CN110635479B - Intelligent aid decision-making method and system for limiting short-circuit current operation mode - Google Patents

Intelligent aid decision-making method and system for limiting short-circuit current operation mode Download PDF

Info

Publication number
CN110635479B
CN110635479B CN201911022581.8A CN201911022581A CN110635479B CN 110635479 B CN110635479 B CN 110635479B CN 201911022581 A CN201911022581 A CN 201911022581A CN 110635479 B CN110635479 B CN 110635479B
Authority
CN
China
Prior art keywords
operation mode
control measure
short
decision control
sample
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.)
Active
Application number
CN201911022581.8A
Other languages
Chinese (zh)
Other versions
CN110635479A (en
Inventor
黄磊
鲍颜红
刘映尚
任先成
周剑
杨君军
徐光虎
徐伟
张建新
涂旺
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.)
China Southern Power Grid Co Ltd
Nari Technology Co Ltd
NARI Nanjing Control System Co Ltd
Original Assignee
China Southern Power Grid Co Ltd
Nari Technology Co Ltd
NARI Nanjing Control System Co Ltd
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 China Southern Power Grid Co Ltd, Nari Technology Co Ltd, NARI Nanjing Control System Co Ltd filed Critical China Southern Power Grid Co Ltd
Priority to CN201911022581.8A priority Critical patent/CN110635479B/en
Publication of CN110635479A publication Critical patent/CN110635479A/en
Application granted granted Critical
Publication of CN110635479B publication Critical patent/CN110635479B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Power Engineering (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses an intelligent aid decision-making method, device and system for limiting a short-circuit current operation mode. And then, carrying out short-circuit current calculation and safety and stability check on the operation mode after the control measure is taken to determine whether a corresponding auxiliary decision control measure can be adopted or not, and when the prediction model or the result thereof is unavailable, calculating to obtain the conventional auxiliary decision control measure of the current operation mode based on the conventional auxiliary decision control measure of the detailed calculation. The method can realize intelligent aid decision for limiting the short-circuit current operation mode and meet the requirements of on-line aid decision calculation speed and accuracy.

