CN111564848B - Intelligent power dispatching method and power utilization load prediction device for micro power grid - Google Patents
Intelligent power dispatching method and power utilization load prediction device for micro power grid Download PDFInfo
- Publication number
- CN111564848B CN111564848B CN202010519383.9A CN202010519383A CN111564848B CN 111564848 B CN111564848 B CN 111564848B CN 202010519383 A CN202010519383 A CN 202010519383A CN 111564848 B CN111564848 B CN 111564848B
- Authority
- CN
- China
- Prior art keywords
- power
- data
- time
- prediction model
- period
- 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
Links
Images
Classifications
-
- 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
- H02J3/04—Circuit arrangements for ac mains or ac distribution networks for connecting networks of the same frequency but supplied from different sources
- H02J3/06—Controlling transfer of power between connected networks; Controlling sharing of load between connected networks
-
- 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
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- 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
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
An intelligent power dispatching method and a power load forecasting device of a micro power grid are used for acquiring power utilization data of the micro power grid for a period of time; inputting the electricity utilization data of the period of time into a pre-established electricity utilization load prediction model to predict the electricity utilization condition at a future moment relative to the period of time, wherein the electricity utilization condition comprises a predicted value and a probability distribution range of the predicted value; and carrying out matching scheduling on the electric energy of the micro-grid according to the predicted electricity utilization condition at a future moment relative to the period of time.
Description
Technical Field
The invention relates to an intelligent power dispatching method and a power utilization load prediction device for a micro power grid.
Background
The micro-grid is a system which combines distributed power generation resources (such as self-powered power generation equipment or standby generator sets, solar power generation devices, wind power generation equipment and other renewable energy power generation devices) owned in a certain area or in certain enterprises and public institutions to supply power to users together, and runs in parallel with a main large-scale power grid through a power distribution network to form combined operation of the large-scale power grid and the small-scale power generation equipment. In a sense, when the distributed power reaches a certain proportion, it can be called a microgrid.
In a microgrid, the electricity usage does not remain constant all the time, for example, there are peaks and valleys in electricity usage at different times of the day, there are peaks and valleys in electricity usage at different times of the week, and so on in a month. In the peak time period of electricity utilization of a micro-grid, if the supply condition does not meet the demand of electricity utilization, a lot of problems are brought, but if the micro-grid is always in a high power supply state, a lot of unnecessary resource waste is generated. Therefore, this is a problem to be solved.
Disclosure of Invention
The application provides an intelligent power dispatching method and a power utilization load prediction device for a micro power grid.
According to a first aspect, an embodiment provides an electrical load prediction apparatus, comprising:
the sensor is used for acquiring electricity utilization data for a period of time;
the processor is used for inputting the electricity utilization data of the period of time into a pre-established electricity utilization load prediction model so as to predict the electricity utilization condition at a future moment relative to the period of time, wherein the electricity utilization condition comprises a normal value, a minimum value and a maximum value;
wherein the power load prediction model is established in the following way:
acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments;
training to obtain the power load prediction model by utilizing the training set; the method specifically comprises the following steps: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; and taking the prediction model trained by the training set as the power load prediction model. .
The power load prediction model is also established in the following way:
acquiring a test set, wherein data in the test set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments;
verifying the power load prediction model obtained by training the training set by using the test set; and performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
In one embodiment, the power usage comprises at least power.
In one embodiment, the electricity consumption data of the micro grid for a period of time comprises power, corresponding voltage, current and phase angle.
In an embodiment, the last several electricity consumption data at the same time include electricity consumption data at the same time of the last multiple days, or electricity consumption data at the same time of the last multiple weeks and the same week.
According to a second aspect, an embodiment provides a smart power scheduling method for a microgrid, including:
acquiring electricity utilization data of a period of time in the micro power grid;
inputting the electricity utilization data of the period of time into a pre-established electricity utilization load prediction model to predict the electricity utilization condition of a future moment relative to the period of time, wherein the electricity utilization condition comprises a predicted value or a predicted value and a probability distribution range thereof;
according to the predicted power utilization situation at a future moment relative to the period of time, carrying out matching scheduling on the electric energy of the micro-grid;
wherein the power load prediction model is established in the following way:
acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments;
training to obtain the power load prediction model by utilizing the training set; the method specifically comprises the following steps: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; and taking the prediction model trained by the training set as the power load prediction model.
