CN105139080A - Improved photovoltaic power sequence prediction method based on Markov chain - Google Patents
Improved photovoltaic power sequence prediction method based on Markov chain Download PDFInfo
- Publication number
- CN105139080A CN105139080A CN201510471216.0A CN201510471216A CN105139080A CN 105139080 A CN105139080 A CN 105139080A CN 201510471216 A CN201510471216 A CN 201510471216A CN 105139080 A CN105139080 A CN 105139080A
- Authority
- CN
- China
- Prior art keywords
- state
- weather
- data
- photovoltaic
- markov chain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Photovoltaic Devices (AREA)
Abstract
The present invention discloses an improved photovoltaic power sequence prediction method based on a Markov chain. The method comprises: carrying out data preprocessing on sample data for photovoltaic power; classifying the sample data according to season characteristics, period characteristics and weather characteristics; establishing different transfer matrices based on the Markov chain; establishing a weather state transfer matrix according to historical weather data; calculating a base portion of the power data; and describing fluctuation characteristics of the power data by using a first-order differential component of the sample data at adjacent time points. The method provided by the present invention achieves the advantages that: photovoltaic output weather characteristics are taken into account in the Markov chain model, and the weather state change is described by using the weather state transfer matrix; the fluctuation characteristics area added into the calculation process from a photovoltaic output state to a photovoltaic output value to reflect the photovoltaic characteristics; and multiple state transfer matrices are generated according to the season characteristics, day characteristics and weather characteristics of power of photovoltaic power generation, and a corresponding matrix is determined and selected according to the climate characteristics and time attributes to generate a state quantity of a target time.
Description
Technical field
The present invention is a kind of photovoltaic power Forecasting Methodology, belongs to new forms of energy photovoltaic electric powder prediction.
Background technology
The fast development of photovoltaic power generation technology facilitates the maximization of photovoltaic plant and also networking.Due to randomness and the undulatory property of photovoltaic generation, the grid-connected quality of power supply to electrical network of large-scale photovoltaic and Stability and dependability is made to create adverse effect.The research method that current photovoltaic is exerted oneself mainly contains artificial neural network, least square support vector machines, combined method etc.The characteristic of research photovoltaic generation power, and then generate its simulated series assessment is grid-connectedly extremely important for the impact of electrical network and the scheduling planning of electrical network.
Simulated series has the following advantages compared to historical series: the phenomenon occurring shortage of data and error in data in (1) historical series most probably, utilize this historical data to carry out assessment result to have deviation (2) due to domestic photovoltaic industry and start late, less to some areas photovoltaic historical data of exerting oneself, data length is inadequate for evaluation requirement, utilize simulated series then can produce the data sequence of random length, the data that can embody photovoltaic power producing characteristics produced after feature extraction research is through to assessment more convenient (3) simulated series, therefore for its result of grid-connected assessment, there is credibility.
In existing research, the modeling method of photovoltaic power simulation is generally divided into two classes.The first kind is first according to the model of irradiation historical data Corpus--based Method method establishment irradiation intensity, then according to the Conversion of Energy relation of photovoltaic array, utilizes irradiation intensity to obtain photovoltaic power simulated data.Because the Conversion of Energy of photovoltaic array is subject to temperature, the impact of the many factors such as the chemical composition of photovoltaic cell, and these class methods are difficult to comprehensive these factors of comprehensively considering, data result can be caused inaccurate.Be difficult to probability statistics model modeling the timing considering formation sequence simultaneously.Equations of The Second Kind method is disregard the conversion process of irradiation intensity to photovoltaic power, directly utilize photovoltaic exert oneself historical data simulation generate photovoltaic go out force data.These class methods save photoelectric conversion process, without the need to considering that the many factors of conversion process improves data accuracy, simplify modeling process simultaneously.
Summary of the invention
For solving the deficiencies in the prior art, the object of the present invention is to provide a kind of improvement photovoltaic power sequence prediction method based on Markov chain, based on original single order Markov chain method, consider the Seasonal Characteristics of photovoltaic generation power, period characteristic, weather features generates multiple state-transition matrix, selects corresponding matrix to generate object time quantity of state through property identification, consider differential characteristic superposition difference component, thus set up the improvement Markov-chain model of photovoltaic power prediction.
In order to realize above-mentioned target, the present invention adopts following technical scheme:
Based on an improvement photovoltaic power sequence prediction method for Markov chain, it is characterized in that, comprise the steps:
1) data prediction is done to history photovoltaic power data, reject misdata, missing data surrounding time point mean value being replaced, forming the sample data collection S for predicting;
2) according to sample data data according to season, period, weather features classification samples data; Sample data is categorized as 32 data set { s
1, s
2... s
32;
3) to step 2) in data set carry out state classification, then set up Markovian state transition matrix P respectively, calculate accumulation state transition matrix Q;
4) suppose that current time is t, the current state of exerting oneself is β
t, photovoltaic numerical value of exerting oneself is α
t, place season is γ
t, the period is θ
t, state of weather is ε
t, undulate quantity is f
t; According to current γ
t, θ
t, ε
t, selected corresponding state-transition matrix
its element is p
i,j, and calculate corresponding accumulation state transition matrix
its element is
5) photovoltaic is set to exert oneself subsequent time state as β
t+1; Generate and obey equally distributed random number μ
t, and judge μ
tspan; If
then β
t+1=1, if
then β
t+1=m+1;
6) undulate quantity σ is determined in sampling
t, at the current numerical value α that exerts oneself
ton superpose undulate quantity σ
tobtain η, judge that whether η is in step 5) middle gained β
t+1corresponding exerts oneself in span; If so, then subsequent time goes out force value α
t+1=η, otherwise regenerate undulate quantity and judge; If t is not the finish time of this period, then ε
t+1=ε
t, otherwise carry out above-mentioned steps again carry out the selection of state of weather according to state of weather transition matrix after.
A kind of aforesaid improvement photovoltaic power sequence prediction method based on Markov chain, it is characterized in that, described step 2) middle Seasonal Characteristics is according to delimiting four seasons, and period characteristic is divided into the morning and afternoon hours, and weather features is divided into heavy rain, rain, cloudy and fine Four types.
The beneficial effect that the present invention reaches: consider that in Markov chain model photovoltaic is exerted oneself weather features, and utilize state of weather transition matrix to describe weather state change; Wave characteristic being joined photovoltaic state of exerting oneself exerts oneself in the computation process of numerical value to photovoltaic, embodies photovoltaic property; Consider the Seasonal Characteristics of photovoltaic generation power, day characteristic and weather features generate multiple state-transition matrix, differentiate through climate characteristic and time attribute and select corresponding matrix, generate object time quantity of state.
Embodiment
The invention will be further described below.Following examples only for technical scheme of the present invention is clearly described, and can not limit the scope of the invention with this.
The present invention relates to a kind of photovoltaic power Forecasting Methodology based on Markov chain, specifically comprise the steps:
Step 1) data prediction is done to history photovoltaic power data, reject its misdata, missing data surrounding time point mean value being replaced, forming the sample data collection S for predicting.
Its concrete real alternate data computing formula is:
wherein d
tfor disappearance power data, d
t-1with d
t+1be respectively certain data previous moment and a rear moment photovoltaic power value.
Step 2) according to sample data historical summary according to season, period, weather features classification samples data.Seasonal Characteristics is throughout the year, and period characteristic is divided into the morning and afternoon hours, and weather features is divided into heavy rain, rain, cloudy and fine Four types, sample data is categorized as 32 data set { s
1, s
2... s
32.Wherein fine day is designated as F, and cloudy day is designated as C, and shower day is designated as S, and heavy rain day is designated as R.Usable probability transition matrix P is changed between weather pattern
wrepresent, it is the square formation of one 4 × 4, and w represents weather pattern, w ∈ { F, C, S, R}.
When element P (F|F) wherein represents that the day before yesterday is fine day, this day is also the conditional probability of fine day, and the implication of other elements by that analogy, carries out weather features judgement according to state of weather transition matrix.
Step 3) state classification is carried out to data set, then set up Markovian state transition matrix respectively, calculate accumulation state transition matrix.
The matrix that Markov chain state-transition matrix is made up of the transition probability between different conditions is called state-transition matrix P, the square formation of its to be each row sum be N × N of 1.
This matrix remains unchanged in state migration procedure.
Be directed to the Markov process that photovoltaic is exerted oneself, its element can be estimated by following formula:
n
ijexpression state i transfers to the frequency of state j through a step, x
twith x
t+1represent the state in t and t+1 moment respectively, i and j is the element in state space.
Accumulated state transition matrix Q then calculates based on state-transition matrix.
Step 4) suppose that current time is t, the current state of exerting oneself is β
t, photovoltaic numerical value of exerting oneself is α
t, place month is γ
t, the period is θ
t, state of weather is ε
t, undulate quantity is f
t.
According to current γ
t, θ
t, ε
t, selected corresponding state-transition matrix
its element is p
i,j, and calculate corresponding accumulation state transition matrix
wherein element is
Step 5) set photovoltaic to exert oneself subsequent time state as β
t+1.Generate and obey equally distributed random number μ
t, and judge μ
tspan, if
then β
t+1=1, if
then β
t+1=m+1.
Step 6) sample and determine undulate quantity σ
t, at the current numerical value α that exerts oneself
ton superpose undulate quantity and obtain η, judge that whether η is at β
t+1span in, if then subsequent time goes out force value α
t+1=η, otherwise regenerate undulate quantity and judge again.If t is not the initial time of this period, then ε
t+1=ε
t, otherwise carry out above-mentioned steps again carry out the selection of state of weather according to state of weather transition matrix after.
The present invention considers that in Markov chain model photovoltaic is exerted oneself weather features, and utilizes state of weather transition matrix to describe weather state change; Wave characteristic being joined photovoltaic state of exerting oneself exerts oneself in the computation process of numerical value to photovoltaic, embodies photovoltaic property; Consider the Seasonal Characteristics of photovoltaic generation power, day characteristic and weather features generate multiple state-transition matrix, differentiate through climate characteristic and time attribute and select corresponding matrix, generate object time quantity of state.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from the technology of the present invention principle; can also make some improvement and distortion, these improve and distortion also should be considered as protection scope of the present invention.
Claims (2)
1., based on an improvement photovoltaic power sequence prediction method for Markov chain, it is characterized in that, comprise the steps:
1) data prediction is done to history photovoltaic power data, reject misdata, missing data surrounding time point mean value being replaced, forming the sample data collection S for predicting;
2) according to sample data data according to season, period, weather features classification samples data; Sample data is categorized as 32 data set { s
1, s
2... s
32;
3) to step 2) in data set carry out state classification, then set up Markovian state transition matrix P respectively, calculate accumulation state transition matrix Q;
4) suppose that current time is t, the current state of exerting oneself is β
t, photovoltaic numerical value of exerting oneself is α
t, place season is γ
t, the period is
, state of weather is ε
t, undulate quantity is f
t; According to current γ
t,
, ε
t, selected corresponding state-transition matrix
its element is p
i,j, and calculate corresponding accumulation state transition matrix
its element is
5) photovoltaic is set to exert oneself subsequent time state as β
t+1; Generate and obey equally distributed random number μ
t, and judge μ
tspan; If
then β
t+1=1, if
then β
t+1=m+1;
6) undulate quantity σ is determined in sampling
t, at the current numerical value α that exerts oneself
ton superpose undulate quantity σ
tobtain η, judge that whether η is in step 5) middle gained β
t+1state corresponding to exert oneself in span; If so, then subsequent time goes out force value α
t+1=η, otherwise regenerate undulate quantity and judge; If t is not the finish time of this period, then ε
t+1=ε
t, otherwise carry out above-mentioned steps again carry out the selection of state of weather according to state of weather transition matrix after.
2. a kind of improvement photovoltaic power sequence prediction method based on Markov chain according to claim 1, it is characterized in that, described step 2) middle Seasonal Characteristics is according to delimiting four seasons, and period characteristic is divided into the morning and afternoon hours, and weather features is divided into heavy rain, rain, cloudy and fine Four types.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510471216.0A CN105139080A (en) | 2015-08-04 | 2015-08-04 | Improved photovoltaic power sequence prediction method based on Markov chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510471216.0A CN105139080A (en) | 2015-08-04 | 2015-08-04 | Improved photovoltaic power sequence prediction method based on Markov chain |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105139080A true CN105139080A (en) | 2015-12-09 |
Family
ID=54724421
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510471216.0A Pending CN105139080A (en) | 2015-08-04 | 2015-08-04 | Improved photovoltaic power sequence prediction method based on Markov chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105139080A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105825298A (en) * | 2016-03-14 | 2016-08-03 | 梁海东 | Electric network metering early-warning system and method based on load characteristic pre-estimation |
| CN105914769A (en) * | 2016-05-03 | 2016-08-31 | 易事特集团股份有限公司 | Optical storage micro-grid energy management method and system in off-network state |
| CN106485093A (en) * | 2016-11-10 | 2017-03-08 | 哈尔滨工程大学 | Based on the solar irradiance time series synthetic method for improving Markov chain |
| CN107194495A (en) * | 2017-04-21 | 2017-09-22 | 北京信息科技大学 | A kind of longitudinal Forecasting Methodology of photovoltaic power excavated based on historical data |
| CN111652449A (en) * | 2020-07-17 | 2020-09-11 | 上海积成能源科技有限公司 | System model for predicting short-term photovoltaic power generation system generated energy based on conditional random field model |
| CN111784543A (en) * | 2020-07-17 | 2020-10-16 | 上海积成能源科技有限公司 | System model for predicting short-term photovoltaic power generation system power generation based on hidden Markov model |
| CN112861363A (en) * | 2021-02-23 | 2021-05-28 | 国网冀北电力有限公司智能配电网中心 | Photovoltaic power generation state estimation method and device and terminal equipment |
| CN114418424A (en) * | 2022-01-26 | 2022-04-29 | 北京航空航天大学 | Photovoltaic power station power generation amount evaluation method considering initial inventory configuration |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103400204A (en) * | 2013-07-26 | 2013-11-20 | 华南理工大学 | Forecasting method for solar photovoltaic electricity generation amount based on SVM (support vector machine) - Markov combination method |
| CN103439572A (en) * | 2013-08-15 | 2013-12-11 | 国家电网公司 | Electricity larceny prevention monitoring method based on photovoltaic power generation power prediction |
-
2015
- 2015-08-04 CN CN201510471216.0A patent/CN105139080A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103400204A (en) * | 2013-07-26 | 2013-11-20 | 华南理工大学 | Forecasting method for solar photovoltaic electricity generation amount based on SVM (support vector machine) - Markov combination method |
| CN103439572A (en) * | 2013-08-15 | 2013-12-11 | 国家电网公司 | Electricity larceny prevention monitoring method based on photovoltaic power generation power prediction |
Non-Patent Citations (5)
| Title |
|---|
| DING M等: "An ANN-based approach for forecasting the power output of photovoltaic system", 《PROCEDIA ENVIRONMENTAL SCIENCES》 * |
| 丁明等: "基于马尔可夫链的光伏发电系统输出功率短期预测方法", 《电网技术》 * |
| 吴振威等: "多时间尺度的光伏出力波动特性研究", 《现代电力》 * |
| 王飞等: "基于天气状态模式识别的光伏电站发电功率分类预测方法", 《中国电机工程学报》 * |
| 蒋平等: "基于改进一阶马尔可夫链的风速时间序列模型", 《电力系统自动化》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105825298A (en) * | 2016-03-14 | 2016-08-03 | 梁海东 | Electric network metering early-warning system and method based on load characteristic pre-estimation |
| CN105825298B (en) * | 2016-03-14 | 2020-05-01 | 梁海东 | Power grid metering early warning system and method based on load characteristic estimation |
| CN105914769A (en) * | 2016-05-03 | 2016-08-31 | 易事特集团股份有限公司 | Optical storage micro-grid energy management method and system in off-network state |
| CN105914769B (en) * | 2016-05-03 | 2019-02-22 | 易事特集团股份有限公司 | The energy management method and system of light storage micro-capacitance sensor under off-network state |
| CN106485093A (en) * | 2016-11-10 | 2017-03-08 | 哈尔滨工程大学 | Based on the solar irradiance time series synthetic method for improving Markov chain |
| CN106485093B (en) * | 2016-11-10 | 2019-02-01 | 哈尔滨工程大学 | Based on the solar irradiance time series synthetic method for improving Markov chain |
| CN107194495A (en) * | 2017-04-21 | 2017-09-22 | 北京信息科技大学 | A kind of longitudinal Forecasting Methodology of photovoltaic power excavated based on historical data |
| CN107194495B (en) * | 2017-04-21 | 2020-05-12 | 北京信息科技大学 | Photovoltaic power longitudinal prediction method based on historical data mining |
| CN111652449A (en) * | 2020-07-17 | 2020-09-11 | 上海积成能源科技有限公司 | System model for predicting short-term photovoltaic power generation system generated energy based on conditional random field model |
| CN111784543A (en) * | 2020-07-17 | 2020-10-16 | 上海积成能源科技有限公司 | System model for predicting short-term photovoltaic power generation system power generation based on hidden Markov model |
| CN112861363A (en) * | 2021-02-23 | 2021-05-28 | 国网冀北电力有限公司智能配电网中心 | Photovoltaic power generation state estimation method and device and terminal equipment |
| CN114418424A (en) * | 2022-01-26 | 2022-04-29 | 北京航空航天大学 | Photovoltaic power station power generation amount evaluation method considering initial inventory configuration |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105139080A (en) | Improved photovoltaic power sequence prediction method based on Markov chain | |
| Huang et al. | Multiple-input deep convolutional neural network model for short-term photovoltaic power forecasting | |
| Mellit et al. | Short-term forecasting of power production in a large-scale photovoltaic plant | |
| WO2017054537A1 (en) | Long-time scale photovoltaic output time sequence modelling method and apparatus | |
| CN110070226A (en) | Photovoltaic power prediction technique and system based on convolutional neural networks and meta learning | |
| CN107194495B (en) | Photovoltaic power longitudinal prediction method based on historical data mining | |
| CN104573879A (en) | Photovoltaic power station output predicting method based on optimal similar day set | |
| CN109165774A (en) | A kind of short-term photovoltaic power prediction technique | |
| Al-Sbou et al. | Nonlinear autoregressive recurrent neural network model for solar radiation prediction | |
| Melzi et al. | Hourly solar irradiance forecasting based on machine learning models | |
| CN105426956A (en) | Ultra-short-period photovoltaic prediction method | |
| CN113496311A (en) | Photovoltaic power station generated power prediction method and system | |
| CN103218674A (en) | Method for predicating output power of photovoltaic power generation system based on BP (Back Propagation) neural network model | |
| CN106485093B (en) | Based on the solar irradiance time series synthetic method for improving Markov chain | |
| CN110311420A (en) | A kind of generation method of scene joint power output timing scene | |
| CN103761578A (en) | Solar irradiation predicting method based on multiple linear regression | |
| WO2024051524A1 (en) | Joint prediction method and apparatus for hydraulic, wind and photovoltaic generation power | |
| CN106503848A (en) | The load forecasting method of many small power station's bulk sale area power grids | |
| Luo et al. | Short-term photovoltaic generation forecasting based on similar day selection and extreme learning machine | |
| CN110084430A (en) | A method of considering space-time characterisation design distributed photovoltaic power output prediction model | |
| Chicco et al. | Characterization of solar irradiance profiles for photovoltaic system studies through data rescaling in time and amplitude | |
| CN115765044A (en) | Wind, light and water power system combined operation and risk analysis method and system | |
| CN114897229A (en) | Ultra-short-term prediction method for irradiance of photovoltaic power station based on AP clustering and Markov chain | |
| CN112949938B (en) | Wind power climbing event direct forecasting method for improving training sample class imbalance | |
| Tu et al. | Artificial neural networks in the estimation of monthly capacity factors of WECS in Taiwan |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151209 |