Methods, models, and early warning systems for predicting faults in dense corridors based on meteorological disasters.

Abstract

This invention relates to the field of dense corridor risk prediction technology, specifically to a method, model, and early warning system for dense corridor failure prediction based on meteorological disasters. The prediction method includes: step S21, collecting historical meteorological characteristic sequences and equipment failure rate sequences through a meteorological disaster acquisition unit; step S22, establishing a historical meteorological disaster sample set; step S23, obtaining meteorological characteristic sequences for the period to be predicted through a meteorological condition acquisition unit; step S24, comparing the meteorological characteristic sequences with historical meteorological characteristic sequences one by one; and step S25, outputting the failure rate sequence of the most similar sample through a failure rate output unit as the predicted failure rate. The prediction model is used to implement the above method, and the early warning system incorporates the above prediction model. This invention can effectively predict the probability of equipment failure or overall failure due to meteorological disasters in operating lines.

Description

Technical Field

[0001] This invention relates to the field of dense corridor risk prediction technology, and more specifically, to a method, model, and early warning system for predicting faults in dense corridors based on meteorological disasters. Background Technology

[0002] A dense UHVDC transmission corridor refers to an important transmission corridor consisting of no fewer than two ±800 kV or higher UHVDC lines, with a minimum gap of no more than 100 meters between the pole conductors of two adjacent UHVDC lines. The operating lines covered by this dense transmission corridor are characterized by large capacity, wide impact range, long operating line length, and large transmitted load power, making them highly susceptible to external factors. Furthermore, a simultaneous fault in any line within the corridor could have a significant impact on the power grid at both the sending and receiving ends, potentially triggering at least a Level 3 or higher power grid security incident.

[0003] In addition, due to the characteristics of dense passageways, such as narrow corridors, complex terrain and social environments, the traditional operation and maintenance mode, which mainly relies on manual inspections, is difficult to achieve all-weather monitoring and all-time equipment status perception. Therefore, it is urgent to apply new technologies and methods to carry out theoretical research and application innovation on intelligent early warning for dense passageways, and to develop more suitable multi-dimensional risk identification and early warning models and comprehensive analysis methods for the scope of safety risk impact, so as to achieve autonomous early warning of risks to passageway equipment and environment. Summary of the Invention

[0004] This invention provides a method for predicting faults in dense transportation corridors based on meteorological disasters. It addresses the problem that existing dense transportation corridors are difficult to predict the probability of faults caused by meteorological disasters. By using a similarity algorithm and building a historical database, it can better predict the probability of faults in the main body or the whole of the equipment in the operating line caused by meteorological disasters.

[0005] The dense channel failure prediction method based on meteorological disasters according to the present invention includes the following steps:

[0006] Step S21: Collect historical meteorological characteristic sequences A of the same dense channel under historical meteorological disasters through the meteorological disaster collection unit. # and the failure rate sequence E of the main body of the equipment # ; For the uth type of meteorological disaster A u eigenvalues; For the i-th device body E i The failure rate;

[0007] Step S22: Establish a historical meteorological disaster sample set F and store it in the historical meteorological disaster database.

[0008]

[0009] Among them, F x The x-th sample in the historical meteorological disaster sample set F, For sample F x Failure rate sequence E # , For sample F x Historical meteorological characteristic sequence A # ;

[0010] Step S23: Obtain the meteorological characteristic sequence A for the period to be predicted through the meteorological condition acquisition unit. * , For the uth type of meteorological disaster A u eigenvalues;

[0011] Step S24: Based on the similarity algorithm, the meteorological feature sequence A is compared using the similarity comparison unit. * Each sample F in the historical meteorological disaster sample set F x Historical meteorological characteristic sequence By making comparisons one by one, historical meteorological characteristic sequences can be obtained. With meteorological characteristic sequence A * Most similar sample F x ;

[0012] Step S25: Output the most similar sample F through the failure rate output unit. x Failure rate sequence And serve as the predicted failure rate of the main equipment during the period to be predicted.

[0013] The above method enables the collection of a large number of historical meteorological characteristic sequences A under different meteorological disasters for the same dense channel. # and the failure rate sequence E of the main body of the equipment # This leads to the establishment of a large database. Based on this database, when predicting the probability of failure of various equipment due to meteorological disasters during the forecast period in dense channels, it is possible to obtain historical data most similar to the meteorological conditions of the forecast period based on similarity judgment, and use the failure rate of each equipment in the historical data as the predicted failure rate. Therefore, it is possible to better predict the impact of meteorological disasters on each equipment during the forecast period.

[0014] As a preferred option, in constructing historical meteorological feature sequence A # At that time, for meteorological disaster A u Meteorological disaster A u Actual strength Su and actual duration T u The product of these is used as its eigenvalue; that is,

[0015] The above approach allows for a better consideration of the impact of the intensity and duration of meteorological disasters on the failure rate of the main equipment.

[0016] As a preferred option, in constructing meteorological feature sequence A * At that time, meteorological disaster A u probability of occurrence Predicted intensity and predicted duration The product of these is used as its eigenvalue; that is,

[0017] The above approach allows for a better comprehensive consideration of the impact of the probability of meteorological disasters, the intensity of the forecast, and the duration of the forecast on the calculated forecast failure rate.

[0018] As a preferred option, for any meteorological disaster A u The probability values ​​for blue alerts are set at 0.55, yellow alerts at 0.7, orange alerts at 0.8, and red alerts at 0.9. Therefore, it is possible to obtain quantified probabilities of occurrence based on relevant data from the weather forecasting system.

[0019] As a preferred method, meteorological feature sequence A is obtained based on Euclidean distance. * With historical meteorological characteristic sequence The similarity value Q between them, and the similarity with the meteorological feature sequence A. * The sample F with the smallest similarity value Q x As the most similar sample; among them,

[0020]

[0021] Where U represents the total number of meteorological disasters included in the forecast.

[0022] Through the above methods, the meteorological characteristic sequence A for the period to be predicted can be obtained more effectively. * Similarity determination with all samples in the historical meteorological disaster sample set F.

[0023] As a preferred embodiment, the following steps are also included.

[0024] Step S26: Obtain and output the overall meteorological failure rate L of the operating line through the meteorological early warning unit, wherein,

[0025]

[0026] Among them, Ri For the main body of the equipment E i The influencing factor that affects the status of the operating line, m is the main body of the equipment E i The total number of categories. Therefore, the overall meteorological failure rate L can be obtained through the meteorological early warning unit, and the overall equipment failure rate E of each category can be obtained through comprehensive analysis. i Failure rate By applying weights, it is possible to better predict the probability of failure of dense corridors (operating lines) due to meteorological disasters.

[0027] Furthermore, the present invention also provides a dense channel failure prediction model based on meteorological disasters, which is used to implement any of the above methods, and includes at least:

[0028] A meteorological disaster data acquisition unit is used to implement step S21;

[0029] A historical meteorological disaster database is used to implement step S22;

[0030] Meteorological condition acquisition unit, which is used to implement step S23;

[0031] A similarity comparison unit is used to implement step S24; and,

[0032] Failure rate output unit, which is used to implement step S25.

[0033] Through the above methods, it is possible to better predict the probability of each piece of equipment in a densely packed passage (operating line) failing due to meteorological disasters.

[0034] Preferably, the dense corridor fault prediction model of the present invention can also include, for example, a meteorological warning unit for acquiring and outputting the overall meteorological fault rate L of the line. Therefore, it can better predict the probability of a dense corridor (operating line) failing due to meteorological disasters.

[0035] In addition, the present invention also provides an intelligent early warning system for dense passages, which includes any of the above-mentioned dense passage fault prediction models and an early warning module.

[0036] Through the above, it is possible to better achieve control over the entire operating line and its associated equipment. i It predicts the probability of anomalies during operation and failures due to meteorological disasters, and calculates the comprehensive failure rate caused by both factors. Therefore, it can effectively predict anomalies in densely packed lines under multiple operating conditions. Attached Figure Description

[0037] Figure 1This is a schematic diagram of the multi-dimensional anomaly prediction method and model in Example 1;

[0038] Figure 2 This is a flowchart illustrating the dense channel fault prediction method in Example 2;

[0039] Figure 3 This is a block diagram of the dense channel fault prediction model in Example 2;

[0040] Figure 4 This is a block diagram of the intelligent early warning system in Example 3. Detailed Implementation

[0041] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0042] Example 1

[0043] Combination Figure 1 As shown, this embodiment provides a multi-dimensional anomaly prediction method, which is well applicable to the prediction and early warning of abnormal states of each type of equipment body or the entire dense channel within a dense channel.

[0044] The anomaly prediction method in this embodiment has the following steps:

[0045] Step S11: Collect data on the main body E of the device through the status acquisition unit. i Status data Indicates the main body of the equipment E i The state data of the j-th evaluation parameter in the data;

[0046] Step S12: Construct the terminology evaluation set P and the corresponding numerical evaluation set P * P = {P} k |k∈N +}, P k The k-th evaluation term in the term evaluation set P; For the evaluation term P k The corresponding range of values;

[0047] Step S13: Construct state data based on the term evaluation set P using the data evaluation unit. Terminology Evaluation Matrix Indicates the main body of the equipment E i The evaluation term for the j-th evaluation parameter in the equation;

[0048] Step S14: Construct a term evaluation matrix using numerical units. Numericalized matrix Indicates the main body of the equipment E i The evaluation value of the j-th evaluation parameter in the equation. Evaluation terms The corresponding value range, Random[*] indicates the operation of taking random values;

[0049] Step S15: Obtain the main body E of the equipment through the status evaluation unit. i abnormal factors

[0050]

[0051] Where, ω ij For the main body of the equipment E i The weighting factor of the j-th evaluation parameter, where M is the weighting factor of the main equipment E. i The total number of evaluation parameters, N is the number of the main equipment E participating in the evaluation. i The total number.

[0052] Among them, the main body of the equipment participating in the anomaly evaluation within the dense channel (operating line) can be represented as equipment body E, E = {E i |i∈N +}, E i Let be the equipment body of the i-th category in the equipment body library E. It is understood that in a dense channel (running line), each category of equipment body can have multiple units.

[0053] In this embodiment, the main equipment library E can include eight elements, namely, poles (E1), insulators (E2), fittings (E3), conductors (E4), foundations (E5), grounding devices (E6), channel environment (E7), and auxiliary facilities (E8).

[0054] It is understandable that, for different equipment entities E i It can construct evaluation parameters of corresponding quantity and type.

[0055] Taking the tower (E1) as an example, this embodiment can construct 9 evaluation parameters, namely, the main material bending (E... 11 ), auxiliary material bending (E) 12 ), tilt (E) 13 ), corrosion (E) 14 ), missing components (E) 15 ), slack in the tension line (E) 16 ), Tower body cracks (E 17 ), welding cracks (E)18 ) and deflection (E) 19 ).

[0056] In this embodiment, the status data can be represented by image data, meaning the status acquisition unit can include, for example, a drone. In essence, the evaluation parameters correspond to different parts of the equipment body; therefore, drone aerial photography can effectively collect status data for all evaluation parameters of all equipment bodies. Furthermore, since using drone photography for power line inspection is widely used in the power industry, the status data in step S11 can be obtained relatively easily or directly.

[0057] In particular, considering that step S13 is difficult to implement using methods such as image processing algorithms, and is usually performed manually, a major drawback of manual comparison is that the evaluations given are usually in the form of linguistic terms, but the meaning of linguistic terms is often vague and difficult to represent in numerical language that can be used by computer programs.

[0058] Therefore, in this embodiment, a unified language evaluation terminology can be constructed through the terminology evaluation set P in step S12, thus enabling all evaluation parameters of all device entities to be evaluated based on the same benchmark. In this embodiment, the terminology evaluation set P can have four elements with semantic degree increasing sequentially with the value of k, such as good (P1), attention (P2), abnormal (P3), and serious (P4); good (P1) means that the collected state data of the corresponding evaluation parameter is relatively good in terms of deviation from the normal state, and so on, serious (P4) means that the collected state data of the corresponding evaluation parameter is seriously deviating from the normal state.

[0059] It is understandable that the term evaluation set P can be stored externally after it is constructed, or it can be stored directly in the data evaluation unit.

[0060] In this embodiment, Can be represented as and The range of values ​​are respectively The upper and lower limits of the value; it is understandable that, K is the maximum value of k. In this embodiment, to The values ​​can be (0, 0.2), (0.2, 0.4), (0.5, 0.7), and (0.7, 1), respectively. It is understandable that the evaluation values... The higher the value, the more severe the deviation from the normal range, or the more abnormal it is.

[0061] It is understandable that, since the evaluation parameters are evaluated based on linguistic terms in step S13, the constructed terminology evaluation matrix... With sufficient fuzziness (i.e. uncertainty), the numerical matrix is ​​determined by randomly selecting values ​​within the interval in step S14. Therefore, it can better offset the uncertainty caused by linguistic terminology. In particular, as will be known below, outliers in this case... Based on statistical principles, this method of obtaining random values ​​can provide sufficient reliability even in situations with multiple data sources.

[0062] In step S15 of this embodiment, statistical principles can be used to achieve the analysis of each device body E. i abnormal factors The acquisition. Specifically, in dense channels, the same type of equipment body E i The quantity of each type of equipment is multiple, therefore, when conducting a certain type of equipment body E i When predicting the probability of anomalies, it is possible to collect data on all such devices E within the same time period. i Status data, by statistically analyzing the frequency of anomalies, can serve as a good basis for predicting the probability of anomalies.

[0063] In this embodiment, through Able to obtain the main body E of all participating equipment. i The sum of the evaluation values ​​of the j-th evaluation parameter, and this sum of evaluation values ​​is combined with the evaluation of the main equipment E participating in the evaluation. i The ratio of the total number N is the single anomaly factor used to evaluate the j-th evaluation parameter.

[0064] Considering the same equipment body E i Different evaluation parameters have different effects on the overall anomaly degree, so a weighting factor ω is introduced in this embodiment. ij This allows for a better consideration of the different impacts.

[0065] Furthermore, in this embodiment, a state history database can be established. The state data collected in step S11 can be stored in the state history database, and then steps S12-S15 can be based on relevant data sources in the state history database. Therefore, the main device E participating in the evaluation can be gradually enriched more effectively. i The total number of cases can thus improve the reliability of the prediction.

[0066] In this embodiment, the weighting factor ω ij The determination is made based on the following steps.

[0067] Step S15a: Obtain ω based on the Analytic Hierarchy Process (AHP). ijSubjective weight

[0068] Step S15b: Obtain ω based on the mean square error method. ij objective weight

[0069] Step S15c: Calculate the impact factor ω ij , α is a composite factor, α∈[0,1].

[0070] Based on the above, it is possible to better combine subjective and objective factors to achieve a better weighting of the factor ω. ij The determination.

[0071] Among them, the Analytic Hierarchy Process (APH) can construct a scale and form a judgment matrix through pairwise comparisons, and obtain the subjective weight of each subject through the judgment matrix. This method is a relatively mature existing technology, so it will not be described in detail in this embodiment.

[0072] In the mean squared error method, the evaluation values ​​of corresponding evaluation parameters for multiple lines are collected based on steps S11-S14. Then, the ratio of the variance of the evaluation value of the current line's corresponding evaluation parameter to the sum of the variances of the evaluation values ​​of all lines' corresponding evaluation parameters is used as the basic objective weight. Subsequently, the basic objective weights of different evaluation parameters are obtained sequentially, and the ratio of the basic objective weight of the corresponding evaluation parameter to the sum of all basic objective weights is used as the objective weight. That's it. The mean square error method is also a relatively mature existing method, so it will not be described in detail in this embodiment.

[0073] Where α = 0.5. Therefore, it can better balance subjective and objective weights.

[0074] In this embodiment, after completing step S15, it is also possible to base the device body E on... i abnormal factors Obtain the overall anomaly score P of the operating line, specifically as follows:

[0075] Step S16: Obtain the overall anomaly score P of the operating line.

[0076]

[0077] Among them, R i For the main body of the equipment E i The influencing factor that affects the status of the operating line, m is the main body of the equipment E i The total number of categories.

[0078] In this embodiment, the overall anomaly score P can be obtained through the status evaluation unit, and the overall anomaly score E of each category of equipment body can be obtained by comprehensively evaluating the anomaly score P. i The system identifies and weights abnormal situations, thus enabling better prediction of overall abnormal situations in dense channels (operating lines).

[0079] In the above specific embodiment, the influence factor R i The determination is made based on the following steps.

[0080] Step S16a: Based on the Analytic Hierarchy Process (AHP), obtain R. i Subjective weight

[0081] Step S16b: Obtain R based on the mean square error method. i objective weight

[0082] Step S16c: Calculate the impact factor R i , β is a composite factor, β∈[0,1].

[0083] As mentioned above, steps S16a-S16c can better combine subjective and objective factors to achieve the desired weighting of factor R. i The determination.

[0084] In the above specific embodiment, β = 0.5. Therefore, it can better balance subjective weights and objective weights.

[0085] Furthermore, after completing steps S15 and / or S16, this embodiment can also be based on anomaly factors. This, along with the overall anomaly score P, forms a corresponding early warning signal and is output. Specifically,

[0086] Step S17: Set the corresponding anomaly factor threshold at a state early warning unit. and the overall anomaly threshold P t The status early warning unit is in response to abnormal factors. Exceeding the corresponding anomaly factor threshold The overall anomaly score P exceeds the corresponding overall anomaly threshold P. t It generates and outputs warning signals in a timely manner.

[0087] Through the above methods, it is possible to provide corresponding early warnings when any abnormality occurs in any type of equipment body or dense channel (operating line), thus enabling better early warning of related abnormal situations.

[0088] Yes, it is understandable that the status warning unit can be implemented through a program, and the warning signal can include, but is not limited to, signals sent to a specific app or mobile terminal.

[0089] Furthermore, this embodiment also provides a multi-dimensional anomaly prediction model, which can better implement the method of this embodiment, and includes:

[0090] A status acquisition unit is used to implement step S11;

[0091] The data evaluation unit is used to store the term evaluation set P and the corresponding numerical evaluation set P from step S12. * And used to implement step S13;

[0092] A numerical unit, which is used to implement step S14; and,

[0093] The status evaluation unit is used to implement steps S15 and S16.

[0094] Through the above methods, it is possible to better collect and process the status data of each main equipment in the dense channel (operating line), thereby enabling the prediction of abnormal states of each type of main equipment and the entire dense channel (operating line).

[0095] The anomaly prediction model also includes a state history database, which stores the state data collected in step S11. This allows steps S12-S15 to be based on relevant data sources from the state history database. Therefore, it can better and progressively enrich the main equipment E participating in the evaluation. i The total number of cases can thus improve the reliability of the prediction.

[0096] The anomaly prediction model also includes a status early warning unit, which generates and outputs an early warning signal, thus implementing step S17. Therefore, it can effectively provide early warning when an anomaly occurs in the main equipment or the entire dense channel (operating line).

[0097] In addition, this embodiment also provides a multi-dimensional anomaly prediction device, which includes a memory and a processor. The anomaly prediction model has a computer program stored in the memory. When the processor executes the corresponding computer program, it implements the above-mentioned anomaly prediction method.

[0098] In addition, this embodiment also provides a computer-readable medium having a computer program stored thereon, which, when executed, implements the above-described anomaly prediction method.

[0099] It is understood that the anomaly prediction model in this embodiment can be implemented entirely through a computer program, which can provide corresponding input and output interfaces to realize data input at the status acquisition unit and output early warning signals.

[0100] Example 2

[0101] Combination Figure 2 As shown, this embodiment provides a method for predicting faults in dense passageways based on meteorological disasters. It is well applicable to predicting the probability of faults occurring in each type of equipment or the entire dense passageway due to meteorological disasters.

[0102] The dense channel fault prediction method in this embodiment includes the following steps:

[0103] Step S21: Collect historical meteorological characteristic sequences A of the same dense channel under historical meteorological disasters through the meteorological disaster collection unit. # and the failure rate sequence E of the main body of the equipment # ; For the uth type of meteorological disaster A u eigenvalues; For the i-th device body E i The failure rate;

[0104] Step S22: Establish a historical meteorological disaster sample set F and store it in the historical meteorological disaster database.

[0105]

[0106] Among them, F x The x-th sample in the historical meteorological disaster sample set F, For sample F x Failure rate sequence E # , For sample F x Historical meteorological characteristic sequence A # ;

[0107] Step S23: Obtain the meteorological characteristic sequence A for the period to be predicted through the meteorological condition acquisition unit. * , For the uth type of meteorological disaster A u eigenvalues;

[0108] Step S24: Based on the similarity algorithm, the meteorological feature sequence A is compared using the similarity comparison unit. * Each sample F in the historical meteorological disaster sample set F x Historical meteorological characteristic sequence By making comparisons one by one, historical meteorological characteristic sequences can be obtained. With meteorological characteristic sequence A * Most similar sample F x ;

[0109] Step S25: Output the most similar sample F through the failure rate output unit. x Failure rate sequence And serve as the predicted failure rate of the main equipment during the period to be predicted.

[0110] The above method enables the collection of a large number of historical meteorological characteristic sequences A under different meteorological disasters for the same dense channel. # and the failure rate sequence E of the main body of the equipment # This leads to the establishment of a large database. Based on this database, when predicting the probability of failure of various equipment due to meteorological disasters during the forecast period in dense channels, it is possible to obtain historical data most similar to the meteorological conditions of the forecast period based on similarity judgment, and use the failure rate of each equipment in the historical data as the predicted failure rate. Therefore, it is possible to better predict the impact of meteorological disasters on each equipment during the forecast period.

[0111] In this embodiment, meteorological disaster A u It can include one or more of the following: lightning, blizzard, strong winds, heavy rain, sandstorms, high temperatures, and cold waves. Understandably, when constructing meteorological feature sequences, it is possible to select corresponding types of meteorological disasters based on the geographical location of dense channels.

[0112] In this embodiment, when constructing historical meteorological feature sequence A # At that time, for meteorological disaster A u It can be used as meteorological disaster A u Actual strength S u and actual duration T u The product of these is used as its eigenvalue; that is,

[0113]

[0114] The above approach allows for a better consideration of the impact of the intensity and duration of meteorological disasters on the failure rate of the main equipment.

[0115] In this embodiment, when constructing meteorological feature sequence A * At that time, it is possible to use meteorological disaster A u probability of occurrence Predicted intensity and predicted duration The product of these is used as its eigenvalue; that is,

[0116]

[0117] The above approach allows for a better comprehensive consideration of the impact of the probability of meteorological disasters, the intensity of the forecast, and the duration of the forecast on the calculated forecast failure rate.

[0118] In this embodiment, for high temperature and cold wave, the intensity can be represented by temperature value (°C), blizzard can be represented by snowfall (mm), rainstorm can be represented by rainfall (mm), strong wind and sandstorm can be represented by wind speed (m / s), lightning can be represented by lightning current intensity (A), and the duration can be represented by minutes.

[0119] It is understandable that in constructing historical meteorological characteristic sequences A # At that time, meteorological disaster A u The intensity and duration are both known values; therefore, the corresponding characteristic values ​​can be obtained relatively well.

[0120] In addition, considering that the intensity of a meteorological disaster changes over time during its duration, the average intensity can be used as the calculated value.

[0121] Furthermore, the meteorological condition acquisition unit can be connected to existing weather forecasting systems, thus enabling it to better acquire meteorological conditions for the forecast period and construct a better meteorological characteristic sequence A. * .

