CN117852920A - Power grid service quality situation awareness method and device based on depth model - Google Patents
Power grid service quality situation awareness method and device based on depth model Download PDFInfo
- Publication number
- CN117852920A CN117852920A CN202410043143.4A CN202410043143A CN117852920A CN 117852920 A CN117852920 A CN 117852920A CN 202410043143 A CN202410043143 A CN 202410043143A CN 117852920 A CN117852920 A CN 117852920A
- Authority
- CN
- China
- Prior art keywords
- power grid
- data set
- load
- medium
- term
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 230000009467 reduction Effects 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 27
- 230000004913 activation Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 230000008447 perception Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000003064 k means clustering Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000010438 heat treatment Methods 0.000 claims description 4
- 238000013135 deep learning Methods 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 238000005057 refrigeration Methods 0.000 claims description 2
- 238000010248 power generation Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a power grid service quality situation awareness method and device based on a depth model. The method comprises the steps of extracting weather parameters of a historical data set, performing dimension reduction analysis on the weather parameters, performing group clustering on main features of the characteristic weather parameters, and respectively establishing a mid-term load prediction model of the power system for each group; and carrying out mid-term load prediction by adopting a mid-term load prediction model, and sensing the service quality situation of the power grid based on the historical data with the maximum similarity matched with the mid-term load prediction value. According to the method, the meteorological parameters are subjected to dimension reduction and clustering, a plurality of medium-term load prediction models are established in a targeted manner, the complexity of the prediction models is reduced, and the accuracy of power grid service quality situation awareness is improved.
Description
Technical Field
The invention relates to the field of power grid service quality situation awareness, in particular to a depth model-based electric energy service quality situation awareness method and device.
Background
With the continuous development of power systems, more and more power electronic equipment and renewable energy power generation systems are connected into a power grid in a large scale, and nonlinear loads and distributed power supplies have adverse effects on the power quality of the power grid. In recent years, the use of various household appliances and precise instruments puts forward higher requirements on the quality of electric energy service, and has important significance in evaluating and predicting the quality of a power grid.
For a small-scale independent power grid in a remote area, the power generation is less interfered by the outside under the condition of stable power generation, however, the power supply capacity of the power grid is lower, the bearing fluctuation capacity is lower, and the supply and demand balance is an important evaluation index of the quality of electric energy service. For a small-scale independent power grid, meeting the supply-demand balance is an important index for evaluating the quality of electric energy service. Under the condition that the power supply capacity of the power grid is stable, the load change power service quality has a critical influence.
While most electricity prediction methods focus on short-term predictions (from minutes to 24 hours) rather than medium-term predictions (from days to months). The partial medium-long term prediction method is affected by uncontrollable factors, on one hand, the prediction result is inaccurate, on the other hand, the model is complex, the factor difference in different areas is large, and the popularization is difficult.
Disclosure of Invention
Aiming at the problems, the invention provides a power grid service quality situation sensing device and a power grid service quality situation sensing method based on a depth model, which can effectively evaluate the power service quality situation through predicting the medium-term load so as to take countermeasures and improve the power service quality. The invention is realized by the following technical contents:
the utility model provides a power grid service quality situation awareness method based on a depth model, which is characterized in that:
s1, establishing a mid-term load prediction model of a power system;
s11, acquiring a historical data set, carrying out normalization processing to obtain a first data set, extracting meteorological parameters of the first data set, and obtaining two main meteorological features through principal component analysis and dimension reduction of the meteorological parameters; the historical data set comprises meteorological parameters, holiday number, daily average day duration, daily average load value and power grid service quality grade, wherein the meteorological parameters comprise highest temperature, lowest temperature, refrigeration duration, heating duration and wind speed;
s12, clustering the first data set based on two main meteorological features; dividing the first data set into a first group, a second group and a third group by adopting a k-means clustering algorithm;
s13, respectively establishing a first medium-term load prediction model corresponding to the first group, a second medium-term load prediction model corresponding to the second group and a third medium-term load prediction model corresponding to the third group based on a neural network algorithm;
s2, acquiring a middle-stage meteorological parameter and an input data set, and acquiring a middle-stage load predicted value by adopting a middle-stage load prediction model based on the middle-stage meteorological parameter and the input data set;
s21, determining a corresponding target group based on the middle-stage weather parameters, wherein the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from three groups according to the middle-stage weather parameters;
s22, based on the input data set, adopting a medium-term load prediction model corresponding to the target group to obtain a medium-term load prediction value; the input data set includes the number of holidays, the day average day duration, of the previous week of the current time.
And S3, sensing the service quality situation of the power grid based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid.
Further, the step S3 specifically includes:
s31, acquiring a second data set consisting of a daily average load value and an meteorological parameter from the first data set, and based on a predicted value of a medium-term load, the daily average load value and the meteorological parameter corresponding to the second data set; adopting a maximum similarity algorithm, sensing a mean load value closest to the predicted value of the medium load as a first reference load, and enabling a power grid service quality grade corresponding to the mean load value closest to the predicted value of the medium load to be a first reference quality grade;
s32, judging whether the predicted value of the medium-term load is smaller than the first reference load, if so, acquiring the first reference quality grade as a perception result of the power grid service quality situation; if not, judging whether the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid;
s33, if the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid, acquiring the first reference quality grade as a perception result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
Further, the sensing of the average daily load value closest to the predicted mid-term load value by using the maximum similarity algorithm is a first reference load, which specifically includes:
s311, acquiring a first set X formed by a predicted value of a medium-term load and a medium-term meteorological parameter, and acquiring a second data set (Y1, Y2, yi, …, ym) from the first data set, wherein Yi is a daily average load value of the i week and a corresponding weather parameter form a data set, and m is the number of the data sets in the first data set; calculating the similarity Si between the first set X and the data set Yi as
S312, selecting a minimum value from S1, S2 … Si and Sm, acquiring a most similar set corresponding to the minimum value, and acquiring a corresponding power grid service quality grade from the first data set according to the most similar set.
Further, the step S13 specifically includes: obtaining a plurality of groups of training samples from the first data set, creating a deep learning neural network model comprising an input layer, an hidden layer and an output layer, wherein the hidden layer is of a three-layer fully-connected structure, training based on the training samples, and constructing a mid-term load prediction model of the power system through multiple iterations; the training samples comprise holiday number, average day duration and average day load value;
the first mid-term load prediction model activation function is:
F1＝x*tanh(ln(1+e x ))
the second mid-term load prediction model activation function is:
wherein a is a preset first parameter;
the third mid-term load prediction model activation function is:
wherein lambda is a preset second parameter.
Further, the determining the corresponding target group specifically includes: based on the middle-stage weather parameters, obtaining two middle-stage main weather features through main component analysis and dimension reduction on the middle-stage weather parameters; corresponding target groups are determined based on the two mid-term primary meteorological features.
The utility model also provides a power grid service quality situation awareness device based on the depth model, which realizes a power grid service quality situation awareness method based on the depth model, and is characterized in that: the system comprises a modeling unit, a prediction unit and a perception unit;
modeling unit: the method comprises the steps of establishing a mid-term load prediction model of the power system;
the prediction unit: based on the acquired middle weather parameters and the input data set, a middle load prediction model is adopted to obtain a middle load predicted value;
the sensing unit: based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid, the service quality situation of the power grid is perceived.
Further, the modeling unit comprises a preprocessing module, a feature extraction module, a clustering module and a modeling module;
the preprocessing module is used for: the method comprises the steps of obtaining a historical data set and carrying out normalization processing to obtain a first data set;
the feature extraction module is used for: the method comprises the steps of extracting weather parameters of a first data set, analyzing the weather parameters through principal components, and reducing the dimension to obtain two principal weather features;
the clustering module is used for: clustering the first dataset based on two primary meteorological features; dividing the first data set into a first group, a second group and a third group by adopting a k-means clustering algorithm;
the modeling module: respectively establishing a first medium-term load prediction model corresponding to the first group, a second medium-term load prediction model corresponding to the second group and a third medium-term load prediction model corresponding to the third group based on a neural network algorithm;
the modeling module comprises a sample selection module and a training module;
the sample selection module obtains a plurality of sets of training samples from the first data set; the training module carries out training based on the training samples and builds a mid-term load prediction model of the power system through multiple iterations.
Further, the prediction unit: based on the acquired middle weather parameters and the input data set, a middle load prediction model is adopted to obtain a middle load predicted value;
the prediction unit comprises a group matching module and an operation module;
the group matching module: determining a corresponding target group based on the middle-stage weather parameters, wherein the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from the three groups according to the current weather data;
the operation module is used for: and obtaining a mid-term load predicted value by adopting a mid-term load predicted model corresponding to the target group based on the input data set.
Further, the sensing unit comprises a calculation module, a comparison module and a judgment module;
the calculation module is used for calculating similarity values of a first set X formed by the predicted value of the medium-term load and the medium-term meteorological parameters and each data group of the first data set;
the comparison module is used for comparing the data set corresponding to the minimum similarity value and determining a first reference load and a first reference quality level;
the judging module determines a power grid service quality situation corresponding to the medium-term load predicted value.
Further, the judging module determines a power grid service quality situation corresponding to a predicted value of a medium load, and specifically, if the predicted value of the medium load is smaller than the first reference load, or the predicted value of the medium load is larger than the first reference load and smaller than the current power supply capacity of the power grid, the first reference quality grade is obtained as a sensing result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
The beneficial effects of the invention are as follows: the invention uses the meteorological parameters and the condition clustering, obtains two main component parameters by reducing the dimension of the meteorological parameters, and reduces modeling parameters; aiming at the weather parameter clustering, a plurality of simpler and more accurate models are established to correspond to the weather parameters, so that the construction difficulty of the models and the prediction accuracy can be reduced. Based on the similarity calculation of the predicted medium-term load and the historical data, the electric energy service quality when the weather parameters and the load values are similar is obtained, meanwhile, the relation between the load and the power supply capacity of the power grid is considered, and the service quality situation of the power grid is accurately perceived.
Drawings
FIG. 1 is a flow chart of a power grid service quality situation awareness method based on a depth model;
FIG. 2 is a flow chart of the present invention for building a mid-load prediction model of an electrical power system;
FIG. 3 is a flow chart of the present invention for deriving a mid-term load prediction value using a mid-term load prediction model;
FIG. 4 is a flow chart of the perceived grid quality of service situation of the present invention;
fig. 5 is a block diagram of a depth model-based power grid service quality situation awareness apparatus according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Embodiment one:
the invention provides a power grid service quality situation awareness method based on a depth model. A depth model-based power grid service quality situation awareness method flow chart is shown in fig. 1. The method comprises the following steps:
s1, establishing a mid-term load prediction model of the power system.
The step S1 specifically comprises the following steps: s11, acquiring a historical data set, carrying out normalization processing to obtain a first data set, extracting meteorological parameters of the first data set, and obtaining two main meteorological features through principal component analysis and dimension reduction of the meteorological parameters.
The historical data set comprises meteorological parameters, holiday number, day time length, day load value and power grid service quality level. The historical data set is provided with a plurality of data groups, and various parameters of each week in the historical period can be embodied. Each set of data represents one week of meteorological parameters, holiday number, day time period, day load value, and grid quality of service level.
The meteorological parameters comprise the highest temperature, the lowest temperature, the refrigerating time, the heating time, the wind speed and the like. The refrigerating time period is a time period which is lower than a comfortable temperature range, the heating time period is a time period which is higher than the comfortable temperature range, and the comfortable temperature range is 18-24 ℃.
S12, clustering the first data set based on two main meteorological features; the first dataset is partitioned into a first group, a second group, and a third group using a k-means clustering algorithm.
S13, respectively establishing a first medium-term load prediction model corresponding to the first group, a second medium-term load prediction model corresponding to the second group and a third medium-term load prediction model corresponding to the third group based on a neural network algorithm;
the step S13 specifically includes: obtaining a plurality of groups of training samples from the first data set, creating a deep learning neural network model comprising an input layer, an hidden layer and an output layer, wherein the hidden layer is of a three-layer fully-connected structure, training based on the training samples, and constructing a mid-term load prediction model of the power system through multiple iterations; the training samples comprise holiday number, average day duration and average day load value;
the first mid-term load prediction model activation function is:
F1＝x*tanh(ln(1+e x ))；
the second mid-term load prediction model activation function is:
wherein a is a preset first parameter;
the third mid-term load prediction model activation function is:
wherein lambda is a preset second parameter;
s2, acquiring a middle-stage meteorological parameter and an input data set, and acquiring a middle-stage load predicted value by adopting a middle-stage load prediction model based on the middle-stage meteorological parameter and the input data set.