Description

Intelligent aid decision-making method and system for limiting short-circuit current operation mode
Technical Field
The invention relates to the technical field of power system automation, in particular to an intelligent aid decision method and system for limiting a short-circuit current operation mode.
Background
Along with the continuous expansion of the scale of the power grid, the tightness degree of the connection between the power grids is enhanced, the short-circuit current of the receiving-end power grid generally approaches or even exceeds the maximum breaking capacity of a circuit breaker, and measures of breaking and tripping the circuit for a long time are forced to be taken, the structural integrity of the main grid is continuously weakened, the regulation and control of an important power transmission section are difficult, and the power supply capacity and the section power transmission capacity of the main grid are restricted. At present, when a power grid operator arranges an operation mode, the maximum possible short-circuit current is calculated according to the worst mode, and the adopted measures for limiting the short-circuit current are too conservative, so that the safety margin of power grid operation and the power supply reliability are reduced. And the short-circuit current and the control measures of the system are calculated on line according to the current actual operation condition, so that the problem that the control measures are too conservative can be avoided.
The current operation mode for limiting the short-circuit current assists decision calculation, and control measures meeting the requirement that the short-circuit current does not exceed the limit are searched by calculating the sensitivity of control measures for switching on series high reactance, shutdown circuits, shutdown generators, bus split operation or circuit out-of-series operation to the short-circuit current of the out-of-limit bus. Because the grid structure of the power grid is weakened by the control measures, and adverse effects may be brought to the safe and stable operation of the system, the safety and stability checking calculation needs to be performed for a plurality of control schemes, the control measures which are relatively favorable for the safety and stability and the section power transmission capacity are adopted in a plurality of selectable control measures, the safety and stability checking calculation amount for the plurality of control schemes is large, and the calculation time cannot meet the online auxiliary decision calculation requirement.
Disclosure of Invention
The invention aims to provide an intelligent aid decision method and an intelligent aid decision system for limiting a short-circuit current operation mode, so that an intelligent aid decision for limiting the short-circuit current operation mode is realized, and the requirements of online aid decision calculation speed and accuracy are met.
The invention is based on the following inventive concept: the method comprises the steps of performing machine learning by using a support vector machine based on a sample set by using out-of-limit short-circuit current and key influence factors thereof, expected fault safety stability margin influenced by auxiliary decision measures and key influence factors thereof, and maximum transmission capacity of related sections and key influence factors thereof as characteristic quantities to obtain a multi-classification model using auxiliary decision control measures as classification labels; the multi-classification model is applied to the current mode for classification judgment to obtain an auxiliary decision control measure suggestion, and short-circuit current calculation and safety and stability check are carried out on the operation mode after the control measure is taken to determine whether the control measure is taken or not. The short-circuit current limiting operation mode online assistant decision calculation can be performed once every 15min on average, a sample set with extremely large volume can be obtained by accumulating historical data, relevant knowledge can be extracted by data mining from massive data based on a machine learning method, an assistant decision suggestion of the current operation mode is directly given, and the requirement of online assistant decision calculation speed is met.
The technical scheme adopted by the invention is as follows:
in one aspect, the invention provides an intelligent aid decision method for limiting a short-circuit current operation mode, which comprises the following steps:
s1, acquiring operation state data of the power grid in the current operation mode;
s2, calculating short-circuit current, checking expected fault safety and stability and calculating maximum transmission power of the section based on the current operation state data to judge whether the short-circuit current is out of limit, responding to the short-circuit current out of limit and the adopted control measures in the current operation mode, turning to S3, otherwise ending the auxiliary decision flow;
s3, selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short circuit current out-of-limit element and the adopted control measure;
s4, extracting characteristic quantities of the current operation mode, judging whether the selected prediction model is suitable for the current operation mode according to a preset applicability judgment rule, responding to the applicability of the selected prediction model, and turning to the step S5, otherwise, turning to the step S73;
s5, obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
s6, calculating short-circuit current, calculating expected fault safety and stability, calculating maximum transmission power of the section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirements: if all the short-circuit current is satisfied, turning to step S71, if the short-circuit current out-of-limit element still exists, turning to step S72, otherwise, turning to step S73;
s71, the assistant decision control measure is suggested as the assistant decision control measure of the current operation mode, and the assistant decision process of the current operation mode is ended;
s72, after accepting the suggestion of the assistant decision control measure, calculating the conventional assistant decision control measure aiming at the current operation mode, taking the obtained assistant decision control measure as an additional assistant decision control measure of the current operation mode, and ending the assistant decision flow of the current operation mode;
and S73, performing conventional auxiliary decision control measure calculation aiming at the current operation mode, taking the auxiliary decision control measure obtained by calculation as the auxiliary decision control measure of the current operation mode, and then ending the auxiliary decision flow of the current operation mode.
In the present invention, the decision-making-assisted control measures may include not only control measures for out-of-limit short-circuit current but also recovery measures for control measures that have already been taken.
Further, in the method of the present invention, if the calculation results in S6 are not all satisfied, the method further includes:
s8, carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode, and respectively carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
s9, updating a training sample set of the assistant decision control measure prediction model by taking the current operation mode and the final assistant decision control measure thereof, and the nearby operation mode and the conventional assistant decision control measure thereof as new samples;
and S10, dividing sample subsets and training models corresponding to the sample subsets based on the updated training sample set to obtain new auxiliary decision control measure prediction models corresponding to the sample subsets for subsequent auxiliary decision control measure calculation.
Optionally, in step S8, the monte carlo sampling of the current operation mode is:
randomly sampling the generator terminal voltage in the range of [0.95, 1.05] of the rated voltage; considering the active output and reactive output constraints of the units, randomly sampling the active output of each unit within the range of 20% of the output of the current mode, and keeping the proportion of the reactive output change to the active output change consistent; randomly sampling the active power of each load within 20 percent of the current load power, and keeping the change proportion of the reactive load and the active load consistent; in order to ensure active power balance, the power of each load in the current mode is adjusted according to the same proportion to ensure power generation and load power balance, and a new random mode sample, namely a nearby operation mode sample of the current operation mode, is generated.
Optionally, the method for constructing the pre-constructed auxiliary decision control measure prediction model includes:
acquiring historical short-circuit current out-of-limit event information as an initial sample set;
in response to the sample set being updated, obtaining an updated sample set;
dividing a sample set into a plurality of sample subsets according to short-circuit current out-of-limit elements and control measures which are adopted before aid decision making;
respectively extracting characteristic quantities of each sample subset with 2 or more auxiliary decision control measures in the sample subsets;
based on the extracted characteristic quantity, taking the auxiliary decision control measures included in the sample subset as classification labels, and carrying out classification prediction model training to obtain an auxiliary decision control measure prediction model of the corresponding sample subset;
only 1 auxiliary decision control measure in the sample subset exists, and the corresponding output of the prediction model is always the corresponding auxiliary decision control measure.
Optionally, dividing the sample set into a plurality of sample subsets is:
according to the short-circuit current out-of-limit element and the control measures which are taken before the aid decision, if a plurality of samples simultaneously have the short-circuit current out-of-limit element and the control measures which are taken, the plurality of samples which are identical in short-circuit current out-of-limit element and the control measures which are taken are divided into the same sample subset, otherwise, the samples are divided into the sample subsets which are identical in short-circuit current out-of-limit element or the control measures which are taken according to the short-circuit current out-of-limit element or the control measures which are taken.
Optionally, the auxiliary decision control measure prediction model adopts a support vector machine classification model;
aiming at each sample subset with 2 auxiliary decision control measures in the sample subset, carrying out two-classification model training by taking the two auxiliary decision control measures as classification labels;
and aiming at each sample subset with 3 or more auxiliary decision control measures in the sample subsets, respectively carrying out two-classification model training aiming at each auxiliary decision control measure. Namely, one sample of the assistant decision control measures is taken as a positive sample, and all samples of the assistant decision control measures are left as negative samples, aiming atiIndividual aid decision control measure co-trainingiAnd (4) classifying the models.
The prediction model training aiming at each sample subset and the training of each two-classification model in various sample subsets can adopt a distributed parallel computing technology, so that the efficiency of model training is improved.
Optionally, in step S5, if there are 3 or more auxiliary decision control measures in the sample subset corresponding to the current operation mode, each of the two classification models is used to perform prediction to obtain whether an auxiliary decision control measure can be taken, and then the auxiliary decision control measure obtained by the two classification models with the highest classification prediction confidence is selected from the two classification models and used as an auxiliary decision control measure suggestion of the current operation mode.
Optionally, for a subset of samples having 2 or more auxiliary decision control measures, if the confidence of the classification prediction result of the two-classification prediction model is smaller than a specified threshold value, the prediction result of the classification model is considered to be unreliable, and then the control measure recommendation of the two-classification prediction model is not considered.
The classification prediction confidence of the two classification models can be obtained by calculating a Sigmoid function value of the distance from the sample to be identified to the optimal classification hyperplane of the support vector machine.
Optionally, in step S3, the extracted sample feature quantities include an out-of-limit element short-circuit current and its key influence factors, a short-circuit current and its key influence factors of an element with a large short-circuit current that has been influenced by a control measure, an expected fault safety margin and key influence factors that have a large influence on safety stability by a control measure and a candidate aid decision control measure, and a maximum transmission power of a relevant section and its key influence factors that have a large influence by a control measure and a candidate aid decision control measure;
key contributors to short circuit current include: the method comprises the following steps that the total load and the total power generation output of an area where an out-of-limit element is located, the on-off state of a unit with short-circuit current sensitivity larger than a threshold value, the series high-resistance switching-on/off state of a line with short-circuit current sensitivity larger than the threshold value, the switching-on/off state of the line with short-circuit current sensitivity larger than the threshold value and a transformer, and the out-of-series or bus splitting running state of the line of a plant where the out-of;
the elements that have taken control measures to affect the short-circuit current to a large extent are: a short-circuit current element with the sensitivity of the control measure to the short-circuit current larger than a specified threshold value;
the expected faults with great influence of control measures on the safety stability are as follows: calculating equivalent impedance between a fault point of a fault i in a safety and stability check fault set and a control measure c, taking the equivalent impedance as an evaluation index of the control measure on the influence of the safety and stability of an expected fault, judging whether the index is smaller than a set threshold value, if so, considering the fault as the expected fault of which the control measure c has a larger influence on the safety and stability of the expected fault, wherein the safety and stability margin of the expected fault i before safety and stability check is calculated for an auxiliary decision;
failsafe stability critical influencing factors are envisioned including: key influencing factors of thermal stability, transient power angle stability and transient voltage stability; the key influence factors of thermal stability comprise the power flow and the switching-on/off state of a fault branch and a post-fault power flow transfer branch; the key influence factors of the transient power angle stability comprise the output and the switching-on and switching-off states of a critical group unit and the rest group units with larger participation factors, and the power flow and the switching-on and switching-off states of the contact branch of the oscillation center; the key influence factors of the transient voltage stability comprise the apparent power and the reactive voltage sensitivity of the transient voltage weak node;
the relevant sections with greater influence of the control measures are: if the expected fault associated with the maximum transmission power calculation of the section is a fault with a large influence on the safety and stability of the control measure, the section is considered to be a related section with a large influence on the control measure;
key influencing factors of the maximum transmission power of the relevant section include: the section of the related section constitutes element tide and a switching-on/off state, section related faults and safety and stability key influence factors thereof.
Optionally, in step S4, the preset applicability judgment rule is:
calculating the standardized Euclidean distances between the current operation mode and the sample subset characteristic quantity of all samples in the sample subset, and if the distances between the current operation mode and a preset number k of nearest samples in the sample subset are smaller than a specified threshold value, judging that the prediction model is suitable for the current operation mode; otherwise, the prediction model is not suitable for the current operation mode, and the process goes to step S73.
On the other hand, the invention also provides an intelligent aid decision-making device for limiting the short-circuit current operation mode, which comprises:
the operation mode data acquisition module is used for acquiring operation state data of the power grid in the current operation mode;
the comprehensive calculation module is used for calculating the short-circuit current, the expected fault safety and stability check calculation and the maximum transmission power of the section based on the current running state data so as to judge whether the short-circuit current exceeds the limit and the adopted control measures exist or not;
the prediction model selection module is used for selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short-circuit current out-of-limit element and the adopted control measure, and ending the auxiliary decision flow if the short-circuit current out-of-limit element and the adopted control measure do not exist;
the prediction module is used for extracting characteristic quantities of the current operation mode, selecting a prediction model suitable for the current operation mode according to a preset applicability judgment rule, and obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
the secondary comprehensive calculation module is used for calculating short-circuit current, calculating expected fault safety and stability check and calculating maximum transmission power of a section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirement or not;
the conventional auxiliary decision control measure calculation module is used for performing the conventional auxiliary decision control measure calculation based on detailed calculation to obtain the conventional auxiliary decision control measure when the calculation results of the secondary comprehensive calculation module are not completely met;
and an aid decision control measure determination module for: in response to the fact that the results of the secondary comprehensive calculation all meet the requirements, the auxiliary decision control measures obtained by the prediction model are suggested to be used as the auxiliary decision control measures of the current operation mode; in response to the fact that a short-circuit current out-of-limit element still exists in the result of the secondary comprehensive calculation, the aid decision control measure suggestion obtained by the prediction model and the conventional aid decision control measure are jointly used as the final aid decision control measure of the current operation mode; and in response to other situations where the results of the calculations are not satisfactory, taking the conventional aid decision control measure as the final aid decision control measure.
Further, the intelligent aid decision device for limiting the short-circuit current operation mode further comprises:
the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode and carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
the sample set updating module is used for updating a training sample set of the auxiliary decision control measure prediction model by taking the current operation mode and the final auxiliary decision control measure thereof, and the nearby operation mode and the conventional auxiliary decision control measure thereof as new samples;
and the prediction model training module is used for carrying out sample subset division and model training corresponding to each sample subset based on the initial sample set or the updated training sample set to obtain a prediction model corresponding to each sample subset, and the prediction model is used for calculating follow-up auxiliary decision control measures.
The method can be used for updating the prediction model, so that the assistant decision control measures can be obtained more efficiently during the assistant decision calculation of the subsequent operation mode.
In a third aspect, the present invention further provides an intelligent aid decision system for limiting a short-circuit current operation mode, including:
the sample characteristic extraction module is used for extracting characteristic quantities of all samples in the sample set according to a pre-specified key characteristic quantity type;
the prediction model training module is used for carrying out prediction model training according to the classification label of each sample in the initial sample or the updated sample set and establishing a mapping relation between the sample characteristic quantity and the classification label;
the auxiliary decision control measure prediction module is used for obtaining the auxiliary decision control measure of the predicted current mode by utilizing the prediction model;
the expected fault safety and stability checking and calculating module is used for forming load flow and stability calculation data based on the power grid equipment model parameters and the state estimation result, and performing expected fault safety and stability checking and calculating of the current mode or the mode after the implementation of the control measures by combining static and transient safety and stability evaluation calculation parameters, equipment safety and stability limit values and expected fault set data to obtain expected fault thermal stability, transient power angle stability, transient voltage stability margin and key influence factors thereof;
the short-circuit current calculation module is used for performing short-circuit current checking calculation of a current mode or a mode after implementation of control measures on the basis of load flow, stable calculation data, short-circuit current calculation parameters, a bus and line short-circuit equipment set and a breaker breaking capacity limit value to obtain bus and line short-circuit currents and judging whether the short-circuit currents of the buses and the lines exceed the rated breaking capacity of the breaker and key influence factors of the short-circuit currents;
the section maximum transmission power calculation module is used for calculating the section limit power of a mode after the implementation of the current mode or the control measure based on section association expected failure safety and stability check calculation data and a section power adjustment mode to obtain the section maximum transmission power;
the conventional auxiliary decision control measure detailed calculation module is used for searching control measures meeting the short-circuit current out-of-limit requirement and recovery measures of the adopted control measures aiming at the short-circuit current out-of-limit element and the adopted short-circuit current control measures by taking the minimum influence degree on safety and stability and the minimum switch of an operation element as control targets;
and the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to generate random samples near the current mode, carrying out auxiliary decision control measure calculation based on a conventional detailed calculation method on each newly generated random sample, and then storing all operation modes and calculation results thereof into a sample set.
Advantageous effects
In the invention, auxiliary decision control measure prediction models corresponding to different current out-of-limit elements and short circuit control measures are obtained through machine learning, and in the control process, when the short circuit current operation mode needs to be limited, the corresponding prediction models can be selected according to the current out-of-limit elements obtained through monitoring and the short circuit control measures adopted, so that the auxiliary decision control measures can be predicted quickly. And then, carrying out short-circuit current calculation and safety and stability check on the operation mode after the control measures are taken to determine whether corresponding auxiliary decision control measures can be taken.
When the predicted assistant decision control measures can not limit or completely limit the short-circuit current operation mode, the method obtains the supplement of the assistant decision control measures of the current operation mode through the calculation of the conventional assistant decision control measures based on detailed calculation, simultaneously obtains the random nearby operation mode of the current operation mode through Monte Carlo sampling, obtains the conventional assistant decision control measures of each operation mode through the calculation of the conventional assistant decision control measures based on the detailed calculation, then expands and updates the original training sample set based on the assistant decision control measures of each operation mode to obtain a new prediction model through training, and further is used for the assistant decision control measure prediction of the subsequent operation mode so as to continuously optimize the prediction model, thereby continuously improving the efficiency and the reliability of the assistant decision control measure calculation.
In conclusion, the intelligent aid decision-making method can realize the intelligent aid decision-making for limiting the short-circuit current operation mode, and can meet the requirements of the online aid decision-making on the calculation speed and accuracy.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of the method of the present invention;
FIG. 2 is a schematic diagram illustrating a predictive model update process according to an embodiment of the invention;
fig. 3 is a schematic diagram illustrating an initial prediction model building process according to an embodiment of the present invention.
Detailed Description
The following further description is made in conjunction with the accompanying drawings and the specific embodiments.
Example 1
Referring to fig. 1, the intelligent aid decision method for limiting the short-circuit current operation mode in the present embodiment includes:
s1, acquiring operation state data of the power grid in the current operation mode;
s2, calculating short-circuit current, checking expected fault safety and stability and calculating maximum transmission power of the section based on the current operation state data to judge whether the short-circuit current is out of limit, responding to the short-circuit current out of limit and the adopted control measures in the current operation mode, turning to S3, otherwise ending the auxiliary decision flow;
s3, selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short circuit current out-of-limit element and the adopted control measure;
s4, extracting characteristic quantities of the current operation mode, judging whether the selected prediction model is suitable for the current operation mode according to a preset applicability judgment rule, responding to the applicability of the selected prediction model, and turning to the step S5, otherwise, turning to the step S73;
s5, obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
s6, calculating short-circuit current, calculating expected fault safety and stability, calculating maximum transmission power of the section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirements: if all the short-circuit current is satisfied, turning to step S71, if the short-circuit current out-of-limit element still exists, turning to step S72, otherwise, turning to step S73;
s71, the assistant decision control measure is suggested as the assistant decision control measure of the current operation mode, and the assistant decision process of the current operation mode is ended;
s72, after accepting the suggestion of the assistant decision control measure, calculating the conventional assistant decision control measure aiming at the current operation mode, taking the obtained assistant decision control measure as an additional assistant decision control measure of the current operation mode, and ending the assistant decision flow of the current operation mode;
and S73, performing conventional auxiliary decision control measure calculation aiming at the current operation mode, taking the auxiliary decision control measure obtained by calculation as the auxiliary decision control measure of the current operation mode, and then ending the auxiliary decision flow of the current operation mode.
Decision-aiding control measures may include not only control measures for out-of-limit short circuit current, but also recovery measures for control measures that have already been taken. Because the assistant decision control measures are provided by the prediction model, and the output classification of the prediction model is based on the sample set, the assistant decision control measures related by the invention are actually various different schemes calculated by the conventional assistant decision control measures based on detailed calculation in the historical assistant decision, so the specific contents of the assistant decision control measures are not repeated in the invention.
The conventional decision-making-assisting control measures based on detailed calculation which can be adopted by the invention are calculated by searching the control measures meeting the short-circuit current out-of-limit requirement and the recovery measures of the adopted control measures aiming at the short-circuit current out-of-limit element and the adopted short-circuit current control measures, wherein the control aims are that the influence degree on safety and stability is minimum and the switch of an operation element is minimum.
Examples 1 to 1
On the basis of embodiment 1, in this embodiment:
the construction method of the initial assistant decision control measure prediction model comprises the following steps:
acquiring historical short-circuit current out-of-limit event information as an initial sample set;
in response to the sample set being updated, obtaining an updated sample set;
dividing a sample set into a plurality of sample subsets according to short-circuit current out-of-limit elements and control measures which are adopted before aid decision making;
respectively extracting characteristic quantities of each sample subset with 2 or more auxiliary decision control measures in the sample subsets;
based on the extracted characteristic quantity, taking the auxiliary decision control measures included in the sample subset as classification labels, and carrying out classification prediction model training to obtain an auxiliary decision control measure prediction model of the corresponding sample subset;
only 1 auxiliary decision control measure in the sample subset exists, and the corresponding output of the prediction model is always the corresponding auxiliary decision control measure.
In order to enable the prediction model to be learned and optimized in continuous application, if the calculation result in S6 is not all satisfied, the method further includes:
s8, carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode, and respectively carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
s9, updating a training sample set of the assistant decision control measure prediction model by taking the current operation mode and the final assistant decision control measure thereof, and the nearby operation mode and the conventional assistant decision control measure thereof as new samples;
and S10, dividing sample subsets and training models corresponding to the sample subsets based on the updated training sample set to obtain new auxiliary decision control measure prediction models corresponding to the sample subsets for subsequent auxiliary decision control measure calculation.
In step S8, the monte carlo sampling of the current operation mode is:
randomly sampling the generator terminal voltage in the range of [0.95, 1.05] of the rated voltage; considering the active output and reactive output constraints of the units, randomly sampling the active output of each unit within the range of 20% of the output of the current mode, and keeping the proportion of the reactive output change to the active output change consistent; randomly sampling the active power of each load within 20 percent of the current load power, and keeping the change proportion of the reactive load and the active load consistent; in order to ensure active power balance, the power of each load in the current mode is adjusted according to the same proportion to ensure power generation and load power balance, and a new random mode sample, namely a nearby operation mode sample of the current operation mode, is generated.
The dividing of the sample set into a plurality of sample subsets is:
according to the short-circuit current out-of-limit element and the control measures which are taken before the aid decision, if a plurality of samples simultaneously have the short-circuit current out-of-limit element and the control measures which are taken, a plurality of samples which are respectively identical to the short-circuit current out-of-limit element and the control measures which are taken are divided into the same sample subset, otherwise, the samples are divided into the sample subsets which are identical to the short-circuit current out-of-limit element or the control measures which are taken according to the short-circuit current out-of-limit element or the control measures which are taken.
The auxiliary decision control measure prediction model adopts a support vector machine classification model;
aiming at each sample subset with 2 auxiliary decision control measures in the sample subset, performing model training by taking the two auxiliary decision control measures as classification labels;
and aiming at each sample subset with 3 or more auxiliary decision control measures in the sample subsets, respectively carrying out two-classification model training aiming at each auxiliary decision control measure.
The prediction model training aiming at each sample subset and the training of each two-classification model in various sample subsets can adopt a distributed parallel computing technology, so that the efficiency of model training is improved.
In step S5, if there are 3 or more auxiliary decision control measures in the sample subset corresponding to the current operation mode, each of the two classification models is used to perform prediction to obtain whether an auxiliary decision control measure can be taken, and then the auxiliary decision control measure obtained by the two classification models with the highest classification prediction confidence is selected from the two classification models and used as an auxiliary decision control measure suggestion of the current operation mode.
And aiming at the situation that 2 or more auxiliary decision control measures exist in the sample subset, if the confidence coefficient of the classification prediction result of the two-classification prediction model is smaller than a specified threshold value, the prediction result of the classification model is considered to be unreliable, and the control measure suggestion of the two-classification model prediction is not considered subsequently.
The classification prediction confidence of the two-classification model can be obtained by calculating a Sigmoid function value of the distance from the sample to be recognized to the optimal classification hyperplane of the support vector machine.
In step S3, the extracted sample feature quantities include an out-of-limit element short-circuit current and its key influence factors, a short-circuit current and its key influence factors of an element with a large short-circuit current that has been influenced by a control measure, an expected failure safety margin and key influence factors that have been influenced by a control measure and a candidate aid decision control measure to safety stability, and a maximum transmission power of a relevant section and its key influence factors that have been influenced by a control measure and a candidate aid decision control measure;
key contributors to short circuit current include: the method comprises the following steps that the total load and the total power generation output of an area where an out-of-limit element is located, the on-off state of a unit with short-circuit current sensitivity larger than a threshold value, the series high-resistance switching-on/off state of a line with short-circuit current sensitivity larger than the threshold value, the switching-on/off state of the line with short-circuit current sensitivity larger than the threshold value and a transformer, and the out-of-series or bus splitting running state of the line of a plant where the out-of;
the elements that have taken control measures to affect the short-circuit current to a large extent are: a short-circuit current element with the sensitivity of the control measure to the short-circuit current larger than a specified threshold value;
the expected faults with great influence of control measures on the safety stability are as follows: calculating equivalent impedance between a fault point of a fault i in a safety and stability check fault set and a control measure c, taking the equivalent impedance as an evaluation index of the control measure on the influence of the safety and stability of an expected fault, judging whether the index is smaller than a set threshold value, if so, considering the fault as the expected fault of which the control measure c has a larger influence on the safety and stability of the expected fault, wherein the safety and stability margin of the expected fault i before safety and stability check is calculated for an auxiliary decision;
failsafe stability critical influencing factors are envisioned including: key influencing factors of thermal stability, transient power angle stability and transient voltage stability; the key influence factors of thermal stability comprise the power flow and the switching-on/off state of a fault branch and a post-fault power flow transfer branch; the key influence factors of the transient power angle stability comprise the output and the switching-on and switching-off states of a critical group unit and the rest group units with larger participation factors, and the power flow and the switching-on and switching-off states of the contact branch of the oscillation center; the key influence factors of the transient voltage stability comprise the apparent power and the reactive voltage sensitivity of the transient voltage weak node;
the relevant sections with greater influence of the control measures are: if the expected fault associated with the maximum transmission power calculation of the section is a fault with a large influence on the safety and stability of the control measure, the section is considered to be a related section with a large influence on the control measure;
key influencing factors of the maximum transmission power of the relevant section include: the section of the related section constitutes element tide and a switching-on/off state, section related faults and safety and stability key influence factors thereof.
In step S4, the preset applicability judgment rule is:
calculating the standardized Euclidean distances between the current operation mode and the sample subset characteristic quantity of all samples in the sample subset, and if the distances between the current operation mode and a preset number k of nearest samples in the sample subset are smaller than a specified threshold value, judging that the prediction model is suitable for the current operation mode; otherwise, the prediction model is not suitable for the current operation mode, and the process goes to step S73.
Examples 1 to 2
Based on the embodiment 1-1, the specific operation flow of the method of the embodiment includes:
1: dividing samples in the sample set into a plurality of subsets according to control measures which are already taken before short-circuit current out-of-limit element and auxiliary decision calculation, and respectively extracting key characteristic quantities for each sample in the sample subsets;
after the sample subsets are divided, for a plurality of samples in one sample set, if a short-circuit current out-of-limit element exists, the out-of-limit elements are the same, and if a control measure which is already adopted exists before the auxiliary decision calculation, the control measure is the same.
The key feature quantities extracted for each sample in the sample subset are as follows:
the key characteristic quantities comprise out-of-limit element short-circuit current and key influence factors, short-circuit current and key influence factors of elements with larger short-circuit current influenced by control measures, expected fault safety stability margin and key influence factors with larger influence on safety stability by the control measures and candidate auxiliary decision control measures, maximum transmission power of relevant sections and key influence factors influenced by the control measures and the candidate auxiliary decision control measures, and all sample characteristic quantities in the sample subset are taken as sample subset characteristic quantities.
2: and respectively carrying out auxiliary decision control measure prediction model training aiming at each sample subset, if different auxiliary decision control measures are included in each sample subset, carrying out classification prediction model training by taking the auxiliary decision control measures as classification labels, and otherwise, not carrying out prediction model training.
And (3) respectively carrying out auxiliary decision control measure prediction model training aiming at each sample subset, and carrying out prediction model training aiming at each sample subset in parallel by adopting a distributed parallel computing technology. The classification prediction model is a support vector machine classification model.
If the class i of the auxiliary decision control measure in the sample subset is more than or equal to 3, i classification labels are required to be trained into two classification models belonging to the class control measure and the non-class control measure respectively, and each two classification model can be trained in parallel by adopting a distributed parallel computing technology.
3: and performing short-circuit current calculation, expected fault safety and stability check calculation and section maximum transmission power calculation of the current operation mode in parallel based on the cluster computing platform. If the current mode has no short-circuit current out-of-limit and control measures which are already taken, the method is directly ended; otherwise, judging the sample subset according to the short-circuit current out-of-limit element and the control measure which is already taken, and obtaining the auxiliary decision control measure suggestion of the current mode based on the prediction model of the sample subset. And if the assistant decision control measure suggestion of the current mode cannot be obtained, performing assistant decision control measure calculation based on a conventional detailed calculation method, and then turning to the step 5.
Before obtaining the auxiliary decision control measure suggestion of the current mode based on the sample subset prediction model, the applicability judgment of the prediction model is needed, and the applicability judgment method comprises the following steps:
and (3) calculating the normalized Euclidean distance between the current operation mode and the sample subset characteristic quantity of all samples in the sample subset, and if the distances between the current sample and k preset nearest samples in the sample subset are all smaller than a specified threshold value, judging that the prediction model is suitable for the current operation mode. Otherwise, the prediction model is not suitable for the current operation mode, and the auxiliary decision control measure suggestion cannot be obtained.
The method for obtaining the auxiliary decision control measure suggestion of the current mode based on the sample subset prediction model specifically comprises the following steps: if only one auxiliary decision control measure exists in the sample subset, the auxiliary decision control measure is directly suggested as the auxiliary decision control measure of the current mode. If a plurality of two-classification prediction models exist in the sample subset, the prediction models are respectively adopted to obtain whether the corresponding control measure suggestions are adopted, and then the measure with the highest classification prediction confidence coefficient is selected from all possible control measures to serve as the auxiliary decision control measure suggestions of the current mode. If the confidence coefficient of the classification prediction model result is smaller than the specified threshold value, the prediction result is considered to be unreliable, and the control measure suggestion of the classification model prediction is not considered subsequently.
4: and (4) adjusting the current operation mode according to the auxiliary decision control measure suggestion, and then performing short-circuit current calculation, expected fault safety and stability check calculation and section maximum transmission power calculation in parallel based on the cluster calculation platform. If the short-circuit current, the expected fault safety margin and the maximum transmission power of the section meet the requirements, obtaining an auxiliary decision control measure of the current mode; otherwise, performing auxiliary decision control measure calculation based on a conventional detailed calculation method;
if the short-circuit current out-of-limit element still exists in the short-circuit current calculation result after the current operation mode is adjusted according to the auxiliary decision control measure suggestion through calculation, the auxiliary decision control measure suggestion is accepted, and the auxiliary decision control measure calculation based on the conventional detailed calculation method is carried out on the basis to serve as an additional auxiliary decision control measure; and if the expected fault safety stability margin and/or the maximum transmission power of the section do not meet the requirements, directly performing auxiliary decision control measure calculation based on a conventional detailed calculation method, and only taking the conventional auxiliary decision control measure as a final auxiliary decision control measure of the current operation mode.
5: monte Carlo sampling is carried out on the current operation mode to generate random samples near the current mode, and auxiliary decision control measure calculation based on a conventional detailed calculation method is carried out on each newly generated random sample.
And then storing the advices (if any) of the auxiliary decision control measures adopted by the current operation mode, detailed calculation results of the conventional auxiliary decision control measures, the current operation mode, the nearby operation mode and the detailed calculation results of the conventional auxiliary decision control measures in a sample set to expand and update the sample set, returning to the step 1, and performing sample subset division and prediction model training again to obtain a new prediction model for predicting the subsequent auxiliary decision control measures.
Example 2
Based on the same inventive concept as embodiment 1, this embodiment is an intelligent aid decision device for limiting a short-circuit current operation mode, including:
the operation mode data acquisition module is used for acquiring operation state data of the power grid in the current operation mode;
the comprehensive calculation module is used for calculating the short-circuit current, the expected fault safety and stability check calculation and the maximum transmission power of the section based on the current running state data so as to judge whether the short-circuit current exceeds the limit and the adopted control measures exist or not;
the prediction model selection module is used for selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short-circuit current out-of-limit element and the adopted control measure, and ending the auxiliary decision flow if the short-circuit current out-of-limit element and the adopted control measure do not exist;
the prediction module is used for extracting characteristic quantities of the current operation mode, selecting a prediction model suitable for the current operation mode according to a preset applicability judgment rule, and obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
the secondary comprehensive calculation module is used for calculating short-circuit current, calculating expected fault safety and stability check and calculating maximum transmission power of a section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirement or not;
the conventional auxiliary decision control measure calculation module is used for performing the conventional auxiliary decision control measure calculation based on detailed calculation to obtain the conventional auxiliary decision control measure when the calculation results of the secondary comprehensive calculation module are not completely met;
and an aid decision control measure determination module for: in response to the fact that the results of the secondary comprehensive calculation all meet the requirements, the auxiliary decision control measures obtained by the prediction model are suggested to be used as the auxiliary decision control measures of the current operation mode; in response to the fact that a short-circuit current out-of-limit element still exists in the result of the secondary comprehensive calculation, the aid decision control measure suggestion obtained by the prediction model and the conventional aid decision control measure are jointly used as the final aid decision control measure of the current operation mode; and in response to other situations where the results of the calculations are not satisfactory, taking the conventional aid decision control measure as the final aid decision control measure.
In order to optimize the prediction efficiency and reliability of the prediction model in the continuous prediction and to obtain the aid decision control measure more efficiently in the aid decision calculation of the subsequent operation mode, the intelligent aid decision device of the embodiment further includes:
the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode and carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
the sample set updating module is used for updating a training sample set of the auxiliary decision control measure prediction model by taking the current operation mode and the final auxiliary decision control measure thereof, and the nearby operation mode and the conventional auxiliary decision control measure thereof as new samples;
and the prediction model training module is used for carrying out sample subset division and model training corresponding to each sample subset based on the initial sample set or the updated training sample set to obtain a prediction model corresponding to each sample subset, and the prediction model is used for calculating follow-up auxiliary decision control measures.
The algorithm for realizing the relevant functions by the modules refers to the embodiment 1, the embodiment 1-1 and the embodiment 1-2.
Example 3
Based on the same inventive concept as that of the embodiment 1 and the embodiment 2, the present embodiment is an intelligent assistant decision system for limiting a short-circuit current operation mode, including:
the sample characteristic extraction module is used for extracting characteristic quantities of all samples in the sample set according to a pre-specified key characteristic quantity type;
the prediction model training module is used for carrying out prediction model training according to the classification label of each sample in the initial sample or the updated sample set and establishing a mapping relation between the sample characteristic quantity and the classification label;
the auxiliary decision control measure prediction module is used for obtaining the auxiliary decision control measure of the predicted current mode by utilizing the prediction model;
the expected fault safety and stability checking and calculating module is used for forming load flow and stability calculation data based on the power grid equipment model parameters and the state estimation result, and performing expected fault safety and stability checking and calculating of the current mode or the mode after the implementation of the control measures by combining static and transient safety and stability evaluation calculation parameters, equipment safety and stability limit values and expected fault set data to obtain expected fault thermal stability, transient power angle stability, transient voltage stability margin and key influence factors thereof;
the short-circuit current calculation module is used for performing short-circuit current checking calculation of a current mode or a mode after implementation of control measures on the basis of load flow, stable calculation data, short-circuit current calculation parameters, a bus and line short-circuit equipment set and a breaker breaking capacity limit value to obtain bus and line short-circuit currents and judging whether the short-circuit currents of the buses and the lines exceed the rated breaking capacity of the breaker and key influence factors of the short-circuit currents;
the section maximum transmission power calculation module is used for calculating the section limit power of a mode after the implementation of the current mode or the control measure based on section association expected failure safety and stability check calculation data and a section power adjustment mode to obtain the section maximum transmission power;
the conventional auxiliary decision control measure detailed calculation module is used for searching control measures meeting the short-circuit current out-of-limit requirement and recovery measures of the adopted control measures aiming at the short-circuit current out-of-limit element and the adopted short-circuit current control measures by taking the minimum influence degree on safety and stability and the minimum switch of an operation element as control targets;
and the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to generate random samples near the current mode, carrying out auxiliary decision control measure calculation based on a conventional detailed calculation method on each newly generated random sample, and then storing all operation modes and calculation results thereof into a sample set.
The algorithm for realizing the relevant functions by the modules refers to the embodiment 1, the embodiment 1-1 and the embodiment 1-2.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (10)

1. An intelligent aid decision method for limiting short-circuit current operation mode is characterized by comprising the following steps:
s1, acquiring operation state data of the power grid in the current operation mode;
s2, calculating short-circuit current, checking expected fault safety and stability and calculating maximum transmission power of the section based on the current operation state data to judge whether the short-circuit current is out of limit, responding to the short-circuit current out of limit and the adopted control measures in the current operation mode, turning to S3, otherwise ending the auxiliary decision flow;
s3, selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short circuit current out-of-limit element and the adopted control measure;
s4, extracting characteristic quantities of the current operation mode, judging whether the selected prediction model is suitable for the current operation mode according to a preset applicability judgment rule, responding to the applicability of the selected prediction model, and turning to the step S5, otherwise, turning to the step S73;
s5, obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
s6, calculating short-circuit current, calculating expected fault safety and stability, calculating maximum transmission power of the section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirements: if all the short-circuit current is satisfied, turning to step S71, if the short-circuit current out-of-limit element still exists, turning to step S72, otherwise, turning to step S73;
s71, the assistant decision control measure is suggested as the assistant decision control measure of the current operation mode, and the assistant decision process of the current operation mode is ended;
s72, after accepting the suggestion of the assistant decision control measure, calculating the conventional assistant decision control measure aiming at the current operation mode, taking the obtained assistant decision control measure as an additional assistant decision control measure of the current operation mode, and ending the assistant decision flow of the current operation mode;
s73, calculating a conventional assistant decision control measure according to the current operation mode, taking the assistant decision control measure obtained by calculation as the assistant decision control measure of the current operation mode, and then ending the assistant decision flow of the current operation mode;
the method for constructing the pre-constructed auxiliary decision control measure prediction model comprises the following steps:
acquiring historical short-circuit current out-of-limit event information as an initial sample set;
in response to the sample set being updated, obtaining an updated sample set;
dividing a sample set into a plurality of sample subsets according to short-circuit current out-of-limit elements and control measures which are adopted before aid decision making;
respectively extracting characteristic quantities of each sample subset with 2 or more auxiliary decision control measures in the sample subsets;
based on the extracted characteristic quantity, taking the auxiliary decision control measures included in the sample subset as classification labels, and carrying out classification prediction model training to obtain an auxiliary decision control measure prediction model of the corresponding sample subset;
only 1 sample subset of assistant decision control measures exists in the sample subsets, and the output of the corresponding prediction model is always the corresponding assistant decision control measure;
the preset applicability judgment rule is as follows:
calculating the standardized Euclidean distances between the current operation mode and the sample subset characteristic quantity of all samples in the sample subset, and if the distances between the current operation mode and a preset number k of nearest samples in the sample subset are smaller than a specified threshold value, judging that the prediction model is suitable for the current operation mode; otherwise, the prediction model is not suitable for the current operation mode, and the process goes to step S73.
2. The method of claim 1, wherein if the results of the calculation in S6 are not all satisfied, the method further comprises:
s8, carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode, and respectively carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
s9, updating a training sample set of the assistant decision control measure prediction model by taking the current operation mode and the final assistant decision control measure thereof, and the nearby operation mode and the conventional assistant decision control measure thereof as new samples;
and S10, dividing sample subsets and training models corresponding to the sample subsets based on the updated training sample set to obtain new auxiliary decision control measure prediction models corresponding to the sample subsets for subsequent auxiliary decision control measure calculation.
3. The method of claim 2, wherein in step S8, the monte carlo sampling of the current operating mode is:
randomly sampling the generator terminal voltage in the range of [0.95, 1.05] of the rated voltage; considering the active output and reactive output constraints of the units, randomly sampling the active output of each unit within the range of 20% of the output of the current mode, and keeping the proportion of the reactive output change to the active output change consistent; randomly sampling the active power of each load within 20 percent of the current load power, and keeping the change proportion of the reactive load and the active load consistent; in order to ensure active power balance, the power of each load in the current mode is adjusted according to the same proportion to ensure power generation and load power balance, and a new random mode sample, namely a nearby operation mode sample of the current operation mode, is generated.
4. The method of claim 1, wherein dividing the sample set into a plurality of sample subsets is:
according to the short-circuit current out-of-limit element and the control measures which are taken before the aid decision, if a plurality of samples simultaneously have the short-circuit current out-of-limit element and the control measures which are taken, the plurality of samples which are identical in short-circuit current out-of-limit element and the control measures which are taken are divided into the same sample subset, otherwise, the samples are divided into the sample subsets which are identical in short-circuit current out-of-limit element or the control measures which are taken according to the short-circuit current out-of-limit element or the control measures which are taken.
5. The method of claim 1 or 2, wherein the aided decision control measure prediction model employs a support vector machine classification model;
aiming at each sample subset with 2 auxiliary decision control measures in the sample subset, performing model training by taking the two auxiliary decision control measures as classification labels;
and aiming at each sample subset with 3 or more auxiliary decision control measures in the sample subsets, respectively carrying out two-classification model training aiming at each auxiliary decision control measure.
6. The method as claimed in claim 5, wherein in step S5, if there are 3 or more auxiliary decision control measures in the sample subset corresponding to the current operation mode, the two classification models are used to perform prediction to obtain whether the auxiliary decision control measure can be taken, and then the auxiliary decision control measure obtained by the two classification models with the highest classification prediction confidence is selected from the prediction results as the auxiliary decision control measure suggestion of the current operation mode.
7. The method according to claim 1 or 2, wherein in step S3, the extracted sample characteristic quantities include an out-of-limit element short-circuit current and its key influencing factors, a short-circuit current and its key influencing factors of an element with a larger short-circuit current that has been influenced by control measures, an expected fault safety margin and key influencing factors that have a larger influence on safety stability by control measures and candidate decision-aid control measures, and a maximum transmission power and its key influencing factors of a larger relevant section that have been influenced by control measures and candidate decision-aid control measures;
key contributors to short circuit current include: the method comprises the following steps that the total load and the total power generation output of an area where an out-of-limit element is located, the on-off state of a unit with short-circuit current sensitivity larger than a threshold value, the series high-resistance switching-on/off state of a line with short-circuit current sensitivity larger than the threshold value, the switching-on/off state of the line with short-circuit current sensitivity larger than the threshold value and a transformer, and the out-of-series or bus splitting running state of the line of a plant where the out-of;
the elements that have taken control measures to affect the short-circuit current to a large extent are: a short-circuit current element with the sensitivity of the control measure to the short-circuit current larger than a specified threshold value;
the expected faults with great influence of control measures on the safety stability are as follows: calculating equivalent impedance between a fault point of a fault i in a safety and stability check fault set and a control measure c, taking the equivalent impedance as an evaluation index of the control measure on the influence of the safety and stability of an expected fault, judging whether the index is smaller than a set threshold value, if so, considering the fault as the expected fault of which the control measure c has a larger influence on the safety and stability of the expected fault, wherein the safety and stability margin of the expected fault i before safety and stability check is calculated for an auxiliary decision;
failsafe stability critical influencing factors are envisioned including: key influencing factors of thermal stability, transient power angle stability and transient voltage stability; the key influence factors of thermal stability comprise the power flow and the switching-on/off state of a fault branch and a post-fault power flow transfer branch; the key influence factors of the transient power angle stability comprise the output and the switching-on and switching-off states of a critical group unit and the rest group units with larger participation factors, and the power flow and the switching-on and switching-off states of the contact branch of the oscillation center; the key influence factors of the transient voltage stability comprise the apparent power and the reactive voltage sensitivity of the transient voltage weak node;
the relevant sections with greater influence of the control measures are: if the expected fault associated with the maximum transmission power calculation of the section is a fault with a large influence on the safety and stability of the control measure, the section is considered to be a related section with a large influence on the control measure;
key influencing factors of the maximum transmission power of the relevant section include: the section of the related section constitutes element tide and a switching-on/off state, section related faults and safety and stability key influence factors thereof.
8. An intelligent aid decision device for limiting short circuit current operation mode is characterized by comprising:
the operation mode data acquisition module is used for acquiring operation state data of the power grid in the current operation mode;
the comprehensive calculation module is used for calculating the short-circuit current, the expected fault safety and stability check calculation and the maximum transmission power of the section based on the current running state data so as to judge whether the short-circuit current exceeds the limit and the adopted control measures exist or not;
the prediction model selection module is used for selecting a pre-constructed auxiliary decision control measure prediction model corresponding to the sample set according to the corresponding short-circuit current out-of-limit element and the adopted control measure, and ending the auxiliary decision flow if the short-circuit current out-of-limit element and the adopted control measure do not exist;
the prediction module is used for extracting characteristic quantities of the current operation mode, selecting a prediction model suitable for the current operation mode according to a preset applicability judgment rule, and obtaining an auxiliary decision control measure suggestion by using the selected auxiliary decision control measure prediction model;
the secondary comprehensive calculation module is used for calculating short-circuit current, calculating expected fault safety and stability check and calculating maximum transmission power of a section according to the current operation mode after the adjustment of the auxiliary decision control measure suggestion, and judging whether each calculation result meets the requirement or not;
the conventional auxiliary decision control measure calculation module is used for performing the conventional auxiliary decision control measure calculation based on detailed calculation to obtain the conventional auxiliary decision control measure when the calculation results of the secondary comprehensive calculation module are not completely met;
and an aid decision control measure determination module for: in response to the fact that the results of the secondary comprehensive calculation all meet the requirements, the auxiliary decision control measures obtained by the prediction model are suggested to be used as the auxiliary decision control measures of the current operation mode; in response to the fact that a short-circuit current out-of-limit element still exists in the result of the secondary comprehensive calculation, the aid decision control measure suggestion obtained by the prediction model and the conventional aid decision control measure are jointly used as the final aid decision control measure of the current operation mode; and in response to a situation where other calculation results are not satisfied, taking the conventional aid decision control measure as a final aid decision control measure;
the method for constructing the pre-constructed auxiliary decision control measure prediction model comprises the following steps:
acquiring historical short-circuit current out-of-limit event information as an initial sample set;
in response to the sample set being updated, obtaining an updated sample set;
dividing a sample set into a plurality of sample subsets according to short-circuit current out-of-limit elements and control measures which are adopted before aid decision making;
respectively extracting characteristic quantities of each sample subset with 2 or more auxiliary decision control measures in the sample subsets;
based on the extracted characteristic quantity, taking the auxiliary decision control measures included in the sample subset as classification labels, and carrying out classification prediction model training to obtain an auxiliary decision control measure prediction model of the corresponding sample subset;
only 1 sample subset of assistant decision control measures exists in the sample subsets, and the output of the corresponding prediction model is always the corresponding assistant decision control measure;
the preset applicability judgment rule is as follows:
calculating the standardized Euclidean distances between the current operation mode and the sample subset characteristic quantity of all samples in the sample subset, and if the distances between the current operation mode and a preset number k of nearest samples in the sample subset are smaller than a specified threshold value, judging that the prediction model is suitable for the current operation mode; otherwise, the prediction model is not considered to be applicable to the current operation mode.
9. The intelligent aid decision-making device for limiting short-circuit current operation mode according to claim 8, further comprising:
the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to obtain nearby operation mode samples of the current operation mode and carrying out conventional auxiliary decision control measure calculation on the nearby operation mode samples;
the sample set updating module is used for updating a training sample set of the auxiliary decision control measure prediction model by taking the current operation mode and the final auxiliary decision control measure thereof, and the nearby operation mode and the conventional auxiliary decision control measure thereof as new samples;
and the prediction model training module is used for carrying out sample subset division and model training corresponding to each sample subset based on the initial sample set or the updated training sample set to obtain a prediction model corresponding to each sample subset, and the prediction model is used for calculating follow-up auxiliary decision control measures.
10. An intelligent aid decision-making system for limiting short-circuit current operation mode is characterized by comprising:
the sample characteristic extraction module is used for extracting characteristic quantities of all samples in the sample set according to a pre-specified key characteristic quantity type;
the prediction model training module is used for carrying out prediction model training according to the classification label of each sample in the initial sample or the updated sample set and establishing a mapping relation between the sample characteristic quantity and the classification label;
the auxiliary decision control measure prediction module is used for obtaining the auxiliary decision control measure of the predicted current mode by utilizing the prediction model;
the expected fault safety and stability checking and calculating module is used for forming load flow and stability calculation data based on the power grid equipment model parameters and the state estimation result, and performing expected fault safety and stability checking and calculating of the current mode or the mode after the implementation of the control measures by combining static and transient safety and stability evaluation calculation parameters, equipment safety and stability limit values and expected fault set data to obtain expected fault thermal stability, transient power angle stability, transient voltage stability margin and key influence factors thereof;
the short-circuit current calculation module is used for performing short-circuit current checking calculation of a current mode or a mode after implementation of control measures on the basis of load flow, stable calculation data, short-circuit current calculation parameters, a bus and line short-circuit equipment set and a breaker breaking capacity limit value to obtain bus and line short-circuit currents and judging whether the short-circuit currents of the buses and the lines exceed the rated breaking capacity of the breaker and key influence factors of the short-circuit currents;
the section maximum transmission power calculation module is used for calculating the section limit power of a mode after the implementation of the current mode or the control measure based on section association expected failure safety and stability check calculation data and a section power adjustment mode to obtain the section maximum transmission power;
the conventional auxiliary decision control measure detailed calculation module is used for searching control measures meeting the short-circuit current out-of-limit requirement and recovery measures of the adopted control measures aiming at the short-circuit current out-of-limit element and the adopted short-circuit current control measures by taking the minimum influence degree on safety and stability and the minimum switch of an operation element as control targets;
the sample generation module is used for carrying out Monte Carlo sampling on the current operation mode to generate random samples near the current mode, carrying out auxiliary decision control measure calculation based on a conventional detailed calculation method on each newly generated random sample, and then storing all operation modes and calculation results thereof into a sample set;
the prediction model training module performs prediction model training and comprises the following steps:
acquiring historical short-circuit current out-of-limit event information as an initial sample set;
in response to the sample set being updated, obtaining an updated sample set;
dividing a sample set into a plurality of sample subsets according to short-circuit current out-of-limit elements and control measures which are adopted before aid decision making;
respectively extracting characteristic quantities of each sample subset with 2 or more auxiliary decision control measures in the sample subsets;
based on the extracted characteristic quantity, taking the auxiliary decision control measures included in the sample subset as classification labels, and carrying out classification prediction model training to obtain an auxiliary decision control measure prediction model of the corresponding sample subset;
only 1 auxiliary decision control measure in the sample subset exists, and the corresponding output of the prediction model is always the corresponding auxiliary decision control measure.
CN201911022581.8A 2019-10-25 2019-10-25 Intelligent aid decision-making method and system for limiting short-circuit current operation mode Active CN110635479B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911022581.8A CN110635479B (en) 2019-10-25 2019-10-25 Intelligent aid decision-making method and system for limiting short-circuit current operation mode