[0122] In this embodiment, for any meteorological disaster A u It allows setting the probability of a blue alert to 0.55, a yellow alert to 0.7, an orange alert to 0.8, and a red alert to 0.9. Therefore, it can effectively obtain quantified probabilities of occurrence based on relevant data from the weather forecast system.

[0123] Understandably, yellow, orange, and red alerts are typically used for meteorological disasters such as lightning, sandstorms, and high temperatures; while blue, yellow, orange, and red alerts are typically used for meteorological disasters such as blizzards, strong winds, heavy rain, and cold waves. Therefore, when the meteorological condition acquisition unit obtains relevant color-coded warning information from the weather forecasting system, it can better perform numerical processing of the color-coded warning signal, thereby better obtaining the relevant prediction failure rate for the forecast period.

[0124] In this embodiment, meteorological feature sequence A can be obtained based on Euclidean distance. * With historical meteorological characteristic sequence The similarity value Q between them, and the similarity with the meteorological feature sequence A. * The sample F with the smallest similarity value Q x As the most similar sample; among them,

[0125]

[0126] Where U represents the total number of meteorological disasters included in the forecast.

[0127] Through the above methods, the meteorological characteristic sequence A for the period to be predicted can be obtained more effectively. * Similarity determination with all samples in the historical meteorological disaster sample set F.

[0128] This embodiment also includes the following steps:

[0129] Step S26: Obtain and output the overall meteorological failure rate L of the operating line through the meteorological early warning unit, wherein,

[0130]

[0131] Same as in Example 1, R i For the main body of the equipment E i The influencing factor that affects the status of the operating line, m is the main body of the equipment E i The total number of categories.

[0132] In this embodiment, the overall meteorological failure rate L can be obtained through the meteorological early warning unit, and the overall meteorological failure rate E of each type of equipment can be obtained through comprehensive analysis. i Failure rate By applying weights, it is possible to better predict the probability of failure of dense corridors (operating lines) due to meteorological disasters.

[0133] Combination Figure 3 As shown, to implement the method in this embodiment, this embodiment also provides a dense channel fault prediction model based on meteorological disasters, which includes:

[0134] A meteorological disaster data acquisition unit is used to implement step S21;

[0135] A historical meteorological disaster database is used to implement step S22;

[0136] Meteorological condition acquisition unit, which is used to implement step S23;

[0137] A similarity comparison unit is used to implement step S24;

[0138] Failure rate output unit, which is used to implement step S25; and,

[0139] The weather warning unit is used to implement step S26.

[0140] Through the above methods, it is possible to better predict the probability of failure of each main unit and the whole of equipment in dense channels (operating lines) due to meteorological disasters.

[0141] In addition, this embodiment also provides a dense channel fault prediction device based on meteorological disasters, which includes a memory and a processor. The dense channel fault prediction model has a computer program stored in the memory. When the processor executes the corresponding computer program, it implements the above-mentioned dense channel fault prediction method.

[0142] In addition, this embodiment also provides a computer-readable medium having a computer program stored thereon, which, when executed, implements the above-described dense channel fault prediction method.

[0143] It is understood that the dense channel fault prediction model in this embodiment can be implemented entirely through a computer program, which can provide corresponding input and output interfaces to enable the access and output of relevant data or signals.

[0144] Example 3

[0145] Combination Figure 4 As shown, this embodiment provides an intelligent early warning system for dense passages, which includes the multi-dimensional anomaly prediction model in embodiment 1 and the dense passage fault prediction model in embodiment 2, as well as an early warning module.

[0146] Among them, the anomaly prediction model is used to obtain the main body of the equipment E. i abnormal factors And the overall anomaly score P of the operating line, and used in the anomaly factor Exceeding the corresponding anomaly factor threshold The overall anomaly score P exceeds the corresponding overall anomaly threshold P. t It generates and outputs warning signals in a timely manner.

[0147] Among them, the failure rate prediction model is used to obtain the equipment body E i And the predicted failure rate of the main equipment components of the operating line that are affected by meteorological disasters during the predicted period. The overall meteorological failure rate L is output.

[0148] Among them, the early warning module is used based on the main body of the device E i abnormal factors and predicted failure rate Obtain the main body of the device E i Comprehensive outliers It is used to obtain the comprehensive anomaly value G of the operating line based on the overall anomaly score P and the overall meteorological failure rate L, and then outputs it.

[0149] Through the above, it is possible to better achieve control over the entire operating line and its associated equipment. iIt predicts the probability of anomalies during operation and failures due to meteorological disasters, and calculates the comprehensive failure rate caused by both factors. Therefore, it can effectively predict anomalies in densely packed lines under multiple operating conditions.

[0150] in, G = 1 - (1 - P) * (1 - L). Therefore, it can achieve better acquisition of relevant comprehensive outliers.

[0151] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for predicting dense passage failure based on weather disasters, comprising the following steps: Step S21, collecting a historical meteorological feature sequence A of the same dense channel under historical meteorological disasters by the meteorological disaster collection unit # and the failure rate sequence E of the device main body # is the characteristic value of the u-th meteorological disaster A u ; is the failure rate of the i-th device main body E i ​​ Step S22, establishing a historical weather disaster sample set F and storing it in a historical weather disaster database, wherein F x is the x-th sample in the historical weather disaster sample set F, is the failure rate sequence E x for the sample F # , is the historical weather feature sequence A x for the sample F # ; Step S23, acquiring a meteorological feature sequence A of a to-be-predicted period by a meteorological condition acquisition unit * , is a feature value of the u-th meteorological disaster A u . Step S24: Based on the similarity algorithm, the meteorological feature sequence A is compared using the similarity comparison unit. * Each sample F in the historical meteorological disaster sample set F x Historical meteorological characteristic sequence By making comparisons one by one, historical meteorological characteristic sequences can be obtained. With meteorological characteristic sequence A * Most similar sample F x ; Step S25, outputting the most similar sample F by the failure rate output unit x of the failure rate sequence and as the predicted failure rate of the device subject of the period to be predicted. 2.The weather disaster based dense tunnel fault prediction method of claim 1, wherein: When constructing the historical meteorological feature sequence A # , for the meteorological disaster A u , take the product of the actual intensity S u and the actual duration T u of the meteorological disaster A u as its characteristic value; That is, 3.The weather disaster based dense tunnel fault prediction method of claim 1, wherein: When constructing the weather feature sequence A * , the product of the occurrence probability u , the predicted intensity , and the predicted duration of the weather disaster A is taken as the characteristic value thereof; That is, 4.The weather disaster-based dense tunnel fault prediction method of claim 3, wherein: For any meteorological disaster A u , the occurrence probability value of the blue warning is set to 0.55, the occurrence probability value of the yellow warning is set to 0.7, the occurrence probability of the orange warning is 0.8, and the occurrence probability value of the red warning is 0.

9. 5.The weather disaster based dense tunnel fault prediction method of claim 1, wherein: Meteorological feature sequence A obtained based on Euclidean distance * With historical meteorological characteristic sequence The similarity value Q between them, and the similarity with the meteorological feature sequence A. * The sample F with the smallest similarity value Q x As the most similar sample; among them, wherein U is the total number of weather disasters involved in the prediction. 6.The weather disaster based dense tunnel fault prediction method of claim 1, wherein: Further comprising the following steps, Step S26, obtaining the overall weather failure rate L of the running line by a weather warning unit and outputting, wherein, wherein R i is the equipment body E i The impact factor affecting the running line state, m is the total number of categories of the equipment body E i .

7. A dense passage failure prediction model based on weather disasters, for realizing the method of any one of claims 1-6, comprising at least: a weather disaster collection unit for realizing step S21; a historical weather disaster database for realizing step S22; a weather condition collection unit for realizing step S23; a similarity comparison unit for realizing step S24; and, a failure rate output unit for realizing step S25. 8.The weather disaster based dense tunnel fault prediction model of claim 7, wherein: Further comprising a weather warning unit for obtaining and outputting the overall weather failure rate L of the line.

9. An intelligent warning system for dense passages, comprising an anomaly prediction model, a dense passage failure prediction model in claim 7 or 8, and a warning module.

Images