrates spawning strategy. Clust. Comput. 27(2), 1753–1773 (2024).

Article MathSciNet MATH Google Scholar

Li, M. et al. Classification and identification of mixed gases based on the combination of semiconductor sensor array with SSA-BP neural network. Meas. Sci. Technol. 34(8), 085110 (2023).

Article ADS MATH CAS Google Scholar

Quan, G. et al. Characterization of flow behaviors by a PSO-BP integrated model for a medium carbon alloy steel. Materials 16(8), 2982 (2023).

Article ADS PubMed PubMed Central MATH CAS Google Scholar

Hu, D. et al. Prediction method of surface settlement of rectangular pipe jacking tunnel based on improved PSO-BP neural network. Sci. Rep. 13(1), 5512 (2023).

Article ADS MathSciNet PubMed PubMed Central MATH CAS Google Scholar

Zhu, Y. et al. Structural safety monitoring of high arch dam using improved ABC-BP model. Math. Probl. Eng. 1, 6858697 (2016).

Google Scholar

Wang, Y. et al. Medium and long-term precipitation prediction using wavelet decomposition-prediction-reconstruction model. Water Resour. Manag. 36(3), 971–987 (2022).

Article MATH Google Scholar

Download references

School of Physics and Information Engineering, Zhaotong University, Zhaotong, 657000, Yunnan, China

Zhang Jing, Gu Qingchuan, Wan Pu, Liu Yinxu, Yang Desheng & Li Xin

You can also search for this author in PubMed Google Scholar

Z.J. and G.Q. wrote the main manuscript text, W.P. and L.Y. prepared the experimental data and Figs. 1–3. Y.D. was responsible for data analysis, and L.X. conducted the literature review. All authors participated in the review and revision of the manuscript.

Correspondence to Zhang Jing or Gu Qingchuan.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Jing, Z., Qingchuan, G., Pu, W. et al. Optical fiber sensor used for measuring the rotation angle of the disconnecting switch. Sci Rep 15, 5477 (2025). https://doi.org/10.1038/s41598-025-89310-8

Download citation

Received: 12 October 2024

Accepted: 04 February 2025

Published: 14 February 2025

DOI: https://doi.org/10.1038/s41598-025-89310-8

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative