A method, system, and medium for predicting an application of a pipe optical fiber

Abstract

The application discloses a pipeline optical fiber application situation prediction method and system and a medium, and comprises the following steps: screening influence parameters of a pipeline optical fiber monitoring system application situation, constructing a multi-level influence parameter system of a distributed optical fiber application situation, and determining the degree value corresponding to the influence parameters; calculating the subjective and objective weights of the influence parameters of the distributed optical fiber application situation based on a combination assignment method, and determining the maximum weight of the influence parameters based on game theory; determining the final degree value of the pipeline optical fiber monitoring system according to the degree value and the maximum weight of the influence parameters; combining the fuzzy comprehensive evaluation theory to analyze the influence of the distributed optical fiber application situation, and obtaining the influence result; constructing a combined neural network model and training the model; and predicting the distributed optical fiber application situation based on the trained combined neural network model, and obtaining the prediction result. The application solves the problems of the prior art, such as insufficient comprehensiveness of pipeline optical fiber application evaluation and low prediction accuracy.

Description

Technical Field

[0001] This invention relates to the field of optical fiber monitoring, and specifically to a method, system, and medium for predicting the application status of optical fibers in pipelines. Background Technology

[0002] Oil and gas pipelines serve as crucial energy transmission channels, but they are susceptible to failure due to external disturbances, corrosion, and issues with pipe materials and construction quality. While my country's design and construction standards are now largely on par with international levels, and significant efforts have been made in management, the performance of oil and gas pipelines still lags behind that of developed countries. Therefore, strengthening the monitoring of oil and gas pipeline safety, and providing timely warnings and alerts for vibrations and destructive forces to prevent accidents, has become an industry consensus. With the continuous construction and development of oil and gas pipelines, pipeline networks are becoming increasingly complex, and operating environments are becoming more diverse. Traditional copper cable monitoring methods for energy pipelines, due to high transmission losses and numerous safety hazards, are gradually failing to meet the needs of modern energy transmission pipelines and are being replaced by more advanced fiber optic monitoring technologies.

[0003] Optical fibers can be used for vibration and temperature monitoring of oil and gas pipelines. Fiber optic cables can be laid near oil and gas pipelines, monitoring the pipeline based on the propagation characteristics of optical signals within the fiber and the effect of vibration on the fiber. When the pipeline vibrates, the optical signal transmitted in the fiber is affected by the external vibration, causing a frequency change. Sensors detect this change in the optical signal and convert it into a vibration-related electrical signal for analysis and processing, thereby inferring the location and intensity of the vibration signal. Fiber optic vibration monitoring features high sensitivity, high resolution, and strong anti-interference capabilities, enabling accurate monitoring and analysis of subtle vibration signals. Fiber optic temperature monitoring is based on the thermal properties of the fiber and the effect of temperature on it. It typically employs the working principle of a fiber Bragg grating (FBG), which utilizes a periodically etched grating structure within the fiber. When the grating is subjected to temperature changes, the wavelength of the reflected light wave shifts. Since the wavelength of the reflected light wave in the fiber is proportional to temperature, the temperature change can be inferred by measuring the wavelength shift of the grating. Fiber optic temperature monitoring technology boasts advantages such as high sensitivity, rapid response, resistance to electromagnetic interference, explosion-proof capability, and remote measurement. It can effectively monitor pipeline temperature changes, providing early warnings of pipeline anomalies. To date, numerous scholars both domestically and internationally have conducted research on distributed fiber optic applications and monitoring principles. In the area of ​​distributed fiber optic applications, scholar Zhang Yu designed a monitoring system for submarine optical-electric composite cables using a distributed fiber optic vibration sensor and Φ-OTDR technology. Ultimately, spatiotemporal spectrograms were used to characterize signal changes at fault and non-fault points, enabling real-time monitoring and location of submarine cable faults. Regarding the principles of distributed fiber optic monitoring, scholar Fan Denghua conducted in-depth analysis and experimental research on two structures: a distributed fiber optic vibration sensor based on the Sagnac structure and an optical time-domain reflectometer based on the Michelson structure. Based on this, a novel distributed fiber optic vibration sensor structure was designed, significantly improving the system's monitoring performance. The application of fiber optics in oil and gas pipelines enables real-time monitoring and remote operation of energy supply, improving energy transmission efficiency and security.

[0004] However, due to the diversity of pipeline burial environments, the quality of optical fiber performance, and the differences in management measures, the application of optical fiber in pipelines still needs to be evaluated through the technical parameters of optical fiber and management indicators, so as to provide a suitable reference for the subsequent burial and management of optical fiber in other similar locations.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] The purpose of this invention is to provide a method, system, and medium for predicting the application status of optical fibers in pipelines. Starting from the need for application status assessment in distributed optical fiber monitoring systems, this invention proposes a fuzzy comprehensive evaluation method for distributed optical fiber application status based on combined weighting using fuzzy mathematics to achieve optical fiber application status assessment. Furthermore, it employs the idea of ​​combined neural networks to make reasonable predictions about the application status of optical fiber monitoring systems. This invention, based on fuzzy comprehensive evaluation and combined neural networks, achieves the assessment and prediction of the application status of optical fibers in pipelines, solving the problems of insufficient comprehensiveness and low prediction accuracy in existing technologies for assessing the application status of optical fibers in pipelines.

[0007] This invention is achieved through the following technical solution:

[0008] In a first aspect, the present invention provides a method for predicting the application status of optical fiber in pipelines, the method comprising:

[0009] The parameters affecting the application of pipeline optical fiber monitoring systems were screened, and a multi-level parameter system for the application of distributed optical fibers was constructed. The qualitative and quantitative parameters in the multi-level parameter system were processed to determine their corresponding degree values.

[0010] The subjective and objective weights of the influencing parameters of distributed optical fiber application are calculated based on the combined assignment method to obtain the subjective and objective weights; based on the subjective and objective weights, the final weights of the influencing parameters are determined based on game theory.

[0011] Based on the degree value and the final weight of the influence parameters, the final degree value of the pipeline optical fiber monitoring system is determined; combined with fuzzy comprehensive evaluation theory, the impact analysis of distributed optical fiber application is carried out to obtain the impact results.

