CN116455766A - Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition - Google Patents

Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition Download PDF

Info

Publication number
CN116455766A
CN116455766A CN202310691480.XA CN202310691480A CN116455766A CN 116455766 A CN116455766 A CN 116455766A CN 202310691480 A CN202310691480 A CN 202310691480A CN 116455766 A CN116455766 A CN 116455766A
Authority
CN
China
Prior art keywords
decomposition
carrying capacity
current carrying
prediction
overhead conductor
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202310691480.XA
Other languages
Chinese (zh)
Other versions
CN116455766B (en
Inventor
王孟夏
姚林渭
黄金鑫
杨明
王明强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202310691480.XA priority Critical patent/CN116455766B/en
Publication of CN116455766A publication Critical patent/CN116455766A/en
Application granted granted Critical
Publication of CN116455766B publication Critical patent/CN116455766B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0876—Network utilisation, e.g. volume of load or congestion level
    • H04L43/0882—Utilisation of link capacity
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Environmental & Geological Engineering (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a method and a system for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition, and belongs to the technical field of electrical engineering. The prediction method comprises the following steps: obtaining a current carrying capacity sequence of the overhead conductor according to the acquired historical meteorological data around the overhead conductor to be predicted; decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components by using a decomposition method after parameter optimization, and respectively carrying out probability prediction on each decomposition component to obtain a prediction result corresponding to each decomposition component; and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment. The invention effectively improves the prediction effect of the current carrying capacity.

Description

Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition
Technical Field
The invention relates to the technical field of electrical engineering, in particular to an overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The current carrying capacity of an overhead conductor is closely related to its operating environment and the surrounding microclimate environment and is limited by the maximum allowable temperature of the conductor, which is different under different meteorological conditions. In order to meet the requirement of long-term safe and stable operation of the power transmission line, electric workers take static heat constant values (Static Thermal Rating, STR) as current carrying limit of the overhead conductor, however, the STR is obtained by evaluation under severe weather conditions (high ambient temperature, low wind speed and strong sunlight intensity), has obvious conservation, limits the current carrying capacity of the power transmission line to a certain extent and does not meet the economic requirement of a smart grid. Aiming at the problems, a dynamic hot fixed value (Dynamic Thermal Rating, DTR) technology is applied to engineering, and the real-time current carrying capacity of the line is obtained by monitoring the microclimate environment and the running state around the power transmission line in real time and calculating according to the meteorological conditions monitored in real time. Engineering applications indicate that DTR of overhead conductors is in most cases higher than STR, and the use of DTR as a current carrying limit for overhead lines can better utilize the actual current carrying capacity of the line.
The current carrying capacity of the line can be fully exerted by constructing the current carrying constraint in the scheduling decision by using the DTR, the capacity of the power system for receiving new energy to generate electricity can be effectively enhanced, in order to integrate the time-varying DTR along with the change of meteorological conditions into the scheduling decision of the power system, operators are helped to predict and fully utilize the current carrying capacity of the overhead line, the current carrying capacity prediction problem of the overhead line is widely focused in recent years, and the current carrying capacity prediction method of the overhead line can be divided into: a point prediction method and a probability prediction method.
Aiming at the prediction research of the current carrying capacity point of the overhead conductor, a learner outputs a predicted value of the current carrying capacity through a trained weight network by constructing a relationship among a deep learning network fitting microclimate element, a line running state and a current carrying capacity sequence; in addition, a learner predicts the current carrying capacity of one day in the future by using a gray model based on period residual correction on the basis of smoothing the original sequence according to the characteristic that the current carrying capacity time sequence has multiple periods; the scholars also predict the current carrying capacity of 24 hours in the future by applying an Ornstein-Uhlenbeck method based on historical microclimate data and historical current carrying capacity data, wherein modeling of DTR fluctuation by Brownian motion is considered in the prediction of time series;
because the overhead conductor current carrying capacity has stronger fluctuation, the accurate prediction is difficult, the point prediction can not give out the current carrying capacity fluctuation interval and probability distribution information which are concerned by operators, and the operators often need to carry out conservation treatment according to experience in practical application.
Aiming at the prediction research of the current carrying capacity probability of the overhead conductor, a learner constructs a weather prediction model based on a particle swarm optimization kernel extreme learning machine, firstly predicts weather elements, and obtains a probability density prediction result of the current carrying capacity by combining a thermal balance model on the basis of known weather element distribution functions; meanwhile, a learner takes intermediate variables such as historical current carrying capacity, historical microclimate data, heat absorption capacity, heat dissipation capacity and the like of a conductor as prediction input, and a probability prediction model based on a deep neural network is constructed to predict the current carrying capacity interval distribution of one hour in the future; in addition, on the basis of carrying out time-interval probability prediction on the current carrying capacity, a learner considers the autocorrelation of the current carrying capacity time sequence according to the Copula theory, and carries out multi-time-interval joint probability prediction on the current carrying capacity;
the scheme for predicting the current carrying capacity probability of the overhead conductor does not effectively process the current carrying capacity sequence, is limited by the characteristics of strong nonlinearity and strong non-stationarity of the current carrying capacity of the overhead conductor, cannot effectively mine deep features in the current carrying capacity data, has larger limitation, and still has a space for improving the current carrying capacity predicting effect.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides the overhead conductor current-carrying capacity prediction method and the overhead conductor current-carrying capacity prediction system based on signal sequence decomposition, which are characterized in that after the current-carrying capacity sequence is subjected to self-adaptive decomposition, a prediction model is built for each decomposition component, and the final probability prediction result is obtained by overlapping the decomposition components, and the current-carrying capacity prediction effect is improved by reducing the strong nonlinearity and the strong non-stationarity of the current-carrying capacity sequence.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the invention provides a method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition.
An overhead conductor current carrying capacity prediction method based on signal sequence decomposition comprises the following steps:
obtaining a current carrying capacity sequence of the overhead conductor according to the obtained historical meteorological data around the overhead conductor to be predicted, wherein the historical meteorological data are meteorological data of a set time period before the current moment;
decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, and respectively carrying out probability prediction based on each decomposition component to obtain a prediction result corresponding to each decomposition component;
and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment.
As a further definition of the first aspect of the invention, decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components comprises:
and decomposing the overhead conductor current carrying capacity sequence based on the variation modal decomposition model optimized by the sparrow search algorithm to obtain a plurality of decomposition components.
As a further limitation of the first aspect of the present invention, a variation mode decomposition model is adopted to decompose the overhead conductor current carrying capacity sequence to obtain a plurality of decomposition components;
taking the ratio of the average value of the absolute value of each decomposition component amplitude to the average amplitude of the original component as a weight value, and carrying out weighted summation on the arrangement entropy values of each decomposition component;
and optimizing the variation modal decomposition model based on a sparrow search algorithm by taking the minimum permutation entropy value of weighted summation as an optimization objective function to obtain a parameter optimizing result of the variation modal decomposition model.
As a further limitation of the first aspect of the present invention, the parameter optimizing result of the variation modal decomposition model includes: the penalty factor and the number of decompositions after optimization.
As a further limitation of the first aspect of the present invention, predicting based on each decomposition component and performing overlap reconstruction on each prediction result, respectively, includes:
and adopting a quantile regression forest model to respectively carry out probability prediction on each decomposition component, and carrying out superposition reconstruction on each decomposition component prediction result to obtain a final probability prediction result.
As a further definition of the first aspect of the invention, the meteorological data comprises: solar intensity, air temperature, wind speed and wind direction.
The second aspect of the invention provides an overhead conductor current carrying capacity prediction system based on signal sequence decomposition.
An overhead conductor current carrying capacity prediction system based on signal sequence decomposition, comprising:
a data acquisition module configured to: obtaining a current carrying capacity sequence of the overhead conductor according to the obtained historical meteorological data around the overhead conductor to be predicted, wherein the historical meteorological data are meteorological data of a set time period before the current moment;
a decomposition prediction module configured to: decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, and respectively carrying out probability prediction on each decomposition component to obtain a prediction result corresponding to each decomposition component;
a reconstruction prediction module configured to: and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment.
As a further definition of the second aspect of the invention, the decomposition prediction module decomposes the overhead conductor current carrying capacity sequence into a plurality of decomposition components, comprising:
and decomposing the overhead conductor current carrying capacity sequence based on the variation modal decomposition model optimized by the sparrow search algorithm to obtain a plurality of decomposition components.
As a further limitation of the second aspect of the present invention, a variation mode decomposition model is adopted to decompose the overhead conductor current carrying capacity sequence to obtain a plurality of decomposition components;
taking the ratio of the average value of the absolute value of each decomposition component amplitude to the average amplitude of the original component as a weight value, and carrying out weighted summation on the arrangement entropy values of each decomposition component;
and optimizing the variation modal decomposition model based on a sparrow search algorithm by taking the minimum permutation entropy value of weighted summation as an optimization objective function to obtain a parameter optimizing result of the variation modal decomposition model.
As a further limitation of the second aspect of the present invention, predicting based on each decomposition component and performing overlap reconstruction on each prediction result, respectively, includes:
and adopting a quantile regression forest model to respectively carry out probability prediction on each decomposition component, and carrying out superposition reconstruction on each decomposition component prediction result to obtain a final probability prediction result.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention creatively provides an overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition, which are used for predicting after decomposing current carrying capacity data, so that the non-stationarity and nonlinearity of the current carrying capacity sequence are effectively reduced, and the current carrying capacity prediction effect is improved.
2. The invention creatively provides a method and a system for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition, which adopt a sparrow search algorithm to optimize a variation modal decomposition model of the current carrying capacity, and realize more accurate and effective judgment on the regularity of the decomposition amount of the current carrying capacity while improving the self-adaptability of the variation modal decomposition model.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
Fig. 1 is a flow chart of a method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to embodiment 1 of the present invention;
fig. 2 is a flowchart of optimizing variation modal decomposition model parameters by using sparrow search algorithm provided in embodiment 1 of the present invention;
FIG. 3 is a result of an iterative process for optimizing a variational modal decomposition model by using the sparrow search algorithm provided in embodiment 1 of the present invention;
FIG. 4 shows the mode frequencies obtained by decomposition of the variation modes according to example 1 of the present invention;
FIG. 5 shows the current carrying capacity of example 1 of the present invention decomposed by a variation mode;
FIG. 6 is a schematic diagram showing the comparison of the prediction results of the present invention provided in example 1 with the prediction results obtained by other prediction models;
fig. 7 is a schematic diagram of an overhead conductor current carrying capacity prediction system based on signal sequence decomposition according to embodiment 2 of the present invention;
wherein VMD is a variation modal decomposition, SSA is a sparrow search algorithm, IMF 1. IMFk is each component after VMD decomposition, QRF 1. IMFk is a quantile regression forest model corresponding to IMF 1. IMFk,Kthe number of the VMD key parameter group decomposition is calculated, and alpha is a penalty factor.
Detailed Description
The invention will be further described with reference to the drawings and examples.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1:
as shown in fig. 1, embodiment 1 of the present invention provides a method for predicting current carrying capacity of an overhead conductor based on signal sequence decomposition, which includes the following steps:
s101: acquiring DTR historical measurement data; wherein the DTR historical measurement data comprises historical weather data around the overhead conductor.
In this embodiment, the historical meteorological data around the overhead conductor includes historical solar intensity, air temperature, wind speed and wind direction around the conductor.
It should be noted here that in other embodiments, the historical weather data around the wire may also include other weather element data.
S102: based on the DTR historical measurement data, calculating the current carrying capacity of the overhead conductor according to formula (1)I max Comprising:
（1）
wherein, the liquid crystal display device comprises a liquid crystal display device,T max for the maximum allowable operating temperature of the conductor, the value of this example is 70 c,R（T max ) At a conductor temperature ofT max The resistance (omega/m) of the conductor per unit length,q s solar heat absorption capacity (W/m) of a conductor per unit length,q c （T max ) Convective heat dissipation (W/m) for a unit length of conductor;q r （T max ) The specific calculation formula of the parameters is as follows:
（2）
（3）
（4）
（5）
wherein, the liquid crystal display device comprises a liquid crystal display device,R（T avg ) Representative conductor at its temperature ofT avg Resistance per unit length (Ω/m), equation (2) describes the linear relationship between conductor temperature and its resistanceT low ，T high ]An effective temperature interval representing a linear relationship between conductor temperature and its resistance, which, in the present invention, is optional,T low andT high are set to 25 ℃ and 75 ℃ respectively, and are found at the maximum allowable temperatureT max In the case of conductor resistances operating downwards, takeT avg =T max The method comprises the steps of carrying out a first treatment on the surface of the In the formula (3), the amino acid sequence of the compound,α s representing the light absorptivity of the conductor,Q se represents the amount of solar radiation per unit area (W/m) after the altitude correction 2 ），θRepresenting the angle of incidence of the sunlight,A’represents the projected area (m 2 M); in the formula (4), the amino acid sequence of the compound,Drepresenting the outer diameter (m) of the conductor,ρ f represents the air density (kg/m) around the conductor 3 ），μ f Represents the dynamic viscosity (kg/m-s) of the surrounding air,V w representing the wind speed (m/s) around the conductor,k f is the thermal conductivity (W/(m-DEGC)) of the surrounding air,k angle representing a wind direction factor related to the angle between the conductor axis and the wind direction,T s is the conductor surface temperature (°c),T a representing the ambient temperature (C) around the conductor,q c1 andq c2 can be used to calculate the heat dissipation from forced convection at low and high wind speeds, respectively, and should be used at a given wind speedq c1 Andq c2 the larger of the results of the calculation therebetween serves as the heat dissipation amount generated by the forced convection,q c3 for calculating the heat dissipation capacity generated by natural convection at zero wind speed, it is generally required to calculate separately according to IEEE standard in practical applicationq c1， q c2 Andq c3， the larger of the calculation results should be selected asq c The method comprises the steps of carrying out a first treatment on the surface of the In the formula (5), the amino acid sequence of the compound,εrepresenting the emissivity of the conductor.
S103: penalty factor α and number of decompositions for current carrying capacity sequence variant modal decomposition (Variational Mode Decomposition, VMD) using a sparrow search algorithm (Sparrow Search Algorithm, SSA)KKey parameters of the VMD (virtual machine direction) are automatically optimized, and the parameter setting of the VMD method is realizedAnd (3) performing self-adaptive optimization, and decomposing the current carrying capacity sequence into a group of relatively stable decomposition components by using the VMD after parameter optimization.
When parameter optimization is carried out, the selection of the fitness function in the optimization algorithm directly influences the algorithm convergence speed and the solution of the optimal result, and the fitness function is constructed by adopting an permutation entropy algorithm which can measure the complexity and the change regularity of the time sequence.
S1031: and decomposing the variation mode.
The variational modal decomposition algorithm is an adaptive and completely non-recursive signal processing method for modal decomposition of non-stationary time series. The adaptivity is characterized in that according to the modal decomposition number of a given sequence, the optimal center frequency and the limited bandwidth of each modal can be selected in a self-adaptive mode in the subsequent searching and solving process, and the effective separation of the inherent modal functions (intrinsic mode function, IMF) and the frequency domain division of signals can be realized, so that the effective decomposition components of the given signals are obtained, and the optimal solution of the variation problem is finally obtained. VMD provides an effective analysis and processing means for non-stationary time series with high complexity and strong nonlinearity based on solid mathematical theory, and can decompose given time series to obtain multiple relatively stationary decomposition components with different frequency ranges, and the core idea of VMD is to construct and solve variational problem.
For each current carrying capacity mode signalu k （t）The analytic spectrum of each mode is obtained through Hilbert transformation, the spectrum of each mode is modulated to a corresponding baseband, and the square L of the demodulation signal gradient is used for 2 Norm regularization Gaussian smoothing estimation is carried out to estimate each mode signalu k （t）With the minimum sum of modal bandwidths as a target and with the original current carrying capacity equal to the sum of modal bandwidths as a constraint condition, an optimization model of a modal expression and the center frequency thereof is established, and the constrained variation problem is as follows:
（6）
（7）
wherein:u k represent the firstkThe mode of the seed crystal is changed,w k for the set of center frequencies of all modes,Kfor the number of all the modes,f（t）is the original current carrying capacity sequence signal,is a unit pulse function>A convolution operation is represented and is performed,tis thattTime of day (I)>For time oftAnd (5) obtaining deviation guide.
To convert the constraint variable problem into an unconstrained variable problem to solve, a quadratic penalty factor alpha and a Lagrange multiplier are introduced into the variable problemThe extended Lagrangian expression is as follows:
（8）
by iterative updating using a multiplier-alternate direction method (Alternating Direction Method of Multipliers, ADMM)u k+1 ，w k+1 Andthe "saddle point" of the extended lagrangian expression is sought, and its update method is as follows:
（9）
（10）
（11）
in the method, in the process of the invention,as a result of the center frequency,w k is the center frequency of the kth mode,nfor the number of iterations, α is a quadratic penalty factor,>for the current residual quantity->Is Fourier transformation of a mode signal at the t moment of the n+1st round iteration under the kth mode; />The center of gravity of the power spectrum of the kth mode function under the n+1th round of iteration times;、/>、/>respectively->、/>、/>Is used for the fourier transform of (a),tis thattTime of day (I)>For the i-th modality +.>Fourier transform of->Lagrangian multiplier for time t, < ->For the nth round iteration +.>Fourier transform of->Is expressed in non-negative frequency +>Interval, update λ.
S1032: sparrow search algorithm.
In the sparrow searching algorithm, the sparrow population is divided into two types, namely a finder and a jointer, wherein the finder has stronger searching capability and is good at searching food, the identities of the finder and the jointer are dynamically changed, and the sparrow population can be continuously adjusted in the population, but the proportion is kept constant, so that the stability of the population structure is ensured; and meanwhile, a certain proportion of individuals are set as the scout for detecting predators so as to ensure the population safety. And calculating and updating the position of the whole group by constructing a corresponding mathematical model according to the initial position of each sparrow and the fitness determined by the fitness function, and simulating the foraging and predation resisting behaviors of the sparrow group.
1) The location update of the discoverer in the SSA algorithm during the iteration process is as follows:
（12）
wherein:vrepresenting the current iteration number;jrepresenting dimensions;r max representing a maximum number of iterations;X i,j representing the current iteration numbervLower firstiSparrow of the first kindjPosition information in the dimension; alpha is (0, 1)]Is a random number of (a);Qrandom numbers which are subject to normal distribution;Lrepresenting a 1 xd full-one matrix; s is S T ∈[0.5,1]Representing a security value;R 2 ∈[0,1]the early warning value is represented by the formula,R 2 <S T indicating that the population is in a safe area, discoverers can randomly forge food,R 2 ≥S T indicating that predators are found around the population, the position needs to be immediately adjusted to a safe area for foraging.
2) The enrollees attempt to compete for their discovered food in monitoring the discoverer's process, where the location of the enrollee is updated as follows:
（13）
wherein:X P representing the position of the sparrow individual with the highest fitness value in the current population;X worst representing the position of the sparrow individual with the lowest adaptability in the current population; a represents a 1×d matrix in which each element is randomly assigned 1 or-1, and A + =A T （AA T ） -1 ；i>n/2 represents the first low fitness valueiIndividual participants do not get food and need to go to other places to find food.
3) The number of sparrows with awareness of danger is usually 10% -20% of the total number in the whole sparrow population, and their initial positions are randomly distributed, and the mathematical model can be expressed as follows:
（14）
wherein, thereinIs the current iteration numbervIs set to be a global optimum position of (c),βthe step control parameter is a random number subjected to normal distribution with a mean value of 0 and a variance of 1,G∈[-1,1]is a random number, and the code is a random number,f i the fitness value of the current sparrow individual,f g andf w respectively areThe current global best and worst fitness values,εis the smallest constant to avoid zero occurrence in the denominator.
S1033: and (5) arranging entropy.
Is decomposed by VMD to obtainKIndividual components {u（1）,u（2）,…,u（K) The IMF component time series may be represented as {s（q）,q=1,2,···,NAnd (3) carrying out phase space reconstruction on the vector to obtain a reconstruction vector:
（15）
wherein: representation ofEmbedding dimension->For delay time, subscript->Reconstructing a vectorX j Can be used as matrixXIs to matrixXEach row of (a) is a reconstruction vectorX j The ascending arrangement is carried out again, and the method comprises the following steps:
（16）
for matrixXA set of symbol sequences is available for each row:
（17）
in the method, in the process of the invention,S（g) For a symbol matrix containing only location information,g=1,2,…,l. Calculating probability of occurrence of each symbol sequenceP 1 ,P 2 ,…,P l Calculating a time sequence { through shannon entropys（q）,q=1, 2, ··, the permutation entropy of N }, is:
（18）
entropy of arrangementAnd (3) carrying out normalization processing, namely:
（19）
wherein, the liquid crystal display device comprises a liquid crystal display device,H p the magnitude of the value reflects the degree of randomness of the time series,H p the larger the description time series is, the more random and vice versa, and in the subsequent calculation analysis of the invention, the embedding dimensionAnd delay time->Calculated by using a pseudo-neighbor method and an autocorrelation method respectively. The arrangement entropy value of each decomposition component is calculated according to the formulas (18) and (19), and is expressed as { respectivelyP E1 ,P E2 ,···,P EK }。
The signal decomposition-based method decomposes an original signal into a smooth-varying component and an unstable component based on characteristics of the signal. The smooth component is usually a low-frequency high-amplitude component, the fluctuation of the smooth component is relatively smooth and has obvious regularity, and therefore, a better prediction effect can be obtained by modeling and predicting the smooth component. The unstable component is often a high-frequency low-amplitude component, the fluctuation of the unstable component is larger and lacks obvious rules, and modeling prediction on the unstable component can generate a certain prediction error, but the influence on the final prediction result is relatively smaller because the amplitude proportion of the unstable component in the original signal is smaller. Based on the above, the invention takes the ratio of the average value of the absolute value of the amplitude of each decomposition component to the average amplitude of the original component as the weight valueP E The values are weighted summed.
Set the original signalf（t）Average amplitude of (a)An average value of absolute values of magnitudes of the decomposition components is { A }, A 1 ,A 2 ,…,A k Let w 1 =A 1 /A,w 2 =A 2 /A,…,w k =A k and/A, the fitness function is expressed as:
（20）
in the invention, byP EW The minimum value is the optimized objective function, and SSA is adopted to decompose the number of VMD key parameter sets and penalty factorK，α]And optimizing.
In this embodiment, historical microclimate measurement data and corresponding current carrying capacity data of a 220kV overhead conductor (the conductor model is LGJ 400/35) are collected, the time range is 20 years, 1 month and 3 months, the time resolution is 15 minutes, decomposition of the optimized VMD is carried out on the current carrying capacity time sequence based on SSA, a flow chart of the optimized VMD is shown in FIG. 2, and the key parameter group decomposition number and penalty factor [ of the VMD method ]K，α]The parameter optimizing boundary setting ranges are ([ 3, 15)],[100,2000]) The optimizing iterative process and result of SSA optimized VMD are shown in FIG. 3, which obtains the minimum fitness value of 0.0306 in the 6 th iteration, and the optimizing results are [10, 1850 respectively]The frequency results of each mode obtained by VMD decomposition are shown in fig. 4, and as can be seen from the graph, the frequency distinguishing characteristics of different IMFs are obvious, so that effective decomposition is realized; fig. 5 is a graph showing the results of IMF processing by VMD decomposition of a partial current carrying capacity sequence.
The SSA is used for automatically optimizing key parameters of VMD decomposition, so that the self-adaptive decomposition of the VMD is realized; meanwhile, the fitness function of the SSA-VMD is constructed based on the weighted permutation entropy, so that the current carrying capacity data can be effectively decomposed, the change rule of each decomposition component is fully excavated, the influence of different amplitude values of each decomposition component on a final prediction result is considered, and a solid data foundation is laid for maximally improving the current carrying capacity prediction effect.
S104: and establishing a corresponding QRF prediction model for each decomposition component, and overlapping and reconstructing the predicted values of different decomposition components to obtain a final current carrying capacity probability predicted result.
In order to evaluate the prediction effect of the provided prediction model, three prediction models, namely a quantile regression model (Quantile Regression, QR), a quantile regression forest model (Quantile Regression Forest, QRF) and the provided SSA-VMD-QRF prediction model, are selected to implement probability prediction on the current carrying capacity of the line, and the performance of the prediction model is checked through comparing and analyzing the prediction results.
S1041: quantile regression model.
The quantile regression algorithm is expanded by a least square algorithm and is a classical nonparametric probability prediction method. Overhead line current carrying capacityτQuantile prediction resulty（τ) Can be expressed as:
（21）
wherein, the liquid crystal display device comprises a liquid crystal display device,x i （i=1,2，…，h) In order to input the variable(s),β（τ）=[β 0 （τ）,β 1 （τ）,…,β h （τ）]is thatτParameter vectors of quantile predictive models.
Parameter vectorβ（τ) The expression is as follows:
（22）
wherein, the liquid crystal display device comprises a liquid crystal display device,y i as the current carrying capacity sample data,mfor the number of samples to be taken,η τ (. Cndot.) is a test function, whose expression is:
（23）
in the estimationβ（τ) Then, the probability prediction of the current carrying capacity can be realized according to the formula (21).
S1042: quantiles regress to the forest model.
The quantile regression forest model is an improved model that combines the ideas of quantile regression and random forests. Compared with the traditional quantile regression model, the quantile regression forest model can consider the nonlinear relation between the input variable and the output variable, so that more data relevance is mined. The model is widely applied to the fields of power prediction such as new energy power generation, load and the like.
The quantile regression forest model is an adaptive neighbor classification and regression process, for eachThe original can be obtainedMA set of weights for the observations +.>. QRF takes the weighted sum of all dependent variable observations as the dependent variable Y conditional mean +.>The QRF decision tree is generated with a standard random forest algorithm and the conditional distribution is obtained by weighting the estimates of the observed dependent variables, wherein the weight of each observation is equal to the weight of the random forest algorithm.
For QRF, defineIs the observation +.>Is a weighted average of (1), namely:
（24）
the specific steps of the QRF algorithm are summarized as follows:
1) Generating a decision tree: first, the generatedcDecision treeAs a basic component of the modelθParameters for a single decision tree, including the number of leaf node samples and the depth of the tree), for eachThe decision tree examines all the observations on each leaf node on the training dataset, unlike the traditional averaging method, where not only the average on the leaf nodes is considered, but the information of all the observations is considered;
2) Given a givenXTraversing all decision trees and calculating the observation weight of each decision tree for each observation valueThese weights measure the importance of each decision tree to each observation, and then average these weights to obtain the weight of each observationw i （X）；
3) For all ofUsing the weights obtained in step 2), an estimate of the distribution function is calculated by equation (24).
Through the steps, the quantile regression forest model can utilize a plurality of decision trees to predict, and the weight of each observed value is considered, so that the conditional distribution of different quantiles of the target variable can be estimated more accurately. Here, the key parameters of the SSA optimized QRF prediction model are applied at the same time, and the optimization flow steps are similar to those of the SSA optimized VMD, and are not repeated here.
In order to develop comparison analysis on the 3 prediction models, the above-mentioned overhead conductor related data is utilized, the first 90% of the data is selected as a training set to train the prediction models, and the last 10% of the data is selected as a testing set to test the validity of the prediction method. Based on the data analysis result and the prediction method mechanism, the SSA-VMD-QRF model adopts corresponding decomposition component data with a history of 2 hours (8 periods) as a prediction input for the prediction of each decomposition component, the prediction result of each decomposition component is superimposed and reconstructed to obtain a probability prediction result of the current carrying capacity of the line (the prediction is carried out forward by taking 1 hour as a prediction time domain rolling), in order to evaluate the prediction performance of the proposed prediction model, the QR model and the QRF model carry out probability prediction on the current carrying capacity of the line by taking the current carrying capacity data with a history of 2 hours (8 periods) as the prediction input, the part of the prediction results obtained by the test set are shown in fig. 6, which shows the probability prediction result of the current carrying capacity of the line for 72 hours (288 prediction periods), and compared with the QR model (the first method) and the QRF model (the second method), the probability prediction interval of the current carrying capacity of the line predicted by the proposed SSA-VMD-QRF model (the third method) is narrowest.
The mean width of the prediction interval (Average Width of the Prediction Intervals, AWPI), continuous ranking probability score (Continuous Ranked Probability Score, CRPS), mean absolute percentage error of 0.5 score relative to true value (Mean Absolute Percentage Error, MAPE), and root mean square error of 0.5 score relative to true value (Root Mean Square Error, RMSE) for the three prediction results in the test set are given in table 1, and the calculation formulas for AWPI, CRPS, MAPE and RMSE are as follows:
（25）
（26）
（27）
（28）
（29）
wherein 0.99 quatile represents a current carrying capacity prediction result of 0.99 quantiles, 0.01 quatile represents a current carrying capacity prediction result of 0.01 quantiles,zfor the number of samples to be taken,f i (y) a probability density function of the current carrying capacity predicted in the ith period, x is a current carrying capacity prediction variable,H(. Cndot.) represents a Heaviside function, the calculation formula of which is shown in formula (29),is the current carrying capacity true value, < >>(0.5) is a current carrying capacity 0.5 quantile predictive value, and as can be seen from table 1, both AWPI, CRPS, MAPE and RMSE of the current carrying capacity predictive result predicted by the proposed predictive model are minimum.
Table 1: AWPI, CRPS, MAPE and RMSE of three prediction methods
From the analysis, the overall prediction performance of the QRF model is superior to that of the QR model, which proves that the QRF model has better prediction performance and is more suitable for the probability prediction of current carrying capacity. The SSA-VMD-QRF prediction model further improves the prediction effect, which shows that the data preprocessing loop introduced into the front end can effectively improve the current carrying capacity prediction effect, decompose the current carrying capacity sequence into a group of stable and highly periodic decomposition components, and effectively solve the influence of the nonlinear and non-stable characteristics of the original sequence on the prediction effect.
The prediction model adopts the variation modal decomposition method after parameter optimization to reduce the complexity of input data, then carries out probability prediction on each decomposition component, and overlaps and reconstructs the prediction result of each decomposition component, thereby realizing the probability prediction of the current carrying capacity of the line, grasping the current carrying capacity of the overhead conductor at the future moment, providing reference information for the assignment of scheduling decisions of operators and being beneficial to realizing the dynamic capacity increase of the overhead line. In addition, the prediction performance of the provided prediction model is evaluated through comparison of prediction results obtained by multiple prediction models.
Example 2:
as shown in fig. 7, embodiment 2 of the present invention provides an overhead conductor current carrying capacity prediction system based on signal sequence decomposition, including:
a data acquisition module configured to: obtaining a current carrying capacity sequence of the overhead conductor according to the obtained historical meteorological data around the overhead conductor to be predicted, wherein the historical meteorological data are meteorological data of a set time period before the current moment;
a decomposition prediction module configured to: decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, and respectively carrying out probability prediction based on each decomposition component to obtain a prediction result corresponding to each decomposition component;
a reconstruction prediction module configured to: and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment.
The working method of the system is the same as the overhead conductor current carrying capacity prediction method based on signal sequence decomposition provided in embodiment 1, and will not be described here again.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. The overhead conductor current carrying capacity prediction method based on signal sequence decomposition is characterized by comprising the following steps of:
obtaining a current carrying capacity sequence of the overhead conductor according to the obtained historical meteorological data around the overhead conductor to be predicted, wherein the historical meteorological data are meteorological data of a set time period before the current moment;
decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, and respectively carrying out probability prediction on each decomposition component to obtain a prediction result corresponding to each decomposition component;
and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment.
2. The method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to claim 1, wherein,
decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, comprising:
and decomposing the overhead conductor current carrying capacity sequence based on the variation modal decomposition model optimized by the sparrow search algorithm to obtain a plurality of decomposition components.
3. The method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to claim 2, wherein,
decomposing the overhead conductor current carrying capacity sequence by adopting a variable-mode decomposition model to obtain a plurality of decomposition components;
taking the ratio of the average value of the absolute value of each decomposition component amplitude to the average amplitude of the original component as a weight value, and carrying out weighted summation on the arrangement entropy values of each decomposition component;
and optimizing the variation modal decomposition model based on a sparrow search algorithm by taking the minimum permutation entropy value of weighted summation as an optimization objective function to obtain a parameter optimizing result of the variation modal decomposition model.
4. A method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to claim 3,
the parameter optimizing result of the variation modal decomposition model comprises the following steps: the penalty factor and the number of decompositions after optimization.
5. The method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to claim 1, wherein,
probability prediction is performed based on each decomposition component, and overlapping reconstruction is performed on each prediction result, including:
adopts quantile regression forest model to respectively make probability prediction for each decomposition component,
and carrying out superposition reconstruction on the prediction results of the decomposition components to obtain a final probability prediction result.
6. A method for predicting the current carrying capacity of an overhead conductor based on signal sequence decomposition according to any one of claims 1 to 5,
meteorological data, comprising: solar intensity, air temperature, wind speed and wind direction.
7. An overhead conductor current carrying capacity prediction system based on signal sequence decomposition, comprising:
a data acquisition module configured to: obtaining a current carrying capacity sequence of the overhead conductor according to the obtained historical meteorological data around the overhead conductor to be predicted, wherein the historical meteorological data are meteorological data of a set time period before the current moment;
a decomposition prediction module configured to: decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, and respectively carrying out probability prediction on each decomposition component to obtain a prediction result corresponding to each decomposition component;
a reconstruction prediction module configured to: and carrying out superposition reconstruction on the prediction results corresponding to the decomposition components to obtain the current carrying capacity probability prediction result in a set time period after the current moment.
8. The overhead conductor current carrying capacity prediction system based on signal sequence decomposition according to claim 7, wherein,
in the decomposition prediction module, decomposing the overhead conductor current carrying capacity sequence into a plurality of decomposition components, comprising:
and decomposing the overhead conductor current carrying capacity sequence based on the variation modal decomposition model optimized by the sparrow search algorithm to obtain a plurality of decomposition components.
9. The overhead conductor current carrying capacity prediction system based on signal sequence decomposition according to claim 8, wherein,
decomposing the overhead conductor current carrying capacity sequence by adopting a variable-mode decomposition model to obtain a plurality of decomposition components;
taking the ratio of the average value of the absolute value of each decomposition component amplitude to the average amplitude of the original component as a weight value, and carrying out weighted summation on the arrangement entropy values of each decomposition component;
and optimizing the variation modal decomposition model based on a sparrow search algorithm by taking the minimum permutation entropy value of weighted summation as an optimization objective function to obtain a parameter optimizing result of the variation modal decomposition model.
10. The overhead conductor current carrying capacity prediction system based on signal sequence decomposition according to claim 7, wherein,
probability prediction is performed based on each decomposition component, and overlapping reconstruction is performed on each prediction result, including:
adopts quantile regression forest model to respectively make probability prediction for each decomposition component,
and carrying out superposition reconstruction on the prediction results of the decomposition components to obtain a final probability prediction result.
CN202310691480.XA 2023-06-13 2023-06-13 Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition Active CN116455766B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310691480.XA CN116455766B (en) 2023-06-13 2023-06-13 Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310691480.XA CN116455766B (en) 2023-06-13 2023-06-13 Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition

Publications (2)

Publication Number Publication Date
CN116455766A true CN116455766A (en) 2023-07-18
CN116455766B CN116455766B (en) 2023-09-08

Family

ID=87133978

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310691480.XA Active CN116455766B (en) 2023-06-13 2023-06-13 Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition

Country Status (1)

Country Link
CN (1) CN116455766B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102810859A (en) * 2012-07-11 2012-12-05 广东电网公司电力科学研究院 Weather forecast-based maximal current carrying capacity prediction method of overhead transmission line
WO2019041857A1 (en) * 2017-09-04 2019-03-07 东南大学 Scene analysis based operating state prediction method of power distribution network including distributed power
CN110321601A (en) * 2019-06-14 2019-10-11 山东大学 A kind of overhead transmission line dynamic current-carrying capability advanced prediction method and system
CN115310354A (en) * 2022-08-01 2022-11-08 山东大学 Method and system for predicting transient current-carrying capacity probability of overhead conductor

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102810859A (en) * 2012-07-11 2012-12-05 广东电网公司电力科学研究院 Weather forecast-based maximal current carrying capacity prediction method of overhead transmission line
WO2019041857A1 (en) * 2017-09-04 2019-03-07 东南大学 Scene analysis based operating state prediction method of power distribution network including distributed power
CN110321601A (en) * 2019-06-14 2019-10-11 山东大学 A kind of overhead transmission line dynamic current-carrying capability advanced prediction method and system
CN115310354A (en) * 2022-08-01 2022-11-08 山东大学 Method and system for predicting transient current-carrying capacity probability of overhead conductor

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
MA\'D EL-DALAHMEH等: ""Lithium-ion Batteries Capacity Degradation Trajectory Prediction Based on Decomposition Techniques and NARXAlgorithm"", 《2022 57TH INTERNATIONAL UNIVERSITIES POWER ENGINEERING CONFERENCE(UPEC)》 *
张斌;林章岁;叶荣;胡臻达;金涛;温步瀛;: "随环境条件变化的输电线路输送容量概率建模研究", 电工电能新技术, no. 12 *

Also Published As

Publication number Publication date
CN116455766B (en) 2023-09-08

Similar Documents

Publication Publication Date Title
Liang et al. A novel wind speed prediction strategy based on Bi-LSTM, MOOFADA and transfer learning for centralized control centers
Zhao et al. Research and application based on the swarm intelligence algorithm and artificial intelligence for wind farm decision system
Jiang et al. A combined forecasting system based on statistical method, artificial neural networks, and deep learning methods for short-term wind speed forecasting
Catalao et al. Hybrid wavelet-PSO-ANFIS approach for short-term wind power forecasting in Portugal
Niu et al. Uncertainty modeling for chaotic time series based on optimal multi-input multi-output architecture: Application to offshore wind speed
Wang et al. Wind speed forecasting based on multi-objective grey wolf optimisation algorithm, weighted information criterion, and wind energy conversion system: A case study in Eastern China
Jiang et al. Research and application of an innovative combined model based on a modified optimization algorithm for wind speed forecasting
Shams et al. Artificial intelligence-based prediction and analysis of the oversupply of wind and solar energy in power systems
CN110889603A (en) Power system economic dispatching method considering wind power correlation based on PCA-Copula theory
Zhang et al. Short-term power load forecasting using integrated methods based on long short-term memory
CN112115648A (en) Transformer top layer oil temperature prediction method based on improved deep learning method
CN109242136A (en) A kind of micro-capacitance sensor wind power Chaos-Genetic-BP neural network prediction technique
Samet et al. Evaluation of neural network-based methodologies for wind speed forecasting
Dong et al. Ensemble wind speed forecasting system based on optimal model adaptive selection strategy: Case study in China
Pinson Introducing distributed learning approaches in wind power forecasting
Wang et al. Hybrid deep learning model for short-term wind speed forecasting based on time series decomposition and gated recurrent unit
Wang et al. A novel wind power prediction model improved with feature enhancement and autoregressive error compensation
Zhuang et al. A multi-scale spatial-temporal graph neural network-based method of multienergy load forecasting in integrated energy system
Al-Ja’afreh et al. An enhanced CNN-LSTM based multi-stage framework for PV and load short-term forecasting: DSO scenarios
Alblawi et al. PV solar power forecasting based on hybrid MFFNN-ALO
Bilal et al. Wind power conversion system model identification using adaptive neuro-fuzzy inference systems: A case study
CN116722545B (en) Photovoltaic power generation prediction method based on multi-source data and related equipment
Boubaker et al. Prediction of Daily Global Solar Radiation using Resilient-propagation Artificial Neural Network and Historical Data: A Case Study of Hail, Saudi Arabia.
CN116455766B (en) Overhead conductor current carrying capacity prediction method and system based on signal sequence decomposition
Yu et al. Correlated load forecasting in active distribution networks using Spatial‐Temporal Synchronous Graph Convolutional Networks

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant