CN113095562A - Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM - Google Patents

Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM Download PDF

Info

Publication number
CN113095562A
CN113095562A CN202110370624.2A CN202110370624A CN113095562A CN 113095562 A CN113095562 A CN 113095562A CN 202110370624 A CN202110370624 A CN 202110370624A CN 113095562 A CN113095562 A CN 113095562A
Authority
CN
China
Prior art keywords
data
historical
meteorological data
lstm
inputting
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
Application number
CN202110370624.2A
Other languages
Chinese (zh)
Inventor
凌晨光
刘军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Anhui Tianneng Cleaning Energy Technology Co ltd
Original Assignee
Anhui Tianneng Cleaning Energy Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Anhui Tianneng Cleaning Energy Technology Co ltd filed Critical Anhui Tianneng Cleaning Energy Technology Co ltd
Priority to CN202110370624.2A priority Critical patent/CN113095562A/en
Publication of CN113095562A publication Critical patent/CN113095562A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21—Design, administration or maintenance of databases
    • G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Evolutionary Computation (AREA)
  • Quality & Reliability (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Operations Research (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Biophysics (AREA)
  • Game Theory and Decision Science (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Development Economics (AREA)
  • Public Health (AREA)
  • Computing Systems (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Molecular Biology (AREA)

Abstract

The invention provides a method and a device for ultra-short term power generation prediction based on Kalman filtering and LSTM, wherein the method comprises the following steps: acquiring historical measured meteorological data, historical predicted meteorological data, historical power data and weather forecast data at a first preset time in the future; carrying out data cleaning on the data, and carrying out data normalization; inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, and inputting weather forecast data at a first preset time in the future after the model training is finished to obtain 16-step Kalman correction predicted meteorological data; inputting the processed historical meteorological data and historical power data into an LSTM model for training; and loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result, and performing error analysis. The method and the device greatly improve the prediction precision and meet the assessment requirements of the power system.

Description

Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM
Technical Field
The invention belongs to the field of photovoltaic power generation, and particularly relates to a method and a device for ultra-short term power generation prediction based on Kalman filtering and LSTM.
Background
At present, the traditional photovoltaic power prediction algorithm such as a support vector machine method, a regression analysis method and the like performs combined prediction on collected meteorological data and historical power data of a photovoltaic power station, the photovoltaic power prediction is greatly influenced by irradiance and cloud amount, the accuracy of the meteorological data directly influences the accuracy of a prediction result, once rainy weather occurs, the accuracy of the prediction result is poor, and the requirement of power assessment cannot be met.
Disclosure of Invention
The embodiment of the application provides a Kalman filtering and LSTM-based power generation ultra-short term prediction method and device, so that the prediction precision is greatly improved, and the assessment requirement of a power system is met.
In a first aspect, an embodiment of the present application provides a method for ultra-short term power generation prediction based on kalman filtering and LSTM, including:
acquiring historical measured meteorological data, historical predicted meteorological data, historical power data and weather forecast data at a first preset time in the future;
carrying out data cleaning on the data, and carrying out data normalization;
inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, and inputting weather forecast data at a first preset time in the future after the model training is finished to obtain 16-step Kalman correction predicted meteorological data;
inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a target error threshold, if the error is smaller than the target error threshold, saving the model, otherwise, continuing training until the target error threshold is reached;
loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result, and performing error analysis;
and performing inverse normalization processing on the prediction result subjected to error analysis finally, and storing the result.
Inputting historical actual measurement meteorological data and historical prediction meteorological data into a Kalman filter for model training, inputting weather forecast data of a first preset time in the future after the model training is finished, and obtaining 16-step Kalman correction prediction meteorological data, wherein the method comprises the following steps of:
the system matrix adopts an identity matrix InThe system equation and the observation equation are expressed as follows:
xt(ti+1)＝xt(ti)+η(ti)
yi O＝Hi[xt(ti)]+εi
system covariance matrix Q (t)i) And an observation matrix R (t)i) Based on η (t)i)＝xt(ti+1)-xt(ti) And
Figure BDA0003009159630000021
the last 7 values of (a), are expressed as follows:
Figure BDA0003009159630000022
Figure BDA0003009159630000023
correcting multi-step meteorological data principle into known observation matrix H based on Kalman filteringiBy the above formula to ti+1Updating the time observation vector
Figure BDA0003009159630000024
Wherein the first preset time is four hours.
Wherein, the data cleaning of the data comprises: and missing value filling and abnormal value processing are carried out on the historical measured meteorological data, the historical predicted meteorological data, the historical power data and the weather forecast data at the first preset time in the future.
In a second aspect, the present application provides a kalman filter and LSTM-based power generation ultra-short term prediction apparatus, including:
the acquiring unit is used for acquiring historical measured meteorological data, historical predicted meteorological data, historical power data and weather forecast data at a first preset time in the future;
the data cleaning unit is used for cleaning the data and normalizing the data;
the first training unit is used for inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, inputting weather forecast data at a first preset time in the future after the model training is finished, and obtaining 16-step Kalman correction predicted meteorological data;
the second training unit is used for inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a target error threshold value, if the error is smaller than the target error threshold value, saving the model, otherwise, continuing the training until the target error threshold value is reached;
the prediction unit is used for loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result and carrying out error analysis;
and the inverse normalization processing unit is used for performing inverse normalization processing on the prediction result subjected to the error analysis finally and storing the result.
Wherein the first training unit is to:
the system matrix adopts an identity matrix InThe system equation and the observation equation are expressed as follows:
xt(ti+1)＝xt(ti)+η(ti)
yi O＝Hi[xt(ti)]+εi
system covariance matrix Q (t)i) And an observation matrix R (t)i) Based on η (t)i)＝xt(ti+1)-xt(ti) And
Figure BDA0003009159630000031
the last 7 values of (a), are expressed as follows:
Figure BDA0003009159630000032
Figure BDA0003009159630000033
correcting multi-step meteorological data principle into known observation matrix H based on Kalman filteringiBy the above formula to ti+1Updating the time observation vector
Figure BDA0003009159630000041
Wherein the first preset time is four hours.
Wherein the data cleansing unit is configured to: and missing value filling and abnormal value processing are carried out on the historical measured meteorological data, the historical predicted meteorological data, the historical power data and the weather forecast data at the first preset time in the future.
In a third aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program is used for implementing the steps of any one of the above methods when executed by a processor.
In a fourth aspect, the present application provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of any one of the above methods when executing the program.
The method and the device for ultra-short term power generation prediction based on Kalman filtering and LSTM have the following beneficial effects:
according to the method, the Kalman filtering is combined with the LSTM model to perform photovoltaic power ultra-short-term prediction, the influence of inaccurate meteorological data on prediction in a traditional prediction mode is reduced, the prediction precision is greatly improved, the assessment requirement of a power system is met, the operation management efficiency of a photovoltaic power station is improved, the power grid scheduling is optimized, and certain economic loss is avoided.
Drawings
FIG. 1 is a schematic flow chart of a Kalman filtering and LSTM-based power generation ultra-short-term prediction method according to an embodiment of the present application;
FIG. 2 is a schematic flow chart illustrating another method for ultra-short term prediction of power generation based on Kalman filtering and LSTM in accordance with an embodiment of the present invention;
FIG. 3 is a diagram of the structure of an LSTM neuron;
FIG. 4 is a schematic structural diagram of an ultra-short term prediction device for power generation based on Kalman filtering and LSTM in an embodiment of the present application;
fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
The present application is further described with reference to the following figures and examples.
In the following description, the terms "first" and "second" are used for descriptive purposes only and are not intended to indicate or imply relative importance. The following description provides embodiments of the invention, which may be combined or substituted for various embodiments, and this application is therefore intended to cover all possible combinations of the same and/or different embodiments described. Thus, if one embodiment includes feature A, B, C and another embodiment includes feature B, D, then this application should also be considered to include an embodiment that includes one or more of all other possible combinations of A, B, C, D, even though this embodiment may not be explicitly recited in text below.
The following description provides examples, and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made in the function and arrangement of elements described without departing from the scope of the disclosure. Various examples may omit, substitute, or add various procedures or components as appropriate. For example, the described methods may be performed in an order different than the order described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.
The photovoltaic power generation has very obvious random characteristics, and with the increasing of photovoltaic power generation capacity, the influence of grid-connected operation on the power system is more and more obvious, but the precision of the traditional prediction model is difficult to meet the requirement of power assessment, so that power generation enterprises are assessed and penalized to cause economic loss, and the safe operation and effective scheduling of the power system are difficult to ensure.
Based on this, as shown in fig. 1-3, the present application provides a method for ultra-short term prediction of power generation based on kalman filtering and LSTM, comprising: s101, acquiring historical actual measurement meteorological data, historical prediction meteorological data, historical power data and weather forecast data at a first preset time in the future; s103, performing data cleaning on the data, and performing data normalization; s105, inputting historical actual measurement meteorological data and historical prediction meteorological data into a Kalman filter for model training, and inputting weather forecast data at a first preset time in the future after the model training is finished to obtain 16-step Kalman correction prediction meteorological data; s107, inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a target error threshold, if the error is smaller than the target error threshold, saving the model, otherwise, continuing training until the target error threshold is reached; s109, loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result, and performing error analysis; and S111, performing inverse normalization processing on the prediction result subjected to the error analysis finally, and storing the result. As described in detail below.
Kalman filtering is a computational method that corrects by feeding back the error between the observed value at the previous time and the actual value to the observed value at the future time. Based on the research results of Galanis, Louka and the like, the Kalman filtering state equation and the observation equation are described as follows:
xt(ti+1)＝Mi[xt(ti)]+η(ti)
Figure BDA0003009159630000061
in the formula:
xt(ti) Unknown Process at time tiTrue state of (3);
Figure BDA0003009159630000062
-an observation vector known at time t;
η(ti) The system noise is defined as white Gaussian noise and is independent of each other, and the covariance matrix corresponding to the system noise is Q (t)i)；
εi-the observed noise, defined as white gaussian noise and independent of each other, has a covariance matrix R (t) corresponding theretoi)；
Mi-a system matrix;
Hi-an observation matrix.
Kalman filtering provides a method based on all observations yORecursive estimation of unknown states xtUpdate to tiThe method of time, with respect to the prediction step of the state vector and its error covariance matrix, is based on the previous time step analysis values as follows:
xf(ti)＝Mi-1[xa(ti-1)]
Figure BDA0003009159630000071
in the formula:
xa(ti-1)——ti-1a time state analysis value;
xf(ti)——tipredicting a state at a moment;
Pf(ti)——tierror covariance matrix prediction;
Pa(ti-1)——ti-1and analyzing the value of the time error covariance matrix.
The following is the update (analysis) step, in which t is comparediThe time observation is mixed with the past information:
Figure BDA0003009159630000072
Pa(ti)＝(I-KiHi)Pf(ti)
wherein
Figure BDA0003009159630000073
KiIs the kalman gain, which describes the index that adjusts the filter according to the possible new states.
Suppose miFor the NWP model at tiThe weather data, such as irradiance data,
Figure BDA0003009159630000074
for predicting deviations, using information about miThe polynomial of (b) represents:
Figure BDA0003009159630000075
in the formula:
a(j,i)(j-0, 1.., n-1) — coefficients estimated using a kalman filter.
With a(j,i)The matrix forms a matrix of coefficients as a state vector, i.e. x (t)i)＝[a0,i a1,i a2,i … an-1,i]TAt the same time adopt
Figure BDA0003009159630000076
As an observation vector, the observation matrix is
Figure BDA0003009159630000077
The order n is taken as 3.
The system matrix adopts an identity matrix InBased on the above, the system equation and the observation equation can be expressed as follows:
xt(ti+1)＝xt(ti)+η(ti)
Figure BDA0003009159630000081
system covariance matrix Q (t)i) And an observation matrix R (t)i) Based on η (t)i)＝xt(ti+1)-xt(ti) And
Figure BDA0003009159630000082
the last 7 values of (a), are expressed as follows:
Figure BDA0003009159630000083
Figure BDA0003009159630000084
correcting multi-step meteorological data principle into known observation matrix H based on Kalman filteringiBy the above formula to ti+1Updating the time observation vector
Figure BDA0003009159630000085
LSTM (Long-Short Term Memory network) is one of Recurrent Neural Networks (RNN), and is characterized by having Memory and forgetting functions, and thus is commonly used for speech recognition, handwritten text prediction, and the like. The LSTM algorithm not only solves the problem of gradient explosion and gradient disappearance of the RNN network, but also can fully utilize historical information. And the photovoltaic power generation power has obvious dependence relation with time, and the method is very suitable for prediction by adopting an LSTM algorithm. The hidden layer of the LSTM network has unique neural elements, including an Input gate (Input gate), a forgetting gate (forkgate), and an Output gate (Output gate).
Meteorological data are corrected through Kalman filtering and used as input of an LSTM model, prediction precision can be effectively improved, and the specific prediction steps are as follows:
1. acquiring historical actual measurement meteorological data, historical prediction meteorological data, historical power data and weather forecast data in four hours in the future.
2. And performing data cleaning on the data, including missing value filling, abnormal value processing and the like, and performing data normalization.
3. Inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, and inputting weather forecast data four hours in the future after the model training is finished to obtain 16-step Kalman correction predicted meteorological data.
4. Inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a minimum error threshold, if the error is smaller than the threshold, saving the model, otherwise, continuing training until the minimum error threshold is reached.
5. And loading an LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result, and performing error analysis.
6. And performing inverse normalization processing on the prediction result subjected to error analysis finally, and storing the result.
According to the method, the Kalman filtering is combined with the LSTM model to perform photovoltaic power ultra-short-term prediction, the influence of inaccurate meteorological data on prediction in a traditional prediction mode is reduced, the prediction precision is greatly improved, the assessment requirement of a power system is met, the operation management efficiency of a photovoltaic power station is improved, the power grid scheduling is optimized, and certain economic loss is avoided.
As shown in fig. 4, an embodiment of the present application provides a kalman filter and LSTM-based power generation ultra-short term prediction apparatus, including: the acquiring unit 201 is configured to acquire historical measured meteorological data, historical predicted meteorological data, historical power data, and weather forecast data at a first preset time in the future; a data cleaning unit 202, configured to perform data cleaning on the data, and perform data normalization; the first training unit 203 is used for inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, inputting weather forecast data at a first preset time in the future after the model training is finished, and obtaining 16-step Kalman correction predicted meteorological data; the second training unit 204 is used for inputting the processed historical meteorological data and historical power data into the LSTM model for training, comparing training results, setting a target error threshold value, if the error is smaller than the target error threshold value, saving the model, otherwise, continuing the training until the target error threshold value is reached; the prediction unit 205 is used for loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result and performing error analysis; and an inverse normalization processing unit 206, configured to perform inverse normalization processing on the prediction result subjected to the error analysis finally, and store the result.
In the present application, the embodiment of the ultra-short term prediction device based on kalman filtering and LSTM power generation is basically similar to the embodiment of the ultra-short term prediction method based on kalman filtering and LSTM power generation, and for the relevant points, reference is made to the introduction of the embodiment of the ultra-short term prediction method based on kalman filtering and LSTM power generation.
It is clear to a person skilled in the art that the solution according to the embodiments of the invention can be implemented by means of software and/or hardware. The "unit" and "module" in this specification refer to software and/or hardware that can perform a specific function independently or in cooperation with other components, where the hardware may be, for example, an FPGA (Field-Programmable Gate Array), an IC (Integrated Circuit), or the like.
Each processing unit and/or module according to the embodiments of the present invention may be implemented by an analog circuit that implements the functions described in the embodiments of the present invention, or may be implemented by software that executes the functions described in the embodiments of the present invention.
The embodiment of the invention also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when the program is executed by a processor, the steps of the ultra-short term prediction method based on Kalman filtering and LSTM power generation are realized. The computer-readable storage medium may include, but is not limited to, any type of disk including floppy disks, optical disks, DVD, CD-ROMs, microdrive, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of media or device suitable for storing instructions and/or data.
Fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present application, such as a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers, as shown in fig. 5. The computer device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The computer apparatus of the present application comprises a processor 401, a memory 402, an input device 403 and an output device 404. The processor 401, memory 402, input device 403, and output device 404 may be connected by a bus 405 or otherwise. The memory 402 has stored thereon a computer program that is executable on the processor 401, and the processor 401, when executing the program, implements the above-described kalman filtering and LSTM power generation ultra-short term prediction method steps.
The input device 403 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the data processing computer apparatus, such as a touch screen, keypad, mouse, track pad, touch pad, pointer stick, one or more mouse buttons, track ball, joystick or other input device. The output devices 404 may include a display device, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibrating motors), among others. Display devices may include, but are not limited to, Liquid Crystal Displays (LCDs), Light Emitting Diode (LED) displays, plasma displays, and touch screens.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.
All functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A method for ultra-short-term prediction of power generation based on Kalman filtering and LSTM is characterized by comprising the following steps:
acquiring historical measured meteorological data, historical predicted meteorological data, historical power data and weather forecast data at a first preset time in the future;
carrying out data cleaning on the data, and carrying out data normalization;
inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, and inputting weather forecast data at a first preset time in the future after the model training is finished to obtain 16-step Kalman correction predicted meteorological data;
inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a target error threshold, if the error is smaller than the target error threshold, saving the model, otherwise, continuing training until the target error threshold is reached;
loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result, and performing error analysis;
and performing inverse normalization processing on the prediction result subjected to error analysis finally, and storing the result.
2. The ultra-short term power generation prediction method based on Kalman filtering and LSTM according to claim 1, wherein the inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, inputting weather forecast data of a first preset time in the future after the model training is completed to obtain 16-step Kalman correction predicted meteorological data, comprises:
the system matrix adopts an identity matrix InThe system equation and the observation equation are expressed as follows:
xt(ti+1)＝xt(ti)+η(ti)
Figure FDA0003009159620000011
system covariance matrix Q (t)i) And an observation matrix R (t)i) Based on η (t)i)＝xt(ti+1)-xt(ti) And
Figure FDA0003009159620000012
the last 7 values of (a), are expressed as follows:
Figure FDA0003009159620000021
Figure FDA0003009159620000022
correcting multi-step meteorological data principle into known observation matrix H based on Kalman filteringiBy the above formula to ti+1Updating the time observation vector
Figure FDA0003009159620000023
3. The ultra-short term prediction method for power generation based on Kalman filtering and LSTM according to claim 1 or 2, characterized in that the first preset time is four hours.
4. The ultra-short term prediction method for power generation based on Kalman filtering and LSTM according to claim 1 or 2, wherein the data cleaning comprises: and missing value filling and abnormal value processing are carried out on the historical measured meteorological data, the historical predicted meteorological data, the historical power data and the weather forecast data at the first preset time in the future.
5. A power generation ultra-short term prediction device based on Kalman filtering and LSTM is characterized by comprising:
the acquiring unit is used for acquiring historical measured meteorological data, historical predicted meteorological data, historical power data and weather forecast data at a first preset time in the future;
the data cleaning unit is used for cleaning the data and normalizing the data;
the first training unit is used for inputting historical measured meteorological data and historical predicted meteorological data into a Kalman filter for model training, inputting weather forecast data at a first preset time in the future after the model training is finished, and obtaining 16-step Kalman correction predicted meteorological data;
the second training unit is used for inputting the processed historical meteorological data and historical power data into an LSTM model for training, comparing training results, setting a target error threshold value, if the error is smaller than the target error threshold value, saving the model, otherwise, continuing the training until the target error threshold value is reached;
the prediction unit is used for loading the trained LSTM model, inputting 16 steps of Kalman correction prediction meteorological data to obtain a prediction result and carrying out error analysis;
and the inverse normalization processing unit is used for performing inverse normalization processing on the prediction result subjected to the error analysis finally and storing the result.
6. The Kalman filtering and LSTM power generation ultra-short term prediction apparatus of claim 5, wherein the first training unit is configured to:
the system matrix adopts an identity matrix InThe system equation and the observation equation are expressed as follows:
xt(ti+1)＝xt(ti)+η(ti)
Figure FDA0003009159620000031
system covariance matrix Q (t)i) And an observation matrix R (t)i) Based on η (t)i)＝xt(ti+1)-xt(ti) And
Figure FDA0003009159620000032
the last 7 values of (a), are expressed as follows:
Figure FDA0003009159620000033
Figure FDA0003009159620000034
correcting multi-step meteorological data principle into known observation matrix H based on Kalman filteringiBy the above formula to ti+1Updating the time observation vector
Figure FDA0003009159620000035
7. The Kalman filtering and LSTM power generation ultra-short term prediction apparatus according to claim 5 or 6, wherein the first preset time is four hours.
8. The Kalman filtering and LSTM power generation ultra short term prediction apparatus according to claim 5 or 6, wherein the data washing unit is configured to: and missing value filling and abnormal value processing are carried out on the historical measured meteorological data, the historical predicted meteorological data, the historical power data and the weather forecast data at the first preset time in the future.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 4.
10. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1-4 are implemented when the program is executed by the processor.
CN202110370624.2A 2021-04-07 2021-04-07 Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM Pending CN113095562A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110370624.2A CN113095562A (en) 2021-04-07 2021-04-07 Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110370624.2A CN113095562A (en) 2021-04-07 2021-04-07 Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM

Publications (1)

Publication Number Publication Date
CN113095562A true CN113095562A (en) 2021-07-09

Family

ID=76674331

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110370624.2A Pending CN113095562A (en) 2021-04-07 2021-04-07 Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM

Country Status (1)

Country Link
CN (1) CN113095562A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113468820A (en) * 2021-07-21 2021-10-01 上海眼控科技股份有限公司 Data training method, device, equipment and storage medium
CN113537591A (en) * 2021-07-14 2021-10-22 北京琥珀创想科技有限公司 Long-term weather prediction method and device, computer equipment and storage medium
WO2023044770A1 (en) * 2021-09-24 2023-03-30 京东方科技集团股份有限公司 Dry pump downtime early warning method and apparatus, electronic device, storage medium, and program
CN117057174A (en) * 2023-10-13 2023-11-14 长江三峡集团实业发展(北京)有限公司 Runoff prediction method for data-missing area
CN117192063A (en) * 2023-11-06 2023-12-08 山东大学 Water quality prediction method and system based on coupled Kalman filtering data assimilation
CN117648873A (en) * 2024-01-30 2024-03-05 国网天津市电力公司电力科学研究院 Ground subsidence prediction method, training method, device and equipment
WO2024104153A1 (en) * 2022-11-14 2024-05-23 苏州元脑智能科技有限公司 Data enhancement method, system and device, and computer-readable storage medium
CN117648873B (en) * 2024-01-30 2024-06-04 国网天津市电力公司电力科学研究院 Ground subsidence prediction method, training method, device and equipment

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170371308A1 (en) * 2016-06-22 2017-12-28 International Business Machines Corporation Real-Time Forecasting of Electricity Demand in a Streams-Based Architecture with Applications
CN108694484A (en) * 2018-08-30 2018-10-23 广东工业大学 A kind of photovoltaic power generation power prediction method
CN108734331A (en) * 2018-03-23 2018-11-02 武汉理工大学 Short-term photovoltaic power generation power prediction method based on LSTM and system
CN110580549A (en) * 2019-09-02 2019-12-17 山东大学 Regional short-term energy power prediction method and system considering weather
CN111310963A (en) * 2018-12-12 2020-06-19 汉能移动能源控股集团有限公司 Power generation data prediction method and device for power station, computer equipment and storage medium
CN111461444A (en) * 2020-04-07 2020-07-28 上海电气风电集团股份有限公司 Prediction method, system, medium and electronic device for unit power of wind power plant
CN111695601A (en) * 2020-05-15 2020-09-22 特变电工西安电气科技有限公司 Photovoltaic power prediction method, device, equipment and readable storage medium
CN111709569A (en) * 2020-06-11 2020-09-25 国网冀北电力有限公司电力科学研究院 Method and device for predicting and correcting output power of wind power plant
CN111832800A (en) * 2020-05-28 2020-10-27 中国电力科学研究院有限公司 Photovoltaic power station power prediction method and device
CN111915092A (en) * 2020-08-11 2020-11-10 东北大学 Ultra-short-term wind power prediction method based on long-time and short-time memory neural network
CN112101626A (en) * 2020-08-14 2020-12-18 安徽继远软件有限公司 Distributed photovoltaic power generation power prediction method and system
CN112580862A (en) * 2020-12-08 2021-03-30 国家电网有限公司 Method and device for predicting short-term real-time power generation power of distributed photovoltaic system

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170371308A1 (en) * 2016-06-22 2017-12-28 International Business Machines Corporation Real-Time Forecasting of Electricity Demand in a Streams-Based Architecture with Applications
CN108734331A (en) * 2018-03-23 2018-11-02 武汉理工大学 Short-term photovoltaic power generation power prediction method based on LSTM and system
CN108694484A (en) * 2018-08-30 2018-10-23 广东工业大学 A kind of photovoltaic power generation power prediction method
CN111310963A (en) * 2018-12-12 2020-06-19 汉能移动能源控股集团有限公司 Power generation data prediction method and device for power station, computer equipment and storage medium
CN110580549A (en) * 2019-09-02 2019-12-17 山东大学 Regional short-term energy power prediction method and system considering weather
CN111461444A (en) * 2020-04-07 2020-07-28 上海电气风电集团股份有限公司 Prediction method, system, medium and electronic device for unit power of wind power plant
CN111695601A (en) * 2020-05-15 2020-09-22 特变电工西安电气科技有限公司 Photovoltaic power prediction method, device, equipment and readable storage medium
CN111832800A (en) * 2020-05-28 2020-10-27 中国电力科学研究院有限公司 Photovoltaic power station power prediction method and device
CN111709569A (en) * 2020-06-11 2020-09-25 国网冀北电力有限公司电力科学研究院 Method and device for predicting and correcting output power of wind power plant
CN111915092A (en) * 2020-08-11 2020-11-10 东北大学 Ultra-short-term wind power prediction method based on long-time and short-time memory neural network
CN112101626A (en) * 2020-08-14 2020-12-18 安徽继远软件有限公司 Distributed photovoltaic power generation power prediction method and system
CN112580862A (en) * 2020-12-08 2021-03-30 国家电网有限公司 Method and device for predicting short-term real-time power generation power of distributed photovoltaic system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
郑培等: "基于卡尔曼滤波修正的多步风电功率预测", 《热能动力工程》, vol. 35, no. 4, pages 1 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113537591A (en) * 2021-07-14 2021-10-22 北京琥珀创想科技有限公司 Long-term weather prediction method and device, computer equipment and storage medium
CN113468820A (en) * 2021-07-21 2021-10-01 上海眼控科技股份有限公司 Data training method, device, equipment and storage medium
WO2023044770A1 (en) * 2021-09-24 2023-03-30 京东方科技集团股份有限公司 Dry pump downtime early warning method and apparatus, electronic device, storage medium, and program
WO2024104153A1 (en) * 2022-11-14 2024-05-23 苏州元脑智能科技有限公司 Data enhancement method, system and device, and computer-readable storage medium
CN117057174A (en) * 2023-10-13 2023-11-14 长江三峡集团实业发展(北京)有限公司 Runoff prediction method for data-missing area
CN117057174B (en) * 2023-10-13 2024-01-26 长江三峡集团实业发展(北京)有限公司 Runoff prediction method for data-missing area
CN117192063A (en) * 2023-11-06 2023-12-08 山东大学 Water quality prediction method and system based on coupled Kalman filtering data assimilation
CN117192063B (en) * 2023-11-06 2024-03-15 山东大学 Water quality prediction method and system based on coupled Kalman filtering data assimilation
CN117648873A (en) * 2024-01-30 2024-03-05 国网天津市电力公司电力科学研究院 Ground subsidence prediction method, training method, device and equipment
CN117648873B (en) * 2024-01-30 2024-06-04 国网天津市电力公司电力科学研究院 Ground subsidence prediction method, training method, device and equipment

Similar Documents

Publication Publication Date Title
CN113095562A (en) Ultra-short term power generation prediction method and device based on Kalman filtering and LSTM
Chen et al. MMDetection: Open mmlab detection toolbox and benchmark
Shi et al. Support vector echo-state machine for chaotic time-series prediction
Zeng et al. SMO-based pruning methods for sparse least squares support vector machines
Zhang et al. Dynamic extreme learning machine and its approximation capability
CN117039895B (en) Wind power prediction method and system for energy storage auxiliary black start
US20220036178A1 (en) Dynamic gradient aggregation for training neural networks
CN113657661A (en) Enterprise carbon emission prediction method and device, computer equipment and storage medium
Xiaolin et al. Multistage pruning of CNN based ECG classifiers for edge devices
CN114118287A (en) Sample generation method, sample generation device, electronic device and storage medium
Bao et al. Using segmentation with multi-scale selective kernel for visual object tracking
CN111679829B (en) Method and device for determining user interface design
CN115796231B (en) Temporal analysis ultra-short term wind speed prediction method
CN113409898A (en) Molecular structure acquisition method and device, electronic equipment and storage medium
CN116151446A (en) Ultra-short-term wind power prediction method and device based on CEEMD and CNN-LSTM models
CN115759283A (en) Model interpretation method and device, electronic equipment and storage medium
CN115148292A (en) Artificial intelligence-based DNA (deoxyribonucleic acid) motif prediction method, device, equipment and medium
US11830207B2 (en) Method, apparatus, electronic device and readable storage medium for point cloud data processing
CN111291820B (en) Target detection method combining positioning information and classification information
CN113961765A (en) Searching method, device, equipment and medium based on neural network model
US20230080876A1 (en) Image processing apparatus, image recognition system, and image processing method
Tang et al. Deep sparse representation via deep dictionary learning for reinforcement learning
CN111931994A (en) Short-term load and photovoltaic power prediction method, system, equipment and medium thereof
CN111753760A (en) Model generation method and device, electronic equipment and storage medium
Ajay et al. NC-Emotions: Neuromorphic hardware accelerator design for facial emotion recognition

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