The power load prediction model is also established in the following way:
acquiring a test set, wherein data in the test set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments;
verifying the power load prediction model obtained by training the training set by using the test set;
and performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
In an embodiment, the power consumption condition at least includes a predicted value of power, or a predicted value of power and a probability distribution range thereof.
In one embodiment, the predicted value of power includes a maximum value, a minimum value, and a normal value of power.
In one embodiment, the electricity consumption data of the micro grid for a period of time comprises power, corresponding voltage, current and phase angle.
In an embodiment, the last several electricity consumption data at the same time include electricity consumption data at the same time of the last multiple days, or electricity consumption data at the same time of the last multiple weeks and the same week.
According to a third aspect, an embodiment provides a computer-readable storage medium comprising a program executable by a processor to implement the method of any of the embodiments herein.
According to the intelligent power scheduling method, the power utilization load prediction device and the computer-readable storage medium of the micro-grid in the embodiments, the power utilization condition of the micro-grid is predicted, so that the electric energy of the micro-grid is subjected to matching scheduling.
Drawings
Fig. 1 is a schematic structural diagram of an electrical load prediction apparatus according to an embodiment;
FIG. 2 is a flowchart of an embodiment of an algorithm for a power load prediction model;
FIG. 3 is a flow chart of an algorithm for a power load prediction model according to another embodiment;
FIG. 4 is a flowchart of an intelligent power scheduling method of a microgrid according to an embodiment;
FIG. 5 is a flow diagram of an embodiment for establishing an electrical load prediction model;
fig. 6 is a flowchart of establishing a power load prediction model according to another embodiment.
Detailed Description
The present invention will be described in further detail with reference to the following detailed description and accompanying drawings. Wherein like elements in different embodiments are numbered with like associated elements. In the following description, numerous details are set forth in order to provide a better understanding of the present application. However, those skilled in the art will readily recognize that some of the features may be omitted or replaced with other elements, materials, methods in different instances. In some instances, certain operations related to the present application have not been shown or described in detail in order to avoid obscuring the core of the present application from excessive description, and it is not necessary for those skilled in the art to describe these operations in detail, so that they may be fully understood from the description in the specification and the general knowledge in the art.
Furthermore, the features, operations, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. Also, the various steps or actions in the method descriptions may be transposed or transposed in order, as will be apparent to one of ordinary skill in the art. Thus, the various sequences in the specification and drawings are for the purpose of describing certain embodiments only and are not intended to imply a required sequence unless otherwise indicated where such sequence must be followed.
The numbering of the components as such, e.g., "first", "second", etc., is used herein only to distinguish the objects as described, and does not have any sequential or technical meaning. The term "connected" and "coupled" as used herein includes both direct and indirect connections (couplings), unless otherwise specified.
This application aims at carrying out a prediction to micro-grid's power consumption load and matches, reaches intelligent electric energy scheduling's effect to save more energy, realize the low carbon environmental protection. In particular, the application considers that although a micro-grid has high or low electric loads in different time periods, however, through research, the power utilization conditions in the micro-grid are considered to be regular, and certainly, the power utilization rules of different micro-grids are different, for example, a hotel may have high power utilization load at night, low power utilization load in the daytime, the power utilization load of a factory is high in daytime and low in night, the time is expanded to one week, the power utilization load is regular, the research of the application shows that the power utilization condition at each moment is related to a period of time before and is related to the power utilization condition at the same moment of several days before, therefore, the method and the device attempt to find the relation through an algorithm, find a rule, and predict the future power utilization situation through the previous power utilization situation, so that the electric energy of the micro-grid is further matched and scheduled.
The present application is described in detail below.
In some embodiments of the present application, an electrical load prediction apparatus is disclosed. Referring to fig. 1, in some embodiments, the electrical load prediction apparatus includes a sensor 10 and a processor 30, which are described in detail below.
The sensor 10 is used to acquire electricity usage data over a period of time. The number of sensors 10 may be one or more. In some embodiments, the electrical data may include one, two, or three of power, corresponding voltage, current. In some embodiments, the electricity usage data may also include a corresponding voltage phase angle or current phase angle.
In some embodiments, the sensor 10 may include a Phasor Measurement Unit (PMU). The PMU is a phasor measurement unit configured using a GPS (Global Positioning System) second pulse as a synchronous clock. The voltage vector of each node of the power grid in the transient process is measured by the PMU, so that the PMU is applied to the fields of dynamic monitoring, state estimation, system protection, regional stability control, system analysis and prediction and the like of the power grid, and the safe operation of the power grid is guaranteed.
The method is characterized in that one or more PMUs are installed in the micro-grid to construct a dynamic monitoring system for the micro-grid, and power consumption data of the micro-grid, such as a power angle, an internal potential, a terminal three-phase fundamental wave voltage phasor, a terminal fundamental wave positive sequence voltage phasor, a terminal three-phase fundamental wave current phasor, a terminal fundamental wave positive sequence current phasor, active power, reactive power, exciting current, exciting voltage, rotor rotating speed and the like of a generator, are uploaded through the PMUs.
In some examples, the PMU may report data every 10ms, collecting data at a relatively high frequency, and thus obtaining a large amount of data for training and learning the model described below. For example, a GSA (Grid State analyzer) developed and established by the science and technology limited of Jian Ke Yun Zhi (Shenzhen) may be used, and the phase acquisition principle is similar to that of the conventional PMU, and the acquisition precision is better than that of the conventional PMU on the market at present. GSA of Jianke Yunzhi (Shenzhen) science and technology Limited company can obtain various parameters of a power grid node, such as alternating current and direct current voltage and current, alternating current frequency, global phase angle and the like, with high precision, and make initial state estimation of the online power grid node in real time. In some examples, the GSA described above further integrates the processor 30 described below, so that the current state of the power grid can be deeply analyzed, for example, the accurate state of the power grid node can be quickly and accurately obtained and the future state can be predicted through the built-in artificial intelligence edge computing capability and multi-channel big data communication function.
The above are some of the descriptions of the sensor 10.
The processor 30 is configured to predict a power consumption situation at a future time according to the power consumption data acquired by the sensor 10 over a period of time, where the power consumption situation includes a predicted value, or the predicted value and a probability distribution range thereof. For example, without taking power as an example, the processor 30 is configured to predict the power consumption situation at a future time according to the power consumption data acquired by the sensor 10 over a period of time, where the power consumption situation includes a predicted value of the power at the future time, or a predicted value of the power at the future time and a probability distribution range thereof. In some embodiments, the processor 30 is configured to predict a future time point of power consumption, such as power, a future time point of normal, minimum, and maximum, based on the power consumption data acquired by the sensor 10 over a period of time. In some embodiments, the processor 30 is configured to predict the power consumption, such as power, the normal value and its probability, the minimum value and its probability, and the maximum value and its probability at a future time based on the power consumption data acquired by the sensor 10 over a period of time. This will be explained in detail below.
In some embodiments, the processor 30 is configured to input the power consumption data of the period of time into a pre-established power load prediction model to predict power consumption conditions at a time in the future of the period of time, wherein the power consumption conditions include normal values, minimum values and maximum values. For example, the processor 30 acquires the electricity consumption data from 50 minutes to 55 minutes at 9 am of the day, and predicts the electricity consumption at 10 am, where the power is taken as an example, that is, normal value, minimum value, and maximum value of the power at 10 am are predicted. In some embodiments, the processor 30 further performs matching scheduling on the electric energy of the microgrid according to the predicted electricity utilization situation at a future time relative to the period of time.
The following describes the establishment of the electrical load prediction model, and in some embodiments, the electrical load prediction model may be established by:
(1) and acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments.
In some embodiments, the last several same-time electricity consumption data include last multiple days of same-time electricity consumption data, or last multiple weeks of same-time electricity consumption data of same week. For example, the same time in the last several days may be the same time as the future time, the same time in the last N days may be the same time in the same week in the last N weeks, and N may be set by the user.
For example, taking data a as an example of data in the training set, data a is power consumption data of a certain time T11 to T12 of a certain wednesday in the microgrid, and labels of the data are power consumption data of a time T13 of a relative time period T1 to T2 in the future, a maximum value of the power consumption data at a time T13 of the last N days (for example, saturday of the last week, sunday of the week, monday of the week, and tuesday of the week), and a minimum value of the power consumption data.
For another example, taking data B as an example of data in the training set, data B is power consumption data of a time period T21 to T22 of a certain thursday in the microgrid, and labels of the data are power consumption data of a time T23 of a future time period T21 to T22, a maximum value of the power consumption data at a time T3 of the same week of the last N weeks (for example, the thursday of the last week), and a minimum value of the power consumption data.
(2) And training to obtain the power load prediction model by using the training set. Specifically, the training can be carried out as follows: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; and taking the prediction model trained by the training set as the power load prediction model.
Specifically, please refer to fig. 2, which is a flowchart of an algorithm of an electricity load prediction model, taking electricity data as an example, a Time Series Signal (TSS) in the diagram refers to a power Signal, and a Discrete Wavelet Transform (DWT) is performed on the Time Series Signal to obtain each item of W1, W2, …, Wm; then, empirical mode decomposition is performed on the W1, W2, … and Wm respectively to obtain a plurality of sub signals (such as IMF and R in the figure), specifically, empirical mode decomposition is performed on W1 to obtain
And R1; empirical mode decomposition of W1
And Rm; analyzing each sub-signal by using Long short-term network (LSTM) to obtain respective prediction result, for example, using Long short-term network
To pair
Analyzed to obtain
Using long and short term networks
To pair
Analyzed to obtain
Using long and short term networks
Analysis of R1 gave
Using long and short term networks
To pair
Analyzed to obtain
Using long and short term networks
To pair
Analyzed to obtain
Using long and short term networks
Analyzing Rm to obtain
Then, the Prediction result of each sub-signal is used as the input of another long-short term neural network LSTM, and the final Prediction Result (PR) is obtained through training.
In other examples, referring to fig. 3, the predicted result of each sub-signal and the corresponding Additional Features (AF) may be input to the long-short term neural network LSTM for training to obtain the final predicted result. Additional features here may be currents, voltages and phase angles or phase angle differences etc. Taking the phase angle as an example: the power is divided into active power and reactive power, and the respective duty ratios are determined by the phase angle difference of the voltage and the current, so the current, the voltage and the respective phase angles can provide more detailed information for power calculation, and are also helpful for power prediction in the model.
The electrical load prediction model is obtained through training of the training set in the steps (1) and (2). In some embodiments, the model may also be corrected. For example, the electrical load prediction model is also established by:
(3) and acquiring a test set, wherein the data in the test set is the electricity utilization data of the micro power grid for a period of time, and the labels of the data are the electricity utilization data of a future moment relative to the period of time, the maximum values of the electricity utilization data of the latest plurality of same moments and the minimum values of the electricity utilization data of the latest plurality of same moments. The last meanings of the same time points have been described in detail above, and are not described herein.
(4) And using the test set to verify an electricity load prediction model obtained by training the training set.
(5) And performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
And (3), completing super-parameter tuning on the electricity load prediction model trained in the step (1) and the step (2) through the steps (3), (4) and (5).
It should be noted that the data of the training set and the test set may also be collected by the sensor 10, for example, by the PMU mentioned above. After the electricity utilization data is collected, the electricity utilization data can be cleaned, because the electricity utilization data acquired by the sensor has the problems of missing, recording errors or abnormal values (such as abnormal values generated due to short circuit and the like) and the like. And then constructing a training set based on the cleaned power utilization data, and constructing a test set.
According to the invention, through the model algorithm and the GSA developed and formulated by Jianke cloud Zhi (Shenzhen) science and technology Limited company, the prediction result has higher accuracy. In addition, as can be seen from the model algorithm and the practical effect verified by the inventor, the established power load prediction model has good performance without complicated characteristic engineering. By taking the predicted power as an example, the power consumption unit can better perform matching scheduling on the electric energy of the micro power grid by predicting the upper and lower ranges of the power consumption unit when predicting the power.
In some embodiments of the present invention, an intelligent power scheduling method for a microgrid is also disclosed, which is specifically described below.
Referring to fig. 4, an embodiment of an intelligent power scheduling method for a microgrid includes the following steps:
step 100: and acquiring power utilization data of the micro power grid for a period of time.
In some embodiments, the electrical data may include one, two, or three of power, corresponding voltage, current. In some embodiments, the electricity usage data may also include a corresponding voltage phase angle or current phase angle.
In some embodiments, step 100 may acquire power usage data via a sensor 10, such as a phasor measurement unit. Specifically, one or more PMUs are installed in the microgrid to construct a dynamic monitoring system for the microgrid, and power utilization data of the microgrid, such as a power angle, an internal potential, a terminal three-phase fundamental wave voltage phasor, a terminal fundamental wave positive sequence voltage phasor, a terminal three-phase fundamental wave current phasor, a terminal fundamental wave positive sequence current phasor, active power, reactive power, exciting current, exciting voltage, rotor speed and the like, of the generator are uploaded through the PMUs. In some examples, the PMU may report data every 10ms, collecting data at a relatively high frequency, and thus obtaining a large amount of data for training and learning the model described below. For example, a GSA (Grid State analyzer) developed and established by the science and technology limited of Jian Ke Yun Zhi (Shenzhen) may be used, and the phase acquisition principle is similar to that of the conventional PMU, and the acquisition precision is better than that of the conventional PMU on the market at present. The GSA of the science and technology Limited company of CostulZhi (Shenzhen) can obtain various parameters of a power grid node, such as alternating current and direct current voltage and current, alternating current frequency, global phase angle and the like, with high precision, and make a preliminary state estimation of the online power grid node in real time. In some examples, the GSA described above further integrates the processor 30 described below, so that the current state of the power grid can be deeply analyzed, for example, the accurate state of the power grid node can be quickly and accurately obtained and the future state can be predicted through the built-in artificial intelligence edge computing capability and multi-channel big data communication function.
Step 200: inputting the electricity utilization data of the period of time in the step 100 into a pre-established electricity utilization load prediction model to predict the electricity utilization situation at a future moment relative to the period of time, wherein the electricity utilization situation comprises a measured value or a predicted value and a probability distribution range thereof. For example, taking power as an example, step 200 inputs the power consumption data of the period of time in step 100 into a pre-established power consumption load prediction model to predict a power consumption situation at a future time relative to the period of time, where the power consumption situation includes a predicted value of power at the future time, or a predicted value of power at the future time and a probability distribution range thereof. In some embodiments, step 200 inputs the power consumption data of the period of time in step 100 into a pre-established power load prediction model to predict power consumption, e.g., power, at a future time relative to the period of time, and normal, minimum and maximum values at the future time. In some embodiments, step 200 inputs the power consumption data of the period of time in step 100 into a pre-established power load prediction model to predict power consumption, such as power, normal value and probability thereof, minimum value and probability thereof, and maximum value and probability thereof at a future time relative to the period of time.
Specifically, for example, in step 200, electricity consumption data of 50 to 55 minutes at 9 am on the day is acquired, and electricity consumption at 10 am is predicted, where power is taken as an example, that is, a normal value, a minimum value, and a maximum value of power at 10 am are predicted.
The following describes the establishment of a power load prediction model.
In some embodiments, referring to fig. 5, the power consumption load prediction model in step 200 may be established by:
step 210: and acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments.
In some embodiments, the last several same-time electricity consumption data include last multiple days of same-time electricity consumption data, or last multiple weeks of same-time electricity consumption data of same week. For example, the same time in the last several days may be the same time as the future time, the same time in the last N days may be the same time in the same week in the last N weeks, and N may be set by the user.
For example, taking data a as an example of data in the training set, data a is power consumption data of a certain time T11 to T12 of a certain wednesday in the microgrid, and labels of the data are power consumption data of a time T13 of a relative time period T1 to T2 in the future, a maximum value of the power consumption data at a time T13 of the last N days (for example, saturday of the last week, sunday of the week, monday of the week, and tuesday of the week), and a minimum value of the power consumption data.
For another example, taking data B as an example of data in the training set, data B is power consumption data of a time period T21 to T22 of a certain thursday in the microgrid, and labels of the data are power consumption data of a time T23 of a future time period T21 to T22, a maximum value of the power consumption data at a time T3 of the same week of the last N weeks (for example, the thursday of the last week), and a minimum value of the power consumption data.
Step 220: and training to obtain the power load prediction model by using the training set. Specifically, the training can be carried out as follows: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; and taking the prediction model trained by the training set as the power load prediction model.
The algorithm of the power consumption load prediction model in step 220 may refer to the description of the algorithm flow in fig. 2 and fig. 3, and is not described herein again.
In step 210 and step 220 above, the electrical load prediction model is obtained through training of the training set. In some embodiments, the model may also be corrected. In some embodiments, referring to fig. 6, the electrical load prediction model trained in steps 210 and 220 is further calibrated by:
step 230: and acquiring a test set, wherein the data in the test set is the electricity utilization data of the micro power grid for a period of time, and the labels of the data are the electricity utilization data of a future moment relative to the period of time, the maximum values of the electricity utilization data of the latest plurality of same moments and the minimum values of the electricity utilization data of the latest plurality of same moments. The last meanings of the same time points have been described in detail above, and are not described herein.