The middle-term weather parameters are predicted weather parameters corresponding to the time to be predicted, and the predicted weather parameters can be obtained from a right-of-way weather website or each weather center corresponding to the national weather office.
The step S2 specifically comprises the following steps: s21, determining a corresponding target group based on the middle-stage weather parameters, wherein the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from the three groups according to the middle-stage weather parameters.
S22, based on the input data set, adopting a medium-term load prediction model corresponding to the target group to obtain a medium-term load prediction value; the input data set comprises the number of holidays and the day average daytime duration of the week before the current time; the mid-term may be two weeks, three weeks or one month, i.e. the mid-term load prediction model is able to predict the load value one month in advance.
The determining the corresponding target group specifically includes: based on the middle-stage weather parameters, obtaining two middle-stage main weather features through main component analysis and dimension reduction on the middle-stage weather parameters; corresponding target groups are determined based on the two mid-term primary meteorological features.
And S3, sensing the service quality situation of the power grid based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid. The power supply capacity of the power grid is the peak power provided by the current power grid. Because the power supply capacity of the power grid mainly uses traditional energy, the power generation facility can determine the power supply capacity of the middle and short periods according to the construction plan. When the medium-short period power supply capacity is not changed greatly, the current power supply capacity of the power grid can be used as a reference factor for evaluating the supply-demand relationship.
Because the predicted value of the medium-term load directly reflects the future load demand, and under the condition that the power supply capacity of the power grid is stable, the meteorological parameters have great influence on the load demand, records similar to the medium-term meteorological parameters and the load in the historical data are obtained, and the electric energy service quality grade corresponding to the records can be obtained and is similar to the predicted medium-term electric energy service quality grade.
The step S3 specifically comprises the following steps:
s31, acquiring a second data set consisting of a daily average load value and an meteorological parameter from the first data set. The second data is the same as the first data set in number, and the difference is that the number of parameters in each data set is different, each data set of the second data set only comprises two parameters of a daily average load value and an meteorological parameter, and the first data set comprises five parameters of a meteorological parameter, a holiday number, a daily average daytime duration, a daily average load value and a power grid service quality grade.
Based on the mid-term load forecast value, the mid-term weather parameter, and the daily average load value and weather parameter corresponding to the second data set; and sensing a mean load value closest to the predicted value of the medium load as a first reference load by adopting a maximum similarity algorithm, wherein the power grid service quality grade corresponding to the mean load value closest to the predicted value of the medium load is a first reference quality grade.
The adoption of the maximum similarity algorithm senses that a daily average load value closest to the predicted value of the medium-term load is a first reference load, and specifically comprises the following steps:
s311, acquiring a first set X formed by a predicted value of a medium-term load and a medium-term meteorological parameter, and acquiring a second data set (Y1, Y2, yi, …, ym) from the first data set, wherein Yi is a daily average load value of the i week and a corresponding weather parameter form a data set, and m is the number of the data sets in the first data set; the similarity between the first set X and the data set Yi is that
Wherein a smaller Si indicates that the first set is more similar to the second set.
S312, selecting a minimum value from S1, S2 … Si and Sm, acquiring a most similar set corresponding to the minimum value, and acquiring a corresponding power grid service quality grade from the first data set according to the most similar set.
S32, judging whether the predicted value of the medium-term load is smaller than the first reference load, if so, acquiring the first reference quality grade as a perception result of the power grid service quality situation; if not, judging whether the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid.
S33, if the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid, acquiring the first reference quality grade as a perception result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
Embodiment two:
the invention also provides a power grid service quality situation awareness device based on the depth model, which can realize a power grid service quality situation awareness method based on the depth model. A structural frame diagram of the depth model-based power grid quality of service situation awareness device is shown in fig. 5.
A power grid service quality situation sensing device based on a depth model comprises a modeling unit, a prediction unit and a sensing unit.
Modeling unit: the method is used for building a mid-term load prediction model of the power system. The modeling unit comprises a preprocessing module, a feature extraction module, a clustering module and a modeling module.
The preprocessing module is used for: the method comprises the steps of obtaining a historical data set and carrying out normalization processing to obtain a first data set; the feature extraction module is used for: the method comprises the steps of extracting weather parameters of a first data set, analyzing the weather parameters through principal components, and reducing the dimension to obtain two principal weather features; the clustering module is used for: clustering the first dataset based on two primary meteorological features; dividing the first data set into a first group, a second group and a third group by adopting a k-means clustering algorithm; the modeling module: and respectively establishing a first mid-term load prediction model corresponding to the first group, a second mid-term load prediction model corresponding to the second group and a third mid-term load prediction model corresponding to the third group based on a neural network algorithm.
The modeling module comprises a sample selection module and a training module; the sample selection module obtains a plurality of sets of training samples from the first data set; the training module carries out training based on the training samples and builds a mid-term load prediction model of the power system through multiple iterations.
The prediction unit: based on the acquired mid-term meteorological parameters and the input data set, a mid-term load prediction model is adopted to obtain a mid-term load predicted value. The prediction unit comprises a group matching module and an operation module.
The group matching module: determining a corresponding target group based on the middle-stage weather parameters, wherein the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from the three groups according to the current weather data; the operation module is used for: and obtaining a mid-term load predicted value by adopting a mid-term load predicted model corresponding to the target group based on the input data set.
The sensing unit: based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid, the service quality situation of the power grid is perceived. The sensing unit comprises a calculation module, a comparison module and a judgment module.
The calculation module is used for calculating similarity values of a first set X formed by the predicted value of the medium-term load and the medium-term meteorological parameters and each data group of the second data set respectively; the comparison module is used for comparing the data set corresponding to the minimum similarity value and determining a first reference load and a first reference quality level; the judging module determines a power grid service quality situation corresponding to the medium-term load predicted value.
The process of determining the power grid service quality situation corresponding to the medium-term load predicted value by the judging module specifically comprises the following steps: if the predicted value of the medium load is smaller than the first reference load or the predicted value of the medium load is larger than the first reference load and smaller than the current power supply capacity of the power grid, acquiring the first reference quality grade as a sensing result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of one of ordinary skill in the art without departing from the spirit of the present invention. Furthermore, the non-illustrated portions referred to in the present invention are the same as or implemented using the prior art.
Claims (10)
1. A power grid service quality situation awareness method based on a depth model is characterized by comprising the following steps of:
s1, establishing a mid-term load prediction model of a power system;
s11, acquiring a historical data set, carrying out normalization processing to obtain a first data set, extracting meteorological parameters of the first data set, and obtaining two main meteorological features through principal component analysis and dimension reduction of the meteorological parameters; the historical data set comprises meteorological parameters, holiday number, daily average day duration, daily average load value and power grid service quality grade, wherein the meteorological parameters comprise highest temperature, lowest temperature, refrigeration duration, heating duration and wind speed;
s12, clustering the first data set based on two main meteorological features; dividing the first data set into a first group, a second group and a third group by adopting a k-means clustering algorithm;
s13, respectively establishing a first medium-term load prediction model corresponding to the first group, a second medium-term load prediction model corresponding to the second group and a third medium-term load prediction model corresponding to the third group based on a neural network algorithm;
s2, acquiring a middle-stage meteorological parameter and an input data set, and acquiring a middle-stage load predicted value by adopting a middle-stage load prediction model based on the middle-stage meteorological parameter and the input data set;
s21, determining a corresponding target group based on the middle-term meteorological parameters; the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from three groups according to the middle-stage weather parameters;
s22, based on the input data set, adopting a medium-term load prediction model corresponding to the target group to obtain a medium-term load prediction value; the input data set comprises the number of holidays and the day average daytime duration of the week before the current time;
and S3, sensing the service quality situation of the power grid based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid.
2. The depth model-based power grid quality of service situation awareness method of claim 1, wherein: the step S3 specifically includes:
s31, acquiring a second data set consisting of a daily average load value and an meteorological parameter from the first data set, and based on a predicted value of a medium-term load, the daily average load value and the meteorological parameter corresponding to the second data set; adopting a maximum similarity algorithm, sensing a mean load value closest to the predicted value of the medium load as a first reference load, and enabling a power grid service quality grade corresponding to the mean load value closest to the predicted value of the medium load to be a first reference quality grade;
s32, judging whether the predicted value of the medium-term load is smaller than the first reference load, if so, acquiring the first reference quality grade as a perception result of the power grid service quality situation; if not, judging whether the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid;
s33, if the predicted value of the medium-term load is smaller than the current power supply capacity of the power grid, acquiring the first reference quality grade as a perception result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
3. The depth model-based power grid quality of service situation awareness method of claim 2, wherein: the adoption of the maximum similarity algorithm senses that a daily average load value closest to the predicted value of the medium-term load is a first reference load, and specifically comprises the following steps:
s311, acquiring a first set X formed by a predicted value of a medium-term load and a medium-term meteorological parameter, and acquiring a second data set (Y1, Y2, yi, …, ym) from the first data set, wherein Yi is a daily average load value of the i week and a corresponding weather parameter form a data set, and m is the number of the data sets in the first data set; calculating the similarity Si between the first set X and the data set Yi as
S312, selecting a minimum value from S1, S2 … Si and Sm, acquiring a most similar set corresponding to the minimum value, and acquiring a corresponding power grid service quality grade from the first data set according to the most similar set.
4. A depth model based power grid quality of service situation awareness method according to claim 3, characterized in that: the step S13 specifically includes: obtaining a plurality of groups of training samples from the first data set, creating a deep learning neural network model comprising an input layer, an hidden layer and an output layer, wherein the hidden layer is of a three-layer fully-connected structure, training based on the training samples, and constructing a mid-term load prediction model of the power system through multiple iterations; the training samples comprise holiday number, average day duration and average day load value;
the first mid-term load prediction model activation function is:
F1＝x*tanh(ln(1+e x ))
the second mid-term load prediction model activation function is:
wherein a is a preset first parameter;
the third mid-term load prediction model activation function is:
wherein lambda is a preset second parameter.
5. The depth model-based power grid quality of service situation awareness method of claim 4, wherein: the determining the corresponding target group specifically includes: based on the middle-stage weather parameters, obtaining two middle-stage main weather features through main component analysis and dimension reduction on the middle-stage weather parameters; corresponding target groups are determined based on the two mid-term primary meteorological features.
6. A depth model-based power grid service quality situation awareness apparatus for implementing the depth model-based power grid service quality situation awareness method according to any one of claims 1 to 5, wherein: the system comprises a modeling unit, a prediction unit and a perception unit;
modeling unit: the method comprises the steps of establishing a mid-term load prediction model of the power system;
the prediction unit: based on the acquired middle weather parameters and the input data set, a middle load prediction model is adopted to obtain a middle load predicted value;
the sensing unit: based on the predicted value of the medium-term load, the medium-term meteorological parameters and the current power supply capacity of the power grid, the service quality situation of the power grid is perceived.
7. The depth model-based power grid quality of service situation awareness apparatus of claim 6, wherein:
the modeling unit comprises a preprocessing module, a feature extraction module, a clustering module and a modeling module;
the preprocessing module is used for: the method comprises the steps of obtaining a historical data set and carrying out normalization processing to obtain a first data set;
the feature extraction module is used for: the method comprises the steps of extracting weather parameters of a first data set, analyzing the weather parameters through principal components, and reducing the dimension to obtain two principal weather features;
the clustering module is used for: clustering the first dataset based on two primary meteorological features; dividing the first data set into a first group, a second group and a third group by adopting a k-means clustering algorithm;
the modeling module: respectively establishing a first medium-term load prediction model corresponding to the first group, a second medium-term load prediction model corresponding to the second group and a third medium-term load prediction model corresponding to the third group based on a neural network algorithm;
the modeling module comprises a sample selection module and a training module;
the sample selection module obtains a plurality of sets of training samples from the first data set; the training module carries out training based on the training samples and builds a mid-term load prediction model of the power system through multiple iterations.
8. The depth model-based power grid quality of service situation awareness apparatus of claim 7, wherein:
the prediction unit: based on the acquired middle weather parameters and the input data set, a middle load prediction model is adopted to obtain a middle load predicted value;
the prediction unit comprises a group matching module and an operation module;
the group matching module: determining a corresponding target group based on the middle-stage weather parameters, wherein the middle-stage weather parameters are weather parameters according to a date to be predicted, and the target group is a group matched from the three groups according to the current weather data;
the operation module is used for: and obtaining a mid-term load predicted value by adopting a mid-term load predicted model corresponding to the target group based on the input data set.
9. The depth model-based power grid quality of service situation awareness apparatus of claim 8, wherein:
the sensing unit comprises a calculation module, a comparison module and a judgment module;
the calculation module is used for calculating similarity values of a first set X formed by the predicted value of the medium-term load and the medium-term meteorological parameters and each data group of the second data set respectively;
the comparison module is used for comparing the similarity values to obtain minimum similarity values, and determining a first reference load and a first reference quality level according to a data set corresponding to the minimum similarity values;
the judging module determines a power grid service quality situation corresponding to the medium-term load predicted value.
10. The depth model-based power grid quality of service situation awareness apparatus of claim 9, wherein:
the judging module determines a power grid service quality situation corresponding to a predicted value of a medium load, and specifically, if the predicted value of the medium load is smaller than the first reference load or the predicted value of the medium load is larger than the first reference load and smaller than the current power grid power supply capacity, the first reference quality grade is obtained as a sensing result of the power grid service quality situation; otherwise, an alarm is sent out to prompt the power grid service quality situation risk.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410043143.4A CN117852920A (en) | 2024-01-11 | 2024-01-11 | Power grid service quality situation awareness method and device based on depth model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410043143.4A CN117852920A (en) | 2024-01-11 | 2024-01-11 | Power grid service quality situation awareness method and device based on depth model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117852920A true CN117852920A (en) | 2024-04-09 |
Family
ID=90530070
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410043143.4A Pending CN117852920A (en) | 2024-01-11 | 2024-01-11 | Power grid service quality situation awareness method and device based on depth model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117852920A (en) |
-
2024
- 2024-01-11 CN CN202410043143.4A patent/CN117852920A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110610280B (en) | Short-term prediction method, model, device and system for power load | |
| CN106408223A (en) | Short-term load prediction based on meteorological similar day and error correction | |
| CN108921339B (en) | Quantile regression-based photovoltaic power interval prediction method for genetic support vector machine | |
| CN110796307B (en) | Distributed load prediction method and system for comprehensive energy system | |
| CN113205207A (en) | XGboost algorithm-based short-term power consumption load fluctuation prediction method and system | |
| CN110380444B (en) | Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula | |
| CN109687447B (en) | Electric power energy consumption prediction method and device | |
| CN110210670A (en) | A kind of prediction technique based on power-system short-term load | |
| CN112149890A (en) | Comprehensive energy load prediction method and system based on user energy label | |
| CN112365056A (en) | Electrical load joint prediction method and device, terminal and storage medium | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| CN115965132A (en) | Distributed photovoltaic digital twin system power prediction method based on GA-BP neural network | |
| CN112633556A (en) | Short-term power load prediction method based on hybrid model | |
| CN115545333A (en) | Method for predicting load curve of multi-load daily-type power distribution network | |
| CN115099511A (en) | Photovoltaic power probability estimation method and system based on optimized copula | |
| CN116826745B (en) | Layered and partitioned short-term load prediction method and system in power system background | |
| CN112101673A (en) | Power grid development trend prediction method and system based on hidden Markov model | |
| CN117852920A (en) | Power grid service quality situation awareness method and device based on depth model | |
| CN115907228A (en) | Short-term power load prediction analysis method based on PSO-LSSVM | |
| CN109034453A (en) | A kind of Short-Term Load Forecasting Method based on multiple labeling neural network | |
| CN117236532B (en) | Load data-based electricity consumption peak load prediction method and system | |
| Kartini et al. | Development hybrid model deep learning neural network (dl-nn) for probabilistic forecasting solar irradiance on solar cells to improve economics value added | |
| CN117595231B (en) | Intelligent power grid distribution management system and method thereof | |
| CN113537575B (en) | Trend load prediction method containing distributed photovoltaic and electric automobile grid connection | |
| CN117318051B (en) | Intelligent energy storage management and control system and method based on Internet of things technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |