Generated power prediction method and device

Abstract

The invention discloses a generation power prediction method and device, and relates to the technical field of generation power prediction. The method comprises the following steps: acquiring fragment vectors corresponding to a plurality of sequence fragments based on power generation data at a plurality of continuous historical moments and meteorological data at N future moments; the first sub-vector of the fragment vector is obtained based on meteorological data of continuous U moments in the sequence fragment. The second sub-vector is obtained based on the generated power of the last N moments in the U moments. Then, the similarity between a search vector obtained based on meteorological data of N future moments and a first sub-vector of each fragment vector is obtained, at least two target fragment vectors with high similarity are obtained, then weighted summation is carried out on second sub-vectors of the at least two target fragment vectors, and the target fragment vectors of the at least two target fragment vectors are obtained; and the generated power of the target equipment at each future moment is obtained. In the process, physical characteristics and an operation mechanism of the photovoltaic power generation system do not need to be deeply researched, so that the calculation complexity of predicting the generated power is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of power generation prediction, and in particular to a method and device for predicting power generation. Background Art

[0002] With the rapid development of renewable energy, photovoltaic power generation, as a clean and renewable energy form, is gradually gaining widespread attention and application. However, photovoltaic power generation is intermittent and volatile, which brings challenges to the stable operation of the power system. Therefore, predicting the output of photovoltaic power generation is crucial to ensure the safe and stable operation and dispatch of the power system.

[0003] In the related art, the photovoltaic output can be predicted by deeply studying the physical characteristics and operation mechanism of the photovoltaic power generation system and establishing a physical model. However, the calculation complexity of the prediction method in the related art is relatively high. Summary of the invention

[0004] The present application provides a method and device for predicting power generation, which can solve the problem of high computational complexity of prediction methods in related technologies. The technical solution is as follows:

[0005] In one aspect, a method for predicting power generation is provided, the method comprising:

[0006] Based on the power generation data sequence of the target device and the meteorological data of N future moments, a plurality of sequence segments are obtained, wherein the power generation data sequence includes: power generation data of a plurality of consecutive historical moments, the power generation data includes: power generation power and meteorological data, and each of the sequence segments includes: the power generation data of (M+N) consecutive moments, the (M+N) moments including the historical moment;

[0007] Based on each of the sequence segments, a corresponding segment vector is obtained, the segment vector comprising: a first sub-vector and a second sub-vector, the first sub-vector being obtained based on the power generation data of U consecutive moments in the sequence segment, and the second sub-vector being obtained based on the power generation power of the last N moments among the U moments, where U is greater than or equal to M and less than or equal to (M+N);

[0008] Based on the similarity between the search vector and the first subvector of each of the segment vectors, obtaining at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors, wherein the search vector is determined based on at least the meteorological data at the N future moments;

[0009] A weighted sum is performed on the second subvectors of the at least two target segment vectors to determine the generated power of the target device at each of the future moments.

[0010] Optionally, before performing weighted summation on the second subvectors of the at least two target segment vectors to obtain the generated power of the target device at each of the future moments, the method further includes:

[0011] For each of the target segment vectors, determining a weight of a second subvector of the target segment vector based on a similarity corresponding to the target segment vector and a sum of similarities corresponding to the at least two target time series;

[0012] The weight of the second subvector of the target segment vector is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum value.

[0013] Optionally, before obtaining at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors based on the similarity between the search vector and the first subvectors of each segment vector, the method further includes:

[0014] Obtaining the distance between the search vector and the first subvector of each of the fragment vectors;

[0015] Based on the distance, a similarity between the search vector and the first sub-vector is determined, wherein the similarity is negatively correlated with the distance.

[0016] Optionally, the distance is a dynamic time warping distance or a shape-based distance.

[0017] Optionally, before acquiring a plurality of sequence segments based on the power generation data sequence of the target device and the meteorological data at N future moments, the method further includes:

[0018] The power generation data sequence of the target device and the meteorological data at N future moments are standardized respectively.

[0019] On the other hand, a power generation prediction device is provided, the device comprising:

[0020] A first acquisition module is used to acquire multiple sequence fragments based on the power generation data sequence of the target device and the meteorological data of N future moments, wherein the power generation data sequence includes: power generation data of multiple consecutive historical moments, the power generation data includes: power generation power and meteorological data, and each of the sequence fragments includes: the power generation data of (M+N) consecutive moments, and the (M+N) moments include the historical moment;

[0021] A second acquisition module is used to acquire a corresponding segment vector based on each of the sequence segments, wherein the segment vector includes: a first subvector and a second subvector, wherein the first subvector is obtained based on the power generation data of U consecutive moments in the sequence segment, and the second subvector is obtained based on the power generation power of the last N moments among the U moments, where U is greater than or equal to M and less than or equal to (M+N);

[0022] A third acquisition module is used to acquire at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors based on a similarity between a search vector and a first subvector of each of the segment vectors, wherein the search vector is determined based on at least the meteorological data at the N future moments;

[0023] The first determination module is used to perform weighted summation on the second subvectors of the at least two target segment vectors to determine the generated power of the target device at each future moment.

[0024] Optionally, the device further comprises:

[0025] a second determination module, configured to determine, for each of the target segment vectors, a weight of a second subvector of the target segment vector based on a similarity corresponding to the target segment vector and a sum of similarities corresponding to the at least two target time series;

[0026] The weight of the second subvector of the target segment vector is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum value.

[0027] Optionally, the device further comprises:

[0028] A third acquisition module, configured to acquire the distance between the search vector and the first subvector of each of the fragment vectors;

[0029] The third determination module is used to determine the similarity between the search vector and the first sub-vector based on the distance, wherein the similarity is negatively correlated with the distance.

[0030] Optionally, the device further comprises:

[0031] The standardization processing module is used to perform standardization processing on the power generation data sequence of the target device and the meteorological data at N future moments respectively.

[0032] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in the above aspects is implemented.

[0033] The beneficial effects brought about by the technical solution provided by this application include at least:

[0034] The present application provides a method and device for predicting power generation. The method can obtain multiple sequence fragments based on the power generation data of multiple consecutive historical moments of the target device and the meteorological data of N future moments, and obtain the corresponding fragment vector based on each sequence fragment. The first subvector in the fragment vector is obtained based on the power generation data of U consecutive moments in the corresponding sequence fragment, and the second subvector is obtained based on the power generation power of the last N moments of the U moments, where U is greater than or equal to M. Then, based on the similarity between the search vector and the first subvector in each fragment vector, the method can obtain at least two target fragment vectors whose similarity is greater than the similarity threshold, and then perform weighted summation on the second subvectors in the at least two target fragment vectors to determine the power generation power of the target device at each future moment. The search vector is obtained based on at least the meteorological data of N future moments. Since the first sub-vector can at least reflect the meteorological conditions at M consecutive moments, the search vector can at least reflect the meteorological conditions at N future moments, and the power generation mainly depends on the meteorological conditions, the method can obtain the power generation under meteorological conditions similar to the meteorological conditions at N future moments (i.e., the second sub-vector) through the first sub-vector and the search vector, and can achieve accurate prediction of the power generation at N future moments. In this process, since there is no need to deeply study the physical characteristics and operating mechanism of the photovoltaic power generation system, the computational complexity of predicting the power generation is reduced. In addition, since the method can determine the power generation of the target device at each future moment based on the second sub-vector in at least two target fragment vectors, the prediction accuracy of the power generation can be further improved.

[0035] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of a method for predicting power generation provided in an embodiment of the present application;

[0037] Figure 2 is a flow chart of another method for predicting power generation provided in an embodiment of the present application;

[0038] Figure 3 It is a structural schematic diagram of a power generation prediction device provided in an embodiment of the present application;

[0039] Figure 4 It is a structural schematic diagram of another power generation prediction device provided in an embodiment of the present application;

[0040] Figure 5 It is a structural schematic diagram of a controller provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0042] The present application embodiment provides a method for predicting power generation, which is applied to a power generation prediction device. Figure 1 , the method comprising:

[0043] Step 101: Acquire multiple sequence segments based on the power generation data sequence of the target device and the meteorological data at N future moments.

[0044] In an embodiment of the present application, the power generation prediction device can obtain a power generation data sequence of a target device and meteorological data at N future moments. The power generation data sequence includes: power generation data at multiple consecutive historical moments. Then, the power generation prediction device can divide the power generation data at multiple consecutive historical moments and the power generation data at N future moments to obtain multiple sequence segments.

[0045] The target device may be a photovoltaic device or a wind power generation device. The power generation data includes: power generation and meteorological data. Each sequence segment includes: power generation data of (M+N) consecutive moments, and (M+N) moments include historical moments. It should be understood that if the consecutive (M+N) moments of a sequence segment include a future moment, the power generation included in the power generation data at the future moment is 0.

[0046] Step 102: Based on each sequence segment, obtain a corresponding segment vector.

[0047] The segment vector corresponding to each sequence segment includes: a first subvector and a second subvector. The first subvector is obtained based on the power generation data of U consecutive moments of the sequence segment. The second subvector is obtained based on the power generation power of the last N moments of the U moments. Among them, U is greater than or equal to M and less than or equal to (M+N). That is, the power generation data of U consecutive moments can be the power generation data of at least M moments of the power generation data of (M+N) moments. The last N moments refer to: the last N moments of the U moments of the sequence segment arranged in order from early to late.

[0048] It can be seen that the first sub-vector can be used to reflect the meteorological conditions at least at M consecutive moments, that is, the first sub-vector can reflect the meteorological conditions at all or part of the (M+N) moments.

[0049] It can be understood that the dimensions and lengths of the first subvectors in any two fragment vectors in the multiple fragment vectors are the same, and the dimensions and lengths of the second subvectors are the same.

[0050] Step 103 : Based on the similarity between the search vector and the first subvector of each fragment vector, at least two target fragment vectors having a similarity greater than a similarity threshold are obtained from the multiple fragment vectors.

[0051] Since the power generation of the target device mainly depends on the meteorological conditions, the power generation prediction device can obtain a search vector based on at least the meteorological data of the N future moments when it needs to predict the power generation of the target device in the future time period. The search vector is at least used to reflect the meteorological conditions of the future time period. The future time period includes the N future moments. Then, the power generation prediction device can obtain meteorological conditions similar to the future time period based on the search vector. Specifically, the power generation prediction device can determine the similarity between the search vector and each first sub-vector, and obtain at least two target segment vectors whose similarity is greater than a similarity threshold from multiple segment vectors based on multiple similarities.

[0052] The length of the search vector is the same as the length of each first sub-vector, and the dimension of the search vector is the same as the dimension of each first sub-vector.

[0053] Step 104: Perform a weighted summation on the second subvectors of at least two target segment vectors to determine the generated power of the target device at each future moment.

[0054] After obtaining meteorological conditions similar to the future time period, the power generation prediction device can obtain the power generation of the target device at each future moment based on the power generation of the target device under the meteorological conditions. Specifically, the power generation prediction device can perform a weighted summation of the second subvectors in the at least two target segment vectors to obtain the power generation of the target device at each future moment.

[0055] In summary, the embodiment of the present application provides a method for predicting power generation, which can obtain multiple sequence segments based on the power generation data of multiple consecutive historical moments of the target device and the meteorological data of N future moments, and obtain the corresponding segment vector based on each sequence segment. The first subvector of the segment vector is obtained based on the power generation data of U consecutive moments in the corresponding sequence segment, and the second subvector is obtained based on the power generation of the last N moments in the U moments, where U is greater than or equal to M. Then, the method can obtain at least two target segment vectors whose similarity is greater than the similarity threshold based on the similarity between the search vector and the first subvector in each segment vector, and then perform weighted summation on the second subvectors in the at least two target segment vectors to determine the power generation of the target device at each future moment. Since the first subvector can at least reflect the meteorological conditions of M consecutive moments, the search vector can at least reflect the meteorological conditions of N future moments, and the power generation mainly depends on the meteorological conditions, the method can achieve accurate prediction of the power generation at N future moments by obtaining the power generation under meteorological conditions similar to the meteorological conditions of at least N future moments based on the first subvector and the search vector. In this process, since there is no need to deeply study the physical characteristics and operation mechanism of the photovoltaic power generation system, the computational complexity of predicting the power generation is reduced. In addition, since the method can determine the power generation of the target device at each future moment based on the second subvector in at least two target fragment vectors, the prediction accuracy of the power generation can be further improved.

[0056] Figure 2 is a flow chart of another method for predicting power generation provided in an embodiment of the present application, which method can be applied to a power generation prediction device. Figure 2 , the method may include:

[0057] Step 201: Standardize the power generation data sequence of the target device and the meteorological data at N future moments respectively.

[0058] In an embodiment of the present application, the power generation prediction device can obtain the power generation data sequence of the target device and the meteorological data of N future moments. The power generation data sequence includes: power generation data of multiple consecutive historical moments. Then, the power generation prediction device can perform standardization processing on the power generation data of the multiple historical moments and the meteorological data of N future moments respectively.

[0059] The target device may be a photovoltaic device or a wind power generation device. N future moments are continuous, and N is an integer greater than or equal to 1. The power generation data includes: power generation power and meteorological data. The meteorological data may include: temperature, perceived temperature, dew point temperature, relative humidity, rainfall, precipitation phase, wind speed, wind direction angle, gust wind speed, cloud cover, visibility, total solar radiation, solar scattered radiation, direct solar radiation, irradiation index, ultraviolet (UV) irradiance, sunrise time, sunset time, weather phenomena and ground pressure, etc. One or more.

[0060] For example, the meteorological data may include: temperature, perceived temperature, dew point temperature, relative humidity, rainfall, precipitation phase, wind speed, wind direction angle, gust wind speed, cloud cover, visibility, total solar radiation, solar scattered radiation, direct solar radiation, irradiance index, ultraviolet (UV) irradiance, sunrise time, sunset time, weather phenomena and ground pressure. The weather phenomena are special weather events such as fog, frost, thunderstorm, lightning, etc.

[0061] It can be seen that both the power generation data and the meteorological data at N future moments are multi-dimensional data. The dimension of the meteorological data depends on the number of types of meteorological parameters (such as temperature, humidity, and air pressure, etc.) in the meteorological data. The dimension of the power generation data is one more than the dimension of the meteorological data. Based on this, the process of standardizing the power generation data sequence (or the meteorological data at N future moments) may include: standardizing the data of each dimension in the power generation data sequence (or the meteorological data at N future moments).

[0062] Taking the standardization of the power generation data sequence as an example, the process of the power generation prediction device executing step 201 is exemplarily described as follows:

[0063] Assuming that the dimension of the power generation data sequence is I, any data v in the i-th dimension of the I-th dimension after standardization i * Can satisfy:

[0064]

[0065] In formula (1), i is the i-th dimension among I dimensions, v i is any data in the i-th dimension data in I dimensions, μ i is the mean of the i-th dimension data, σ i is the standard deviation of the i-th dimension data.

[0066] Since the power generation data sequence and the meteorological data at N future moments can be standardized respectively, the ranges of the power generation data sequence and the meteorological data at N future moments are scaled, and the impact of the dimension on the prediction of the power generation power can be reduced.

[0067] In the embodiment of the present application, the power generation prediction device can establish a communication connection with a meteorological data server and a power sensor for detecting the power generation of the target device. The power generation prediction device can obtain the power generation of the target device at a historical moment through the power sensor, and can request the meteorological data collected by the meteorological data server at a historical moment, as well as the predicted meteorological data at a future moment.

[0068] Step 202: Acquire multiple sequence segments based on the standardized power generation data sequence and the meteorological data at N future moments.

[0069] In an embodiment of the present application, the power generation prediction device can divide the standardized power generation data sequence and the meteorological data of N future moments in chronological order to obtain multiple sequence segments.

[0070] Among them, any two adjacent sequence segments in the plurality of sequence segments have an intersection. That is, any two adjacent sequence segments may include power generation data at the same time. Each sequence segment includes: power generation data at (M+N) consecutive moments. The (M+N) moments include historical moments. M is an integer greater than 1.

[0071] In the embodiment of the present application, for any sequence segment, the (M+N) moments of the sequence segment may all be historical moments, or the (M+N) moments may include both historical moments and future moments. If the (M+N) moments include future moments, the power generation data of the future moments included in the sequence segment may have a power generation power of 0.

[0072] Optionally, the power generation prediction device may use a sliding window to divide the standardized power generation data sequence and the meteorological data of N future moments to obtain multiple sequence segments. Specifically, the power generation prediction device may start from the earliest historical moment, use a sliding window to slide through multiple consecutive moments, and use each moment included in the sliding window and the power generation data at each moment as a sequence segment.

[0073] The multiple consecutive moments include: multiple historical moments and N future moments arranged in order from early to late. The length of the sliding window is the length of (M+N) moments. The sliding step length of the sliding window is smaller than the length of the sliding window. In this way, at least every two adjacent sequence segments can have an intersection.

[0074] Step 203: Based on each sequence segment, obtain a corresponding segment vector.

[0075] For each sequence segment in the plurality of sequence segments, the power generation prediction device can obtain a first sub-vector based on the power generation data of U consecutive moments in the sequence segment, and obtain a second sub-vector based on the power generation power of the last N moments in the power generation data of the U consecutive moments, thereby obtaining a segment vector corresponding to the sequence segment. Wherein, U is greater than or equal to M and less than or equal to (M+N). That is, the power generation data of U consecutive moments can be the power generation data of at least M moments of the power generation data of (M+N) moments. The last N moments refer to: the last N moments of the U moments arranged in order from early to late.

[0076] It can be seen that the first sub-vector can be used to reflect the meteorological conditions at least at M consecutive moments, that is, the first sub-vector can reflect the meteorological conditions at all or part of the (M+N) moments.

[0077] In the embodiment of the present application, each element in the second sub-vector may be the power generation power of the last N moments in the power generation data of consecutive U moments. Each element in the first sub-vector may be: all or part of the power generation data of consecutive U moments.

[0078] For example, for each sequence segment, the power generation prediction device can use the power generation data of U consecutive moments in the sequence segment as the elements in the first sub-vector; or use part of the power generation data of the U moments (such as meteorological data) as the elements in the first sub-vector, so as to obtain the first sub-vector.

[0079] It can be seen that the length of the first sub-vector can be greater than or equal to M and less than or equal to (M+N), that is, the length of the first sub-vector can be less than the length of the corresponding sequence fragment. And the dimension of the first sub-vector can be greater than or equal to 1, and less than or equal to the dimension of the corresponding fragment sequence. The length of the second sub-vector can be N, and the dimension of the second sub-vector is 1. Among them, the length of the first sub-vector refers to the total number of moments to which the first sub-vector includes power generation data, and the length of the second sub-vector can refer to: the total number of power generation powers included in the second sub-vector.

[0080] It should be understood that if the length of the first sub-vector is greater than M, the dimension of the first sub-vector is greater than the difference between the dimension of the corresponding fragment sequence and the dimension of the second sub-vector, and the first sub-vector includes meteorological data at a future time, then the power generation power in the first sub-vector at the future time is 0.

[0081] That is, if the first sub-vector includes power generation data at a future time, the power generation power in the power generation data is filled with 0.

[0082] Step 204: Obtain the similarity between the search vector and the first subvector of each segment vector.

[0083] Since the power generation of the target device mainly depends on the meteorological conditions, the power generation prediction device can obtain the search vector based on at least the meteorological data of the N future moments when it needs to predict the power generation of the target device in the future time period. Then, the power generation prediction device can determine the similarity between the search vector and each first subvector to obtain meteorological conditions similar to the future time period from the meteorological conditions of the historical time periods where the multiple sequence segments are located.

[0084] The search vector is at least used to reflect the weather conditions of a future period. The future period includes the N future moments. The length of the search vector is the same as the length of each first sub-vector, and the dimension of the search vector is the same as the dimension of each first sub-vector. The search vector and the first sub-vector include the same type of meteorological parameters in the meteorological data.

[0085] It can be understood that if the length of the first sub-vector is M+N, the number of future moments is N, and the dimension of the first sub-vector is the dimension of the meteorological data, the power generation prediction device can obtain the search vector based on the meteorological data of N future moments and the meteorological data of M consecutive historical moments earlier than the N future moments. The latest historical moment among the M historical moments is earlier than the earliest future moment among the N future moments. Each element of the search vector is the meteorological data of each moment in the N future moments and the M historical moments.

[0086] In an embodiment of the present application, for each fragment vector among multiple fragment vectors, the power generation prediction device can determine the distance between the first sub-vector in the fragment vector and the search vector, and then based on the distance, obtain the similarity between the search vector and the first sub-vector.

[0087] The similarity is negatively correlated with the distance. That is, the smaller the distance, the higher the similarity between the search vector and the first subvector; the larger the distance, the lower the similarity between the search vector and the first subvector. For example, the difference between 1 and the distance is used to represent the similarity.

[0088] Optionally, the distance between the first subvector and the search vector may be a dynamic time warping (DTW) distance, a shape-based distance (SBD), a Euclidean distance, a cosine distance, a Chebyshev distance, or the like.

[0089] For example, the distance between the first subvector and the search vector may be the DTW distance or the SBD. Since the DTW distance or the SBD not only considers the distance but also the shape of the sequence during the calculation process, it is more suitable for the application scenario of predicting the generated power.

[0090] Step 205 : Based on the similarity between the search vector and the first subvector of each fragment vector, at least two target fragment vectors whose similarity is greater than a similarity threshold are obtained from the multiple fragment vectors.

[0091] After obtaining the similarity between the search vector and each first sub-vector, the power generation prediction device can select at least two target segment vectors having a similarity greater than the similarity threshold from the multiple segment vectors based on the similarity threshold.

[0092] Among them, the meteorological conditions of the historical period in which the sequence segment corresponding to each target segment vector is located are relatively similar to the meteorological conditions of the future period. The similarity threshold can be pre-stored by the power generation prediction device. Alternatively, after the power generation prediction device obtains multiple similarities, the multiple similarities can be arranged in order from large to small, and the T+1th similarity can be determined as the similarity threshold. That is, at this time, the power generation prediction device can use each segment vector in the first T segment vectors of the multiple segment vectors arranged in order from large to small according to the similarity as a target segment vector.

[0093] Since the number of target segment vectors is at least two, that is, when predicting the power generation of the target equipment, the power generation prediction device does not only consider the power generation under the meteorological conditions most similar to the meteorological conditions in the future time period, but considers the power generation under multiple similar meteorological conditions. Therefore, it can ensure that the stability and accuracy of the prediction are high.

[0094] Step 206: Perform a weighted summation on the second subvectors of at least two target segment vectors to determine the generated power of the target device at each of the N future moments.

[0095] After obtaining meteorological conditions similar to the future time period, the power generation prediction device can obtain the power generation of the target device at each future moment based on the power generation of the target device under the meteorological conditions. Specifically, the power generation prediction device can perform weighted summation on the second subvectors of at least two target sequences to obtain a target vector. The length of the target vector is equal to the total number of N future moments, and each element of the target vector is the power generation at a future moment. In this way, the power generation of the target device at each future moment can be obtained.