Step 240: and using the test set to verify an electricity load prediction model obtained by training the training set.
Step 250: and performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
The power load prediction model trained in the steps 210 and 220 is super-optimized through the steps 230 to 250.
It should be noted that the data of the training set and the test set may also be collected by the sensor 10, for example, by the PMU mentioned above. After the electricity utilization data is collected, the electricity utilization data can be cleaned, because the electricity utilization data acquired by the sensor has the problems of missing, recording errors or abnormal values (such as abnormal values generated due to short circuit and the like) and the like. And then constructing a training set based on the cleaned power utilization data, and constructing a test set.
Step 300: and matching and scheduling the electric energy of the micro-grid according to the predicted electricity utilization situation at a future moment relative to the period of time in step 200. For example, if the peak period of the power consumption is predicted to be reached, the micro-grid is controlled to increase the power supply, and if the valley period of the power consumption is predicted to be reached, the micro-grid is controlled to decrease the power supply.
According to the invention, through the model algorithm and the GSA developed and formulated by Jianke cloud Zhi (Shenzhen) science and technology Limited company, the prediction result has higher accuracy. In addition, as can be seen from the model algorithm and the practical effect verified by the inventor, the established power load prediction model has good performance without complicated characteristic engineering. By taking the predicted power as an example, the power consumption unit can better perform matching scheduling on the electric energy of the micro power grid by predicting the upper and lower ranges of the power consumption unit when predicting the power.
Reference is made herein to various exemplary embodiments. However, those skilled in the art will recognize that changes and modifications may be made to the exemplary embodiments without departing from the scope hereof. For example, the various operational steps, as well as the components used to perform the operational steps, may be implemented in differing ways depending upon the particular application or consideration of any number of cost functions associated with operation of the system (e.g., one or more steps may be deleted, modified or incorporated into other steps).
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. Additionally, as will be appreciated by one skilled in the art, the principles herein may be reflected in a computer program product on a computer readable storage medium, which is pre-loaded with computer readable program code. Any tangible, non-transitory computer-readable storage medium may be used, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CD-to-ROM, DVD, Blu-Ray discs, etc.), flash memory, and/or the like. These computer program instructions may be loaded onto a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions which execute on the computer or other programmable data processing apparatus create means for implementing the functions specified. 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 means for implementing the function specified. The 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.
While the principles herein have been illustrated in various embodiments, many modifications of structure, arrangement, proportions, elements, materials, and components particularly adapted to specific environments and operative requirements may be employed without departing from the principles and scope of the present disclosure. The above modifications and other changes or modifications are intended to be included within the scope of this document.
The foregoing detailed description has been described with reference to various embodiments. However, one skilled in the art will recognize that various modifications and changes may be made without departing from the scope of the present disclosure. Accordingly, the disclosure is to be considered in an illustrative and not a restrictive sense, and all such modifications are intended to be included within the scope thereof. Also, advantages, other advantages, and solutions to problems have been described above with regard to various embodiments. However, the benefits, advantages, solutions to problems, and any element(s) that may cause any element(s) to occur or become more pronounced are not to be construed as a critical, required, or essential feature or element of any or all the claims. As used herein, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, system, article, or apparatus. Furthermore, the term "coupled," and any other variation thereof, as used herein, refers to a physical connection, an electrical connection, a magnetic connection, an optical connection, a communicative connection, a functional connection, and/or any other connection.
Those skilled in the art will recognize that many changes may be made to the details of the above-described embodiments without departing from the underlying principles of the invention. Accordingly, the scope of the invention should be determined only by the claims.
Claims (6)
1. An electrical load prediction apparatus comprising:
the sensor is used for acquiring electricity utilization data for a period of time;
the processor is used for inputting the electricity utilization data of the period of time into a pre-established electricity utilization load prediction model so as to predict the electricity utilization condition at a future moment relative to the period of time, wherein the electricity utilization condition comprises a normal value, a minimum value and a maximum value;
wherein the power load prediction model is established in the following way:
acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments; the last plurality of same time points are the same time points relative to the future time point, and the method comprises the following steps: the same time of the last N days, or the same time of the same week of the last N weeks; the plurality of recent electricity consumption data at the same time comprise electricity consumption data at the same time of a plurality of days in the recent past or electricity consumption data at the same time of a plurality of weeks in the recent past and a same week;
training to obtain the power load prediction model by utilizing the training set; the method specifically comprises the following steps: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; taking the prediction model trained by the training set as the power load prediction model;
the power load prediction model is also established in the following way:
acquiring a test set, wherein data in the test set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments; the last plurality of same time points are the same time points relative to the future time point, and the method comprises the following steps: the same time of the last N days, or the same time of the same week of the last N weeks; the plurality of recent electricity consumption data at the same time comprise electricity consumption data at the same time of a plurality of days in the recent past or electricity consumption data at the same time of a plurality of weeks in the recent past and a same week;
verifying the power load prediction model obtained by training the training set by using the test set;
and performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
2. The electrical load forecasting device of claim 1, wherein the electrical condition comprises at least power.
3. The electrical load forecasting device of claim 1, wherein the electrical load data for a period of time of the microgrid comprises power, corresponding voltage, current, and phase angle.
4. An intelligent power dispatching method for a micro power grid is characterized by comprising the following steps:
acquiring electricity utilization data of a period of time in the micro power grid;
inputting the electricity utilization data of the period of time into a pre-established electricity utilization load prediction model to predict the electricity utilization condition at a future moment relative to the period of time, wherein the electricity utilization condition comprises a predicted value or a predicted value and a probability distribution range thereof; the power utilization condition comprises a predicted value of power, and the predicted value of the power comprises a maximum value, a minimum value and a normal value of the power; the power utilization condition comprises a predicted value of power and a probability distribution range thereof, and the predicted value of the power comprises a maximum value, a minimum value and a normal value of the power;
according to the predicted power utilization situation at a future moment relative to the period of time, matching and scheduling the electric energy of the micro power grid;
wherein the power load prediction model is established in the following way:
acquiring a training set, wherein data in the training set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments; the last plurality of same time points are the same time points relative to the future time point, and the method comprises the following steps: the same time of the last N days, or the same time of the same week of the last N weeks; the plurality of the latest electricity consumption data at the same time comprise electricity consumption data at the same time of the latest days or electricity consumption data at the same time of the latest weeks and the same week;
training to obtain the power load prediction model by utilizing the training set; the method specifically comprises the following steps: the method comprises the steps that a prediction model based on integrated deep learning is built in advance, input power utilization data are decomposed through an empirical mode decomposition algorithm to obtain sub-signals with different frequencies, a deep circulation neural network is used for analyzing and predicting each sub-signal, and output obtained after each sub-signal is analyzed and predicted is integrated to serve as predicted power utilization conditions; taking the prediction model trained by the training set as the power load prediction model;
the power load prediction model is also established in the following way:
acquiring a test set, wherein data in the test set is power utilization data of the micro power grid for a period of time, and the label of the data is power utilization data of a future moment relative to the period of time, the maximum value of the power utilization data of the latest plurality of same moments and the minimum value of the power utilization data of the latest plurality of same moments; the last plurality of same time points are the same time points relative to the future time point, and the method comprises the following steps: the same time of the last N days, or the same time of the same week of the last N weeks; the plurality of the latest electricity consumption data at the same time comprise electricity consumption data at the same time of the latest days or electricity consumption data at the same time of the latest weeks and the same week;
verifying the power load prediction model obtained by training the training set by using the test set;
and performing super-parameter tuning on the power load prediction model by taking the error between the label of the test concentrated data and the normal value of the power consumption situation obtained by predicting the power load prediction model as a standard so as to obtain the power load prediction model after super-parameter tuning.
5. The intelligent power scheduling method of claim 4 wherein the micro-grid's period of power usage data includes power, corresponding voltage, current and phase angle.
6. A computer-readable storage medium, characterized by comprising a program which is executable by a processor to implement the method of claim 4 or 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010519383.9A CN111564848B (en) | 2020-06-09 | 2020-06-09 | Intelligent power dispatching method and power utilization load prediction device for micro power grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010519383.9A CN111564848B (en) | 2020-06-09 | 2020-06-09 | Intelligent power dispatching method and power utilization load prediction device for micro power grid |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111564848A CN111564848A (en) | 2020-08-21 |
| CN111564848B true CN111564848B (en) | 2022-05-03 |
Family
ID=72073810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010519383.9A Active CN111564848B (en) | 2020-06-09 | 2020-06-09 | Intelligent power dispatching method and power utilization load prediction device for micro power grid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111564848B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240072A (en) * | 2021-04-14 | 2021-08-10 | 安徽工业大学 | Deep learning-based prediction method for direct-current micro-grid cluster |
| CN117977579A (en) * | 2024-03-28 | 2024-05-03 | 西华大学 | DWT-LSTM neural network-based adjustable load prediction method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102388643B (en) * | 2011-09-19 | 2013-12-04 | 华为技术有限公司 | Load forecast method, device and energy-saving control communication system |
| CN106651200A (en) * | 2016-12-29 | 2017-05-10 | 中国西电电气股份有限公司 | Electrical load management method and system for industrial enterprise aggregate user |
| CN106952181A (en) * | 2017-03-08 | 2017-07-14 | 深圳市景程信息科技有限公司 | Electric Load Prediction System based on long Memory Neural Networks in short-term |
| CN111191811A (en) * | 2018-11-14 | 2020-05-22 | 中兴通讯股份有限公司 | Cluster load prediction method and device and storage medium |
| CN110070229A (en) * | 2019-04-26 | 2019-07-30 | 中国计量大学 | The short term prediction method of home electrical load |
| CN110781947B (en) * | 2019-10-22 | 2023-04-07 | 北京交通大学 | Power load prediction model training and power load prediction method and device |
| CN111160619A (en) * | 2019-12-06 | 2020-05-15 | 北京国电通网络技术有限公司 | Power load prediction method based on data derivation |
-
2020
- 2020-06-09 CN CN202010519383.9A patent/CN111564848B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111564848A (en) | 2020-08-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sossan et al. | Achieving the dispatchability of distribution feeders through prosumers data driven forecasting and model predictive control of electrochemical storage | |
| Sossan et al. | Unsupervised disaggregation of photovoltaic production from composite power flow measurements of heterogeneous prosumers | |
| Milanović | Probabilistic stability analysis: the way forward for stability analysis of sustainable power systems | |
| Giannakis et al. | Monitoring and optimization for power grids: A signal processing perspective | |
| US8880368B2 (en) | Electric power system waveform search | |
| Abunima et al. | Reliability modeling of PV systems based on time-varying failure rates | |
| US20160043548A1 (en) | Rolling stochastic optimization based operation of distributed energy systems with energy storage systems and renewable energy resources | |
| Suslov et al. | Distributed power quality monitoring | |
| CN111564848B (en) | Intelligent power dispatching method and power utilization load prediction device for micro power grid | |
| CN108763399B (en) | Multi-source data modeling method suitable for power distribution network containing D-PMU | |
| CN111983386A (en) | Distributed state estimation system of power distribution network | |
| Zjavka | Power quality statistical predictions based on differential, deep and probabilistic learning using off‐grid and meteo data in 24‐hour horizon | |
| Negri et al. | MPC-based control for a stand-alone LVDC microgrid for rural electrification | |
| Karrari et al. | Sizing of hybrid energy storage systems using recurring daily patterns | |
| Michaelson et al. | Reduction of forced outages in islanded microgrids by compensating model uncertainties in PV rating and battery capacity | |
| Cristaldi et al. | Model‐based maximum power point tracking for photovoltaic panels: parameters identification and training database collection | |
| Angioni et al. | Bayesian distribution system state estimation in presence of non-Gaussian pseudo-measurements | |
| Zjavka | Power quality daily predictions in smart off-grids using differential, deep and statistics machine learning models processing NWP-data | |
| Liu et al. | Review on D-PMU based applications for active electricity distribution system | |
| Graditi et al. | Advanced forecasting method to the optimal management of a DC microgrid in presence of uncertain generation | |
| Dahale et al. | Recursive gaussian process over graphs for integrating multi-timescale measurements in low-observable distribution systems | |
| Cheng et al. | Dynamic parameter identification of generators for smart grid development | |
| Ponnala et al. | Effective monitoring of power system with phasor measurement unit and effective data storage system | |
| Guo et al. | Multi-time interval forecasting-aided state estimation incorporating phasor measurements for power systems with renewable energy sources | |
| Jindal et al. | Internet of Energy for Smart Cities: Machine Learning Models and Techniques |
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 |