[0012] Based on the impact results, a combined neural network model is constructed and trained; based on the trained combined neural network model, the application of distributed optical fiber is predicted, and the prediction results are obtained.

[0013] Furthermore, the qualitative and quantitative impact parameters in the multi-level impact parameter system are processed separately to determine their corresponding degree values, including:

[0014] The quantitative impact parameters in the multi-level impact parameter system are normalized to determine the first degree value corresponding to the quantitative impact parameters;

[0015] Membership degrees are calculated for qualitative influence parameters in a multi-level influence parameter system to determine the second degree value corresponding to the qualitative influence parameters.

[0016] Furthermore, based on the combined assignment method, the subjective and objective weights of the influencing parameters of distributed optical fiber applications are calculated to obtain subjective and objective weights. Based on these subjective and objective weights, the final weights of the influencing parameters are determined using game theory, including:

[0017] Based on the improved analytic hierarchy process, the subjective weights of the influencing parameters of distributed optical fiber applications are determined.

[0018] Based on the entropy weight method, the objective weights of the influencing parameters of distributed optical fiber applications are determined.

[0019] Based on subjective and objective weights, the final weights of the influencing parameters are determined using game theory.

[0020] Furthermore, based on the improved analytic hierarchy process, the subjective weights of the influencing parameters of distributed optical fiber applications are determined, including:

[0021] Taking into account industry standards, the importance scores of each grassroots impact parameter were obtained;

[0022] The importance scores of each grassroots impact parameter are quantified into comparison factors, and a first judgment matrix is ​​constructed based on each comparison factor.

[0023] Based on the first judgment matrix, the row mean obtained after column normalization is the desired subjective weight.

[0024] Furthermore, based on the entropy weight method, the objective weights of the influencing parameters of distributed optical fiber applications are determined, including:

[0025] For the selected n influencing parameters and m sets of influencing parameter data, the second judgment matrix is ​​constructed after normalizing the influencing parameter data.

[0026] The range transformation method is used to normalize the data in the second judgment matrix before calculating the ratio of data indicators;

[0027] Based on the normalized data, the information entropy of each influencing parameter is obtained, and the entropy weight of each influencing parameter is calculated as the objective weight.

[0028] Furthermore, the calculation expression that affects the result is:

[0029]

[0030] In the formula, V is the level factor, which affects the result; h is the score of the four application level, with a value of [1,2,3,4]; k is an undetermined coefficient, with k=1. These represent the membership values ​​for each level.

[0031] Furthermore, a combined neural network model is constructed, including:

[0032] A combined neural network model was constructed using the controlled variable method. The combined neural network model was a CNN-BiGRU combined neural network model that incorporated an attention mechanism.

[0033] The optimization steps in building a combinatorial neural network model are as follows:

[0034] Given the importance of the number of layers in the BiGRU network, we tested the predictive effect of increasing the model depth by continuously increasing the number of layers in the BiGRU network.

[0035] With the basic parameters of the attention and CNN modules kept constant, we tested the impact of different numbers of BiGRU layers on the prediction results.

[0036] Secondly, the present invention provides a pipeline optical fiber application prediction system, which uses the aforementioned pipeline optical fiber application prediction method; the system includes:

[0037] The influencing parameter construction unit is used to screen the influencing parameters of the pipeline optical fiber monitoring system and construct a multi-level influencing parameter system for distributed optical fiber applications.

[0038] The degree value determination unit is used to process the qualitative and quantitative influence parameters in the multi-level influence parameter system respectively and determine their corresponding degree values;

[0039] The weight calculation unit is used to calculate the subjective and objective weights of the impact parameters of distributed optical fiber application based on the combined assignment method, and obtain the subjective weights and objective weights; based on the subjective weights and objective weights, the final weights of the impact parameters are determined based on game theory.

[0040] The impact analysis unit is used to determine the final degree value of the pipeline optical fiber monitoring system based on the degree value and the final weight of the impact parameters; combined with fuzzy comprehensive evaluation theory, it performs impact analysis on the application of distributed optical fiber and obtains the impact results.

[0041] The prediction unit is used to construct and train a combined neural network model based on the impact results; based on the trained combined neural network model, it predicts the application of distributed optical fiber and obtains the prediction results.

[0042] Furthermore, the weight calculation unit includes:

[0043] The first weight determination sub-unit is used to determine the subjective weights of the influencing parameters of distributed optical fiber applications based on the improved analytic hierarchy process.

[0044] The second weight determination subunit is used to determine the objective weights of the influencing parameters of distributed optical fiber applications based on the entropy weight method.

[0045] The final weight calculation subunit is used to determine the final weight of the influencing parameters based on game theory, according to subjective and objective weights.

[0046] Thirdly, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for predicting the application status of optical fiber in pipelines.

[0047] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0048] 1. This invention discloses a method, system, and medium for predicting the application status of optical fibers in pipelines. Starting from the need for application status assessment of distributed optical fiber monitoring systems, this invention proposes a fuzzy comprehensive evaluation method for distributed optical fiber application status based on combined weighting using fuzzy mathematics to achieve optical fiber application status assessment. Furthermore, it employs the idea of ​​combined neural networks to make reasonable predictions about the application status of optical fiber monitoring systems. This invention, based on fuzzy comprehensive evaluation and combined neural networks, achieves the assessment and prediction of the application status of optical fibers in pipelines, solving the problems of insufficient comprehensiveness and low prediction accuracy in existing technologies for assessing the application status of optical fibers in pipelines.

[0049] 2. This invention takes into account the influence of the complex structure and operating environment of the fiber optic monitoring system, and establishes a comprehensive and reasonable parameter system for the application of the pipeline fiber optic monitoring system, so as to obtain a more accurate and effective system application information.

[0050] 3. The pipeline fiber optic monitoring system application evaluation and analysis method based on cooperative game theory and fuzzy comprehensive evaluation method of the present invention can more accurately evaluate the application status of the pipeline fiber optic monitoring system.

[0051] 4. This invention uses a combined neural network to predict the application status of fiber optic monitoring systems, which can effectively improve prediction capabilities and system performance, providing users with more reliable monitoring and decision support. Attached Figure Description

[0052] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0053] Figure 1 This is a flowchart of a method for predicting the application of optical fiber in pipelines according to the present invention;

[0054] Figure 2 This is a flowchart illustrating the screening process for influencing parameters in the application of the pipeline fiber optic monitoring system of this invention.

[0055] Figure 3 This invention provides a parameter system that influences the application of the pipeline fiber optic monitoring system.

[0056] Figure 4 This is a detailed flowchart of a method for predicting the application status of optical fiber in pipelines according to the present invention;

[0057] Figure 5 This is a diagram of the CNN structure of the present invention;

[0058] Figure 6 This is a structural diagram of the BiGRU of the present invention;

[0059] Figure 7 This is a graph showing the prediction results of the present invention;

[0060] Figure 8 This is a block diagram of a pipeline optical fiber application prediction system according to the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0062] Example 1

[0063] like Figure 1 As shown, the present invention provides a method for predicting the application status of optical fiber in pipelines, the method comprising:

[0064] Step 1: Based on pipeline fiber optic management specifications and related academic papers, screen the influencing parameters of pipeline fiber optic monitoring system application and construct a multi-level influencing parameter system for distributed fiber optic applications; process the qualitative and quantitative influencing parameters in the multi-level influencing parameter system respectively and determine their corresponding degree values.

[0065] In this embodiment, step 1 mainly involves screening the influencing parameters of distributed optical fiber applications and constructing the influencing parameter system, as detailed below:

[0066] (1) Parameter selection approach

[0067] The selection of influencing parameters is crucial for determining the project's impact on outcomes. They are not only a means of measuring success but also the foundation for strategy development, decision-making, and progress monitoring. The appropriate selection of influencing parameters directly affects the correctness and rationality of the final impact. Therefore, to develop a scientific and systematic system of influencing parameters and rationally select various influencing parameters in the pipeline fiber optic monitoring process, this project, based on the standards for risk assessment of oil and gas pipeline integrity evaluation, scientific and technological achievement evaluation, the effectiveness evaluation of fiber optic sensing technology, and the application of distributed fiber optic early warning technology in oil and gas pipeline inspection, initially selects and formulates indicators (i.e., influencing parameters) according to the principle of indicators. Then, mathematical methods are used to analyze, screen, and optimize the influencing parameter system, correcting or deleting undesirable influencing parameters, ultimately establishing a complete influencing parameter system. Specific steps are as follows: Figure 2 As shown.

[0068] (2) Constructing a multi-level influencing parameter system for distributed optical fiber applications

[0069] A multi-level influencing parameter system for distributed optical fiber applications is a comprehensive evaluation of the performance, reliability, security, and scalability of optical fiber systems in specific fields. This system focuses on performance indicators such as data transmission rate, response time, accuracy, and precision, as well as system stability, data integrity, security, and scalability. Simultaneously, it comprehensively considers factors such as return on investment, maintenance costs, and user satisfaction to fully assess the actual application of the optical fiber system. This multi-level influencing parameter system can effectively guide the optimization and improvement of optical fiber systems, enhance their applicability and performance in various fields, and promote the further application and development of optical fiber technology. Based on the above process, a multi-level influencing parameter system for distributed optical fiber applications is constructed as follows: Figure 3 As shown.

[0070] Specifically, the influencing parameters include operating environment control S 11 Emergency Response Plan S 12 Alarm Record S 13 Wait, see details Figure 3 .

[0071] (3) Calculation of the impact parameters of the application of the pipeline fiber optic monitoring system

[0072] 3a) Normalize the quantitative influence parameters in the multi-level influence parameter system to determine the first degree value corresponding to the quantitative influence parameters;

[0073] Fiber optic monitoring systems for pipelines are playing an increasingly important role in engineering and structural monitoring. The impact analysis of their application involves various quantitative influence parameters and complex data processing methods. During the analysis, due to the different dimensions of the various influence parameters, their values ​​also differ significantly. To avoid the loss of a large amount of data due to direct evaluation analysis, the state variables of the influence parameters are first processed, transforming them into relative application assessment values. Furthermore, the influence parameters are mainly categorized as either more stable with larger values ​​or more stable with smaller values. Their relative application assessment values ​​also need to be calculated according to the characteristics of each influence parameter. The specific calculation formula is as follows:

[0074] 1) For influence parameters where smaller values ​​generally indicate better application performance, the formula for normalizing these parameters is as follows:

[0075]

[0076] For influence parameters where larger values ​​generally indicate better application performance, the formula for normalizing these parameters is as follows:

[0077]

[0078] Where: p i Qmax represents the degree of application, i.e., the first degree value, with a value in [0,1]; Qmax and Qmin are the limit and minimum values ​​under the normal range of the m-th influencing parameter, respectively, and Qi is the measured data value.

[0079] 3b) Calculate the membership degree of the qualitative influence parameters in the multi-level influence parameter system to determine the second degree value corresponding to the qualitative influence parameters.

[0080] For descriptive state-related impact parameters in static impact parameters, such as equipment operating environment control and early warning management mechanisms, there is a lack of clear classification standards between the various assessment levels of these impact parameters, which may belong to two adjacent state levels. Fuzzy statistical experimental methods are used to obtain the membership degree of these impact parameters. A pre-designed questionnaire is distributed to experts and on-site operators using an expert scoring method, as shown in the table below:

[0081] Table 1. Expert Deterioration Survey Form

[0082]

[0083] The meaning of Table 1: This represents the expert's level of certainty regarding the assessment level, and each level must be checked with exactly one checkmark. After organizing the survey results, the membership degree of the static indicators is obtained, calculated using the following expression:

[0084]

[0085] In the formula, l ij.k Expert k (k = 1, 2, ..., D) considers the evaluation level of this static indicator to be v. i A scale of understanding (i = 1, 2, 3, 4); l ij ′ represents the evaluation level of this static indicator as v by all experts. i The average scale of the degree of grasp; ij For l ij The normalized value of ′; finally, the results of the expert survey were obtained.

[0086] For qualitative state-related parameters among the static influence parameters, such as fiber optic burial depth records and system downtime handling contingency plans, the evaluation level of these parameters is determined by their presence or absence, thus corresponding to "good" and "unacceptable" levels, respectively. Different qualitative state-related parameters are classified into application status levels based on their degree of impact on the pipeline fiber optic monitoring system. For example, the absence of real-time waterfall chart transmission will not affect the normal operation of the system, but will only affect the accuracy of system monitoring, thus corresponding to "good" and "qualified" levels, respectively.

[0087] Step 2: Calculate the subjective and objective weights of the influencing parameters of distributed optical fiber application based on the combined assignment method to obtain the subjective and objective weights; determine the final weights of the influencing parameters based on game theory according to the subjective and objective weights.

[0088] Step 2 specifically includes:

[0089] Step 21: Based on the improved analytic hierarchy process, determine the subjective weights of the influencing parameters of distributed optical fiber applications;

[0090] Step 22: Determine the objective weights of the influencing parameters of distributed optical fiber application based on the entropy weight method;

[0091] Step 23: Determine the final weights of the influencing parameters based on game theory, according to subjective and objective weights.

[0092] Specifically, the Analytic Hierarchy Process (AHP) is currently the most commonly used method for determining weights in the field of state assessment research. It possesses high practicality and simplicity. However, the simplified scale often deviates from people's ideal expectations, potentially causing errors in weight calculation. Furthermore, the ambiguity caused by dispersed comparisons necessitates a final consistency check. Based on this, this invention proposes an improved Analytic Hierarchy Process (AHP) for determining the weights of the indicator layer; step 21 is as follows:

[0093] This invention uses a score-based method to replace the 1-9 scale method used in traditional AHP: First, by comprehensively considering industry standards, the importance score S of each grassroots influence parameter is obtained;

[0094] Secondly, the importance scores of each grassroots influencing parameter are quantified into a comparison factor c. ij (i, j = 1, 2, ..., n), based on each comparison factor c ij Construct the first judgment matrix C1:

[0095]

[0096] c ij The mathematical expression is:

[0097] c ij =S i / S j (5)

[0098] In the formula, a ij To construct the value of the i-th row and j-th column of the scaling matrix, S i Let be the importance score of the i-th influencing factor.

[0099] Then, based on the first judgment matrix, the row mean is obtained after column normalization, which is the desired subjective weight w, and its mathematical expression is:

[0100]

[0101] In the formula, c ij * represents the column-normalized value of the data in the i-th column and j-th row.

[0102] Specifically, entropy is a measure of the degree of disorder in a system. The entropy weighting method uses entropy values ​​to measure the dispersion of each underlying influencing parameter, thereby determining the degree of influence (i.e., weight) of that parameter on the overall result. Step 22 is as follows:

[0103] For the selected n influencing parameters and m sets of influencing parameter data, the influencing parameter data is normalized and then a second judgment matrix C2(m×n) is constructed:

[0104]

[0105] In the formula, X ij For the j-th influencing parameter, there is the i-th set of data;

[0106] The range transformation method (i.e., after processing, the worst value of each attribute is 0 and the best value is 1) is used to normalize the data in the second judgment matrix C2 before calculating the data index ratio z. ij Its mathematical expression is:

[0107]

[0108] In the formulas, equations (7) and (8) are the range transformation formulas for benefit-type data and cost-type data, respectively, X *X max X min These represent the range-processed value, the maximum value in the column, and the minimum value in the column, respectively. ij For the j-th influence parameter (when z ij When it is 0, take Inz. ij =0).

[0109] Based on the normalized data, the information entropy ej of each influencing parameter is calculated, and finally the entropy weight w of each influencing parameter is calculated, which is the objective weight to be obtained. Its mathematical expression is:

[0110]

[0111] Specifically, the central idea of ​​game theory is to combine several different types of weighting methods as a whole and use cooperative game theory to determine the combined weights, thereby ensuring the rationality and accuracy of the weight determination. Let W(i) be the weight obtained by method i (i = 1, 2, ..., n). The specific steps to finally obtain the combined weight W are as follows:

[0112] a. Calculation of the consistency correlation coefficient:

[0113] Let W(m_i) denote the combined weights obtained by the m-1 methods other than the i-th method. The mathematical expression for the consistency correlation coefficient L(i) is:

[0114]

[0115] In the formula, n is the number of influencing parameters, and "—" indicates that the logarithmic values ​​are averaged.

[0116] b. Calculate the combined weights W'(i):

[0117]

[0118] c. The combined weights can be obtained recursively, that is, the number of weights is decreased by 1 after each call to the formula, until the number of weights is 2.

[0119] d. When the number of weights is 2, the mathematical expression for W'(i) is:

[0120]

[0121] Step 3: Determine the final degree value of the pipeline fiber optic monitoring system based on the degree value and the final weight of the influencing parameters; combine fuzzy comprehensive evaluation theory to conduct an impact analysis on the application of distributed fiber optics from both management and technical perspectives, and obtain the impact results.

[0122] First, the fuzzy comprehensive evaluation theory, as follows:

[0123] In 1965, American scholar Zadeh first used membership functions to express the fuzziness of things. The emergence of this theory broke the classical set theory at the time, laid the theoretical foundation for the development of fuzzy theory, and provided a very good direction for the study of fuzzy phenomena.

[0124] The data on various application scenarios of pipeline fiber optic monitoring systems exhibit a degree of dispersion and fuzziness in their influencing parameters, and the extent of their impact on these scenarios is uncertain, a phenomenon known as fuzzy uncertainty. The essential purpose of the fuzzy comprehensive evaluation method is to utilize membership functions to transform qualitative evaluations of the research object into quantitative evaluations. By comprehensively analyzing the relevant influencing parameters of the research object, a scientifically reasonable result summarizing the impact of these indicators on the application scenarios can be obtained.

[0125] The main steps of fuzzy comprehensive evaluation theory include the following:

[0126] (1) Determine the set of status assessment factors

[0127] Let U be the set of evaluation factors for the research subjects:

[0128] U = {u1, u2, u3, ..., u} m} (17)

[0129] In the formula, m is the number of influencing parameters.

[0130] (2) Determine the comment set

[0131] The evaluation set is the collection V of possible final evaluation results for the research subjects.

[0132] V = {v1, v2, v3, ..., v} n} (18)

[0133] In the formula V i Let be the i-th evaluation result of the device; n is the number of evaluation results.

[0134] (3) Determine the set of weight coefficients for status indicators

[0135] By analyzing the importance of each state evaluation factor in set U to the system application using certain mathematical methods, the weight coefficient set W is determined accordingly.

[0136] W = {w1, w2, w3, ..., w m} (19)

[0137] In the formula w i Let be the weight of the i-th influencing parameter.

[0138] (4) Determine the fuzzy membership matrix of the influencing parameters

[0139] Based on the evaluation factor set U of the research subjects, the membership degree of each state index to the research subjects with respect to the comment set V is determined by quantifying each influencing parameter one by one, thereby obtaining the fuzzy membership matrix.

[0140]

[0141] In the formula r mi For the m-th influencing parameter, the comment set v i The subordinate relationship.

[0142] (5) Fuzzy comprehensive evaluation results (i.e., impact results)

[0143] The obtained comprehensive weight vector and fuzzy membership matrix are synthesized using appropriate fuzzy operators, and the fuzzy comprehensive influence result B is calculated.

[0144]

[0145] In the formula, “o” represents the fuzzy synthesis operator.

[0146] (6) Evaluation Result Analysis

[0147] There are three common methods for analyzing fuzzy comprehensive evaluation results:

[0148] (a) Maximum membership principle: At this point, the application status of the equipment belongs to the i-th evaluation result;

[0149] (b) Weighted average method: Assign values ​​to n evaluation results and use a weighted average method to obtain the evaluation score of equipment application;

[0150] (c) Median method: The application status of the equipment at this time belongs to the i-th evaluation result.

[0151] Second, establish the application evaluation factor set U and the state set.

[0152] a. Establish a set of factors for evaluating application status.

[0153] The state set V established in this invention is a set of evaluations of the possible state levels of the submarine cable as a whole and its components. Combining industry standards and the previous classification standards, and taking into account the opinions and experience of on-site maintenance personnel and experts, the operational health status of submarine cables is divided into four identification levels: normal, attention, abnormal, and critical, corresponding to v1, v2, v3, and v4, respectively, i.e., n=4.

[0154] The application of pipeline fiber optic monitoring systems is evaluated primarily from technical and management perspectives. This invention focuses on cross-linked polyethylene (XLPE) fiber optic composite submarine cables. Based on the above, a multi-level index model for evaluating the application status is established by studying the influencing parameters at these two levels. This model consists of three levels: the evaluation target level reflecting the application status of the pipeline fiber optic monitoring system, the intermediate level divided according to the evaluation perspective, and the specific influencing parameter level.

[0155] The application status assessment factor set indicates the perspectives from which the application status of the assessment object needs to be evaluated. Based on the established multi-level influence parameter system for application status assessment, the influence parameters in the selected influence parameter layers are used as the assessment factor set S, and S is used to... i (i = 1, 2, ..., n) represents the i-th part; taking the management level as an example, s 11 To s 19 These respectively represent the operating environment control, emergency plan, alarm record, system page display, fiber optic calibration, data management, fiber optic laying profile, early warning management, and operation inspection of the pipeline fiber optic monitoring system.

[0156] b. Establish an application status evaluation level set U

[0157] The evaluation level set V established in this invention is a set of evaluation comments on the overall application status of the pipeline fiber optic monitoring system and its application at both the management and technical levels. Combining industry standards, operating specifications, and fiber optic-related research literature, and referring to the opinions and experience of on-site maintenance personnel and experts, the application status of the pipeline fiber optic monitoring system is divided into four levels: Excellent, Good, Qualified, and Unsatisfactory, corresponding to v1, v2, v3, and v4 respectively, i.e., n=4.

[0158] Third, evaluation factors with veto power.

[0159] In the practical application of fiber optic monitoring systems, the parameter system established in the above steps, where certain indicators fail to meet certain conditions, directly impacts the system's performance. Based on existing fiber optic monitoring system operating specifications and the maintenance experience of on-site personnel, some high-weighted parameters, such as the maximum monitoring distance of the fiber optic system, if altered to a level that does not meet actual monitoring needs (i.e., the system cannot monitor the entire pipeline length), will result in extremely poor system performance or even complete failure to meet normal application requirements. Therefore, this invention introduces a "veto power evaluation factor" in its fuzzy comprehensive evaluation method. However, for some parameters with multi-level thresholds, such as system response time, if it consistently exceeds the highest threshold, alarms from the fiber optic system may not receive timely feedback, potentially impacting pipeline operation. Therefore, this invention introduces a "secondary veto power evaluation factor." If these evaluation indicators reach a substandard level, subsequent evaluations are unnecessary; the pipeline fiber optic monitoring system's application is directly assessed as substandard.

[0160] The veto power evaluation factors of this invention include monitoring distance and fiber optic attenuation, while secondary veto power evaluation factors should include positioning accuracy and response time. In subsequent status assessments, if any veto power factor exceeds a set threshold or is at a substandard level, or if a secondary veto power factor consistently exceeds the highest threshold, no further assessment is required, and the fiber optic monitoring system application is assessed as substandard. Furthermore, due to the special nature of the veto power, these influencing parameters are no longer subject to weighting and fuzzy comprehensive evaluation.

[0161] Fourth, the membership function of the indicator layer.

[0162] Membership degree is a mathematical method in fuzzy theory, and selecting a suitable membership function is a key point in fuzzy comprehensive evaluation methods. The previous section analyzed the method of weighting parameters related to application conditions. Next, we clarify the membership functions that match each influencing parameter and perform fuzzification processing on the influencing parameters related to the application of pipeline optical fibers determined above. Commonly used quantitative index membership functions in engineering include trapezoidal, ridge, triangular, and normal distributions. Practice has shown that the triangular membership function is too coarse in handling edge information of state levels, while the trapezoidal membership function easily loses membership boundary information, resulting in unsatisfactory processing effects on fuzzy and uncertain information. Therefore, this paper combines the advantages and disadvantages of the two membership functions, using a distribution function combining semi-trapezoidal and triangular distributions to establish a quantitative state index membership function. For descriptive and qualitative state index membership functions, fuzzy statistical experiments are often used. Through the idea of ​​set-valued statistics, the fuzzy shadow function formed by the superposition of multiple expert evaluation intervals is processed for evaluation, which can more accurately reflect the evaluation criteria of the experts.

[0163] Fifth, the influence of the synthesis of results

[0164] Based on the constructed multi-level application influence parameter system, this invention establishes a three-level fuzzy comprehensive evaluation model to assess the overall application status of the pipeline fiber optic monitoring system at both the management and technical levels. First, fuzzy comprehensive evaluation is performed on each influence parameter to obtain the membership state matrix at the technical and management levels. Then, fuzzy comprehensive evaluation is performed on this matrix to finally obtain the overall application status of the fiber optic system. The evaluation and analysis method flow is as follows: Figure 4 As shown.

[0165] By calculating using equations (17) and (18), the membership values ​​of each influencing parameter at different state levels are obtained, and finally, the membership matrix R is obtained. i for:

[0166]

[0167] In the formula R i Let r be the fuzzy evaluation matrix for the i-th part; mi For the m-th influencing parameter, the comment v i The subordinate relationship.

[0168] Weighted average fuzzy computation sorts multiple influencing parameters according to their importance, comprehensively considering the impact of major and minor factors on the result. It is often used in fuzzy comprehensive evaluation result calculation, and its calculation formula is as follows:

[0169] B i =w i ·R i =[b1,b2,b3,b4] (23)

[0170] In the formula B i w represents the membership matrix for different application levels of the i-th component. i The weight vector w between the m state indices of the i-th part i =[w1,w2,…,w m b1, b2, b3, b4 are the membership values ​​of this part corresponding to excellent, good, qualified, and unqualified, respectively.

[0171] Based on this, the comprehensive state membership matrix B = [B1, B2, ... B] of each part of the fiber optic monitoring system is obtained. i The sum of the weight vectors of T and its components is W = [W1, W2, ..., W...]. n The membership matrix D for evaluating the application status of the fiber optic monitoring system is calculated as follows:

[0172] D = W·B = [d1,d2,d3,d4] (24)

[0173] In the formula, d1, d2, d3, and d4 are the membership values ​​corresponding to the application status of the fiber optic monitoring system as excellent, good, qualified, and unqualified, respectively.

[0174] The impact of the application status of the fiber optic monitoring system is obtained by using a weighted average method on the membership values ​​of each level of the application status. The calculation expression is as follows:

[0175]

[0176] In the formula, V is the grading factor; h is the score of the four application level, with values ​​of [1, 2, 3, 4]; k is an undetermined coefficient, which is taken as k = 1 in this paper; These represent the membership values ​​for each level.

[0177] The final scores are converted into a percentage system, and the final application level of the fiber optic monitoring system is determined based on the table.

[0178] Step 4: Based on the impact results, construct and train a combined neural network model; the combined neural network model is a CNN-BiGRU combined neural network model that integrates an attention mechanism; based on the trained combined neural network model, predict the application of distributed optical fiber and obtain the prediction results.

[0179] Specifically, in the prediction model framework of step 4, compared with traditional neural networks, CNNs can extract the inherent features of the dataset more efficiently and accurately, but cannot effectively utilize the temporal features present in the dataset; while a single BiGRU neural network can effectively extract and utilize the temporal features present in the dataset, it is relatively inefficient in utilizing the correlation features present in the dataset. The GRU and BiGRU structures are as follows: Figure 5 , Figure 6 As shown, using a single CNN or BiGRU neural network to process long-term sequences or multi-dimensional input data can lead to problems such as insufficient data feature mining and loss of sequence or related features. Therefore, this invention complements the advantages of both, constructing a CNN-BiGRU combined neural network and integrating an attention mechanism to process the dataset, thereby obtaining the best prediction results.

[0180] The prediction process is as follows:

[0181] a. Data Processing

[0182] The data of each index in the parameter system affecting optical fiber application established in this invention are taken and imported with a daily step size. To solve the problem of large errors caused by abnormal data in the dataset, the mean square value method is used to process the dataset, and its mathematical expression is as follows:

[0183]

[0184] In the formula, N is the number of data in this category, and x i Let x be the data value for the i-th evaluation. If |x i If -u|>5δ, then determine x i If it is an outlier, remove it.

[0185] To address the negative impact of large differences in the scale and dimensions of the input data, the maximum and minimum values ​​of each data type in the dataset are used as benchmarks for dataset normalization before model training. The mathematical expression for this normalization is as follows:

[0186]

[0187] In the formula, D n D represents the data obtained after normalization; D represents the data before processing. MIN D MAX These are the minimum and maximum values ​​for this type of data, respectively.

[0188] b. Predictive Evaluation Indicators

[0189] To evaluate the prediction accuracy of the constructed CNN-BiGRU combined neural network model with an attention fusion mechanism, the maximum percentage of prediction error σ was selected. E-max Root mean square error R MSE Mean percentage error M APE The performance of the model is evaluated, and its mathematical expression is as follows:

[0190]

[0191] In the formula, y i y is the true value of the i-th sample point; pi is the predicted value of the i-th sample point; N is the number of sample points.

[0192] To ensure the scientific rigor and consistency of the model training and prediction processes, the R-value of the model prediction results is... MSE and M APE All values ​​are averages obtained by inputting the test set data into the model. The model prediction results and actual values ​​of the test set are selected to plot the prediction curve to further demonstrate and verify the superiority of the selected model.

[0193] c. Model Training

[0194] This invention employs a controlled variable method to optimize the model. Given the importance of the number of BiGRU network layers, the predictive effect of increasing model depth is tested by continuously increasing the number of BiGRU network layers. The basic parameters of the attention and CNN modules are kept constant, and the impact of different numbers of BiGRU layers on the prediction results is tested. Experiments show that when the number of BiGRU network layers is 2, all influencing parameters reach their optimal values. Afterward, the error rate begins to rise, indicating model overlearning. The results are shown in Table 3.

[0195] Table 3 Results of Model Network Layer Optimization Experiment

[0196] GRU network layers <![CDATA[σ E-max ]]> <![CDATA[M APE ]]> <![CDATA[R MSE ]]> 1 3.81％ 1.63％ 1.47％ 2 3.35％ 1.49％ 1.34％ 3 3.92％ 1.55％ 1.40％ 4 7.61％ 4.23％ 3.44％

[0197] The CNN neural network uses 10 convolutional kernels, a kernel size of 2, a stride of 1, and employs the same convolution method, followed by valid max pooling after consecutive convolutions. The BiGRU neural network uses a sliding window data reading mode, with a time step of 10 and a batch size of 30. The Adam algorithm iteratively updates the weights, continuously updating the weights and biases of each neuron through momentum and adaptive learning rate, thereby optimizing the output value of the loss function. To address potential overfitting, Dropout is used to randomly discard network nodes with a certain probability during training. An attention mechanism is integrated into the training of both CNN and BiGRU to improve model accuracy. The mathematical expression of the model's loss function is:

[0198]

[0199] In the formula, F loss , λ act(t) , λ pred(t) Let and n be the model's loss function, the actual application score and the predicted score at time t, and the number of training samples, respectively.

[0200] In practical implementation, to verify the feasibility and accuracy of the method established in this invention, this invention takes the DVS monitoring system in a certain area of ​​Suining, Sichuan as the research object, selects indicators at the management and technical levels, and constructs an influencing parameter system as follows: Figure 3 As shown, the application evaluation method proposed in this invention was used to evaluate and analyze the impact parameters at the management and technical levels. The scores of the impact parameters and the corresponding weights of each impact parameter were calculated through the above process, as shown in Tables 3 and 4 below:

[0201] Table 3. Values ​​and corresponding weights of parameters influencing management level.

[0202] Influencing parameters Degree value Weight Operating environment management 0.85 0.15 Emergency Plan 0.9 0.1 Alarm Records 0.7 0.1 System page display 0.6 0.05 Fiber optic calibration 0.9 0.15 Data Management 1 0.15 Image of fiber optic cable laying 0.8 0.15 Early warning management 0.8 0.05 Operational Inspection 0.9 0.1

[0203] Table 4. Degrees and corresponding weights of parameters influencing management level

[0204] Influencing parameters Degree value Weight Monitoring distance 0.9 0.15 Positioning accuracy 0.7 0.15 False alarm rate 0.8 0.15 Response time 0.7 0.15 Fiber attenuation 0.8 0.15 Self-control capability 0.8 0.05 Data processing capabilities 0.6 0.05 Equipment reliability 0.8 0.05 Intrusion detection types 0.4 0.1

[0205] Using a weighted average method, comprehensively considering the degree values ​​and corresponding weights of each influencing parameter, the final score (degree value) for the management level of the fiber optic monitoring system was calculated to be 0.8525, which translates to 85.25 on a percentage scale. Combined with the evaluation levels in Table 2, the application level of the management level of the fiber optic monitoring system is rated as excellent. The evaluation score (degree value) for the application level of the technical level of the fiber optic monitoring system is 0.735, which translates to 73.5 on a percentage scale, corresponding to a good evaluation level. In practical applications, excellent management measures can compensate for deficiencies in the system's performance. Therefore, during operation, the management level of the fiber optic monitoring system is more important than the technical level. Based on the above calculations, the weight of the management level is 0.6, and the weight of the technical level is 0.4. Therefore, the overall application score of the fiber optic monitoring system is 80.55, corresponding to a good evaluation level.

[0206] The selected dataset is evaluated and quantified using the application status assessment model proposed in this invention to form the input feature set. This input feature set is then imported into a CNN-BiGRU combined neural network model to predict the application status of pipeline optical fibers. The prediction results are as follows: Figure 7 As shown, the three evaluation indicators of the results are: the maximum percentage of prediction error σ. E-max Root mean square error R MSE Mean percentage error M APE They achieved relatively good rates of 3.35%, 1.34%, and 1.49%, respectively. From... Figure 6 It can be seen that the overall application score of the fiber optic monitoring system fluctuated between 80 and 90 during this period. This was mainly due to the operation of the on-site personnel, but no serious abnormal conditions were encountered. Comparing the actual data of the submarine cable during this period, it can be seen that the trend of the predicted results is consistent with the trend of the application of the fiber optic monitoring system, and at the same time, it has good accuracy.

[0207] Based on literature review and relevant research, this invention combines relevant research results and regulations such as "Definition and Test Method of Linear and Deterministic Properties of Single-Mode Optical Fiber and Cable (G.650.1)," "GB 50312-2007 Acceptance Specification for Integrated Cabling Engineering (including explanatory notes)," and "Technical Conditions for Fiber Optic Monitoring Systems for Direct-Buried Heating Pipelines" (T / CDHA11-2022). Considering the complex structure and operating environment of fiber optic monitoring systems, this invention establishes a comprehensive and reasonable parameter system for influencing the application status of pipeline fiber optic monitoring systems, thereby obtaining more accurate and effective information on system application. 2) The pipeline fiber optic monitoring system application status evaluation and analysis method based on cooperative game theory and fuzzy comprehensive evaluation can more accurately assess the application status of pipeline fiber optic monitoring systems. 3) Using combined neural networks to predict the application status of fiber optic monitoring systems can effectively improve prediction capabilities and system performance, providing users with more reliable monitoring and decision support.

[0208] Example 2

[0209] like Figure 8 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a pipeline optical fiber application prediction system, which uses a pipeline optical fiber application prediction method from Embodiment 1; the system and the pipeline optical fiber application prediction method from Embodiment 1 have a one-to-one functional correspondence; the system includes:

[0210] The influencing parameter construction unit is used to screen the influencing parameters of the pipeline optical fiber monitoring system and construct a multi-level influencing parameter system for distributed optical fiber applications.

[0211] The degree value determination unit is used to process the qualitative and quantitative influence parameters in the multi-level influence parameter system respectively and determine their corresponding degree values;

[0212] The weight calculation unit is used to calculate the subjective and objective weights of the impact parameters of distributed optical fiber application based on the combined assignment method, and obtain the subjective weights and objective weights; based on the subjective weights and objective weights, the final weights of the impact parameters are determined based on game theory.

[0213] The impact analysis unit is used to determine the final degree value of the pipeline optical fiber monitoring system based on the degree value and the final weight of the impact parameters; combined with fuzzy comprehensive evaluation theory, it performs impact analysis on the application of distributed optical fiber and obtains the impact results.

[0214] The prediction unit is used to construct and train a combined neural network model based on the impact results; based on the trained combined neural network model, it predicts the application of distributed optical fiber and obtains the prediction results.

[0215] As a further implementation, the weight calculation unit includes:

[0216] The first weight determination sub-unit is used to determine the subjective weights of the influencing parameters of distributed optical fiber applications based on the improved analytic hierarchy process.

[0217] The second weight determination subunit is used to determine the objective weights of the influencing parameters of distributed optical fiber applications based on the entropy weight method.

[0218] The final weight calculation subunit is used to determine the final weight of the influencing parameters based on game theory, according to subjective and objective weights.

[0219] The execution process of each unit can be carried out according to the steps of the pipeline optical fiber application prediction method in Example 1, and will not be described in detail in this example.

[0220] Meanwhile, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for predicting the application status of optical fiber in pipelines.

[0221] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0222] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0223] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0224] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0225] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting the application status of optical fiber in pipelines, characterized in that, The method includes: The parameters affecting the application of the pipeline optical fiber monitoring system are screened to construct a multi-level parameter system for the application of distributed optical fiber; the qualitative and quantitative parameters in the multi-level parameter system are processed to determine their corresponding degree values. The subjective and objective weights of the influencing parameters of distributed optical fiber application are calculated based on the combined assignment method to obtain the subjective and objective weights; based on the subjective and objective weights, the final weights of the influencing parameters are determined based on game theory. The final degree value of the pipeline fiber optic monitoring system is determined based on the degree value and the final weight of the influencing parameters. By combining fuzzy comprehensive evaluation theory, an impact analysis of the application of distributed optical fiber is conducted, and the impact results are obtained. Based on the aforementioned impact results, a combined neural network model is constructed and trained; based on the trained combined neural network model, the application of distributed optical fibers is predicted, and the prediction results are obtained.

2. The method for predicting the application status of optical fiber in pipelines according to claim 1, characterized in that, The qualitative and quantitative influence parameters in the multi-level influence parameter system are processed separately to determine their corresponding degree values, including: The quantitative influence parameters in the multi-level influence parameter system are normalized to determine the first degree value corresponding to the quantitative influence parameters. Membership degree calculation is performed on the qualitative influence parameters in the multi-level influence parameter system to determine the second degree value corresponding to the qualitative influence parameters.

3. The method for predicting the application status of optical fiber in pipelines according to claim 1, characterized in that, The subjective and objective weights of the influencing parameters of distributed optical fiber application are calculated based on the combined assignment method to obtain the subjective weight and objective weight. Based on the aforementioned subjective and objective weights, the final weights of the influencing parameters are determined using game theory, including: Based on the improved analytic hierarchy process, the subjective weights of the influencing parameters of distributed optical fiber applications are determined. Based on the entropy weight method, the objective weights of the influencing parameters of distributed optical fiber applications are determined. Based on the subjective weights and the objective weights, the final weights of the influencing parameters are determined using game theory.

4. The method for predicting the application status of optical fiber in pipelines according to claim 3, characterized in that, Based on the improved analytic hierarchy process (AHP), the subjective weights of the influencing parameters of distributed optical fiber applications are determined, including: Taking into account industry standards, the importance scores of each grassroots impact parameter were obtained; The importance scores of each grassroots impact parameter are quantified into comparison factors, and a first judgment matrix is ​​constructed based on each comparison factor. Based on the first judgment matrix, the row mean obtained after column normalization is the desired subjective weight.

5. The method for predicting the application status of optical fiber in pipelines according to claim 3, characterized in that, Based on the entropy weight method, the objective weights of the influencing parameters of distributed optical fiber applications are determined, including: For the selected n influencing parameters and m sets of influencing parameter data, the second judgment matrix is ​​constructed after normalizing the influencing parameter data. The data index ratios are calculated after the data in the second judgment matrix are normalized using the range transformation method. Based on the normalized data, the information entropy of each influencing parameter is obtained, and the entropy weight of each influencing parameter is calculated as the objective weight.

6. The method for predicting the application status of optical fiber in pipelines according to claim 1, characterized in that, The calculation expression for the impact result is as follows: In the formula, V is the level factor, which affects the result; h is the score of the four application level, with a value of [1,2,3,4]; k is an undetermined coefficient, with k=1. These represent the membership values ​​for each level.

7. The method for predicting the application status of optical fiber in pipelines according to claim 1, characterized in that, Constructing a combined neural network model includes: A combined neural network model was constructed using the controlled variable method. The combined neural network model is a CNN-BiGRU combined neural network model that incorporates an attention mechanism. The optimization steps in building a combinatorial neural network model are as follows: Given the importance of the number of layers in the BiGRU network, we tested the predictive effect of increasing the model depth by continuously increasing the number of layers in the BiGRU network. With the basic parameters of the attention and CNN modules kept constant, we tested the impact of different numbers of BiGRU layers on the prediction results.

8. A pipeline fiber optic application prediction system, characterized in that, The system includes: The influencing parameter construction unit is used to screen the influencing parameters of the pipeline optical fiber monitoring system and construct a multi-level influencing parameter system for distributed optical fiber applications. The degree value determination unit is used to process the qualitative and quantitative influence parameters in the multi-level influence parameter system respectively and determine their corresponding degree values. The weight calculation unit is used to calculate the subjective and objective weights of the influence parameters of distributed optical fiber application based on the combined assignment method, so as to obtain the subjective weight and the objective weight; and to determine the final weight of the influence parameter based on game theory according to the subjective weight and the objective weight. The impact analysis unit is used to determine the final degree value of the pipeline optical fiber monitoring system based on the degree value and the final weight of the impact parameters; and to conduct an impact analysis on the application of distributed optical fiber by combining fuzzy comprehensive evaluation theory to obtain the impact results. The prediction unit is used to construct a combined neural network model and train the model based on the influence results; based on the trained combined neural network model, it predicts the application of distributed optical fiber and obtains the prediction results.

9. The pipeline optical fiber application prediction system according to claim 8, characterized in that, The weight calculation unit includes: The first weight determination sub-unit is used to determine the subjective weights of the influencing parameters of distributed optical fiber applications based on the improved analytic hierarchy process. The second weight determination subunit is used to determine the objective weights of the influencing parameters of distributed optical fiber applications based on the entropy weight method. The final weight calculation subunit is used to determine the final weight of the influencing parameters based on game theory, according to the subjective weight and the objective weight.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a pipeline optical fiber application prediction method as described in any one of claims 1 to 8.

Images