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911022581.8A CN110635479B (en) 2019-10-25 2019-10-25 Intelligent aid decision-making method and system for limiting short-circuit current operation mode

Publications (2)

Publication Number Publication Date
CN110635479A CN110635479A (en) 2019-12-31
CN110635479B true CN110635479B (en) 2021-04-02

Family

ID=68977697

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911022581.8A Active CN110635479B (en) 2019-10-25 2019-10-25 Intelligent aid decision-making method and system for limiting short-circuit current operation mode

Country Status (1)

Country Link
CN (1) CN110635479B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111064182B (en) * 2019-11-25 2022-06-10 国网浙江省电力有限公司湖州供电公司 Short-circuit current calculation method based on power grid characteristics
CN111262238B (en) * 2020-02-12 2021-09-03 浙江大学 Machine learning-based method for predicting short-circuit current of power distribution network containing IIDG
CN111900724B (en) * 2020-07-14 2021-11-30 国电南瑞科技股份有限公司 Online decision-making method, device and storage medium for transient stability emergency control of power system
CN112701664B (en) * 2020-11-18 2022-06-24 中国南方电网有限责任公司 Method and system for reliably issuing short-circuit current control measures
CN116739550A (en) * 2023-02-06 2023-09-12 广州辰创科技发展有限公司 Intelligent auxiliary decision-making method and system for rush repair

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101916990B (en) * 2010-08-18 2012-10-17 华北电力科学研究院有限责任公司 Optimized check method of short-circuit current limiting measure
CN102982230B (en) * 2012-10-26 2016-09-21 中国电力科学研究院 A kind of short circuit current based on node impedance sensitivity exceeds standard aid decision-making method
CN105913148A (en) * 2016-04-11 2016-08-31 国家电网公司 On-line platform-based power grid accident pre-arranged plan automatic realization method
WO2019132919A1 (en) * 2017-12-28 2019-07-04 Total Solar International A multi-agent shared machine learning approach for real-time battery operation mode prediction and control

Also Published As

Publication number Publication date
CN110635479A (en) 2019-12-31

Similar Documents

Publication Publication Date Title
CN110635479B (en) Intelligent aid decision-making method and system for limiting short-circuit current operation mode
Guo et al. Probabilistic framework for assessing the accuracy of data mining tool for online prediction of transient stability
Srivastava et al. A robust fault detection and location prediction module using support vector machine and Gaussian process regression for AC microgrid
Bai et al. A novel parameter identification approach via hybrid learning for aggregate load modeling
CN103995215B (en) A kind of smart power grid fault diagnostic method based on multi-level feedback adjustment
CN111523785A (en) Power system dynamic security assessment method based on generation countermeasure network
CN112330165B (en) Power grid transient stability evaluation method and system based on feature separation type neural network
CN111900724B (en) Online decision-making method, device and storage medium for transient stability emergency control of power system
Mazhari et al. A quantile regression-based approach for online probabilistic prediction of unstable groups of coherent generators in power systems
CN111049131B (en) Regional power grid online fault handling plan generation method and system
CN114006413A (en) Transient stability control method and system for power system based on graph neural network
CN104763576B (en) A kind of pump-storage generator protection auxiliary signal anomalous discrimination and modification method
CN111769545B (en) Load flow evaluation method and device under sequential multiple-break scene
CN111864744B (en) Online switching method and system for control modes of speed regulator of high-proportion hydroelectric system
CN107179688B (en) Power system reliability analysis method considering Monte Carlo state sampling truncation
CN110503272B (en) Short-circuit-limiting current auxiliary decision-making method and system meeting safety and stability requirements
CN110824297A (en) Single-phase earth fault discrimination method and device based on SVM (support vector machine)
CN111799773A (en) Method and system for predicting power grid cascading failure set
Moaddabi et al. General prefault transient stability estimation index using wide area phasor measurements
CN115983714A (en) Static security assessment method and system for edge graph neural network power system
CN115954956A (en) Method and system for evaluating access capacity of distributed power supply of power distribution network
Sawhney et al. On-line transient stability assessment using artificial neural network
CN113589079A (en) Testing method of self-healing system of power distribution network, electronic equipment and storage medium
Bhattacharya et al. Deep fault analysis and subset selection in solar power grids
CN112713584B (en) Control method and device suitable for out-of-set faults of expected faults of stability control system

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