[0096] The power generation prediction device performs weighted summation on at least two second sub-vectors, which means: performing weighted summation on elements with the same index in at least two second sub-vectors. Correspondingly, in the obtained target vector, the multiple elements arranged in ascending order of index are the power generation at N future moments arranged in descending order of time.

[0097] In an embodiment of the present application, for each target fragment vector, the power generation prediction device can determine the weight of the second subvector of the target fragment vector based on the similarity corresponding to the target fragment vector and the sum of multiple similarities corresponding to multiple target fragment vectors.

[0098] The weight is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum of the multiple similarities. The similarity corresponding to each target segment vector refers to: the similarity of the first subvector of the target segment vector.

[0099] For example, assuming that the total number of target fragment vectors is J, the weight w of the second subvector of the jth target fragment vector among the J target fragment vectors is j Can satisfy:

[0100]

[0101] In formula (2), exp(x) represents the natural constant e raised to the power of x. j is the similarity corresponding to the j-th target segment vector.

[0102] Assuming that the total number of elements included in each second sub-vector is K, the target vector also includes K elements, and the kth element a in the target vector k Can satisfy:

[0103]

[0104] In formula (3), refers to the kth element in the jth target fragment vector.

[0105] In the embodiment of the present application, if the power generation of the target device at a future moment is predicted based on the power generation data after the normalization processing, the power generation of the target device at each future moment obtained by the power generation prediction device after executing step 206 is also the power generation after the normalization processing. At this time, if it is necessary to obtain the power generation within the original data range, the power generation prediction device can perform inverse normalization processing on the power generation.

[0106] It is understandable that the order of the steps of the method for predicting power generation provided in the embodiment of the present application can be appropriately adjusted, and the steps can be increased or decreased accordingly according to the situation. For example, step 201 can be deleted according to the situation. Any technician familiar with the technical field can easily think of a method of change within the technical scope disclosed in this application, which should be included in the protection scope of this application, so it will not be repeated.

[0107] In summary, the embodiment of the present application provides a method for predicting power generation, which can obtain multiple sequence segments based on the power generation data of multiple consecutive historical moments of the target device and the meteorological data of N future moments, and obtain the corresponding segment vector based on each sequence segment. The first subvector of the segment vector is obtained based on the power generation data of U consecutive moments in the corresponding sequence segment, and the second subvector is obtained based on the power generation of the last N moments in the U moments, where U is greater than or equal to M. Then, based on the similarity between the search vector and the first subvector in each segment vector, the method can obtain at least two target segment vectors whose similarity is greater than the similarity threshold, and then perform weighted summation on the second subvectors in the at least two target segment vectors to determine the power generation of the target device at each future moment. Since the first subvector can at least reflect the meteorological conditions of M consecutive moments, the search vector can at least reflect the meteorological conditions of N future moments, and the power generation mainly depends on the meteorological conditions, the method can achieve accurate prediction of the power generation at N future moments by obtaining the power generation under meteorological conditions similar to the meteorological conditions of at least N future moments (i.e., the second subvector) based on the first subvector and the search vector. In this process, since there is no need to deeply study the physical characteristics and operation mechanism of the photovoltaic power generation system, the computational complexity of predicting the power generation is reduced. In addition, since the method can determine the power generation of the target device at each future moment based on the second subvector in at least two target fragment vectors, the prediction accuracy of the power generation can be further improved.

[0108] The present application embodiment provides a power generation prediction device, which can be used to execute the power generation prediction method provided by the above method embodiment. Figure 3 , the device 300 comprises:

[0109] The first acquisition module 301 is used to acquire multiple sequence segments based on the power generation data sequence of the target device and the meteorological data of N future moments. The power generation data sequence includes: power generation data of multiple consecutive historical moments, the power generation data includes: power generation power and meteorological data, and each sequence segment includes: power generation data of (M+N) consecutive moments, and the (M+N) moments include historical moments.

[0110] The second acquisition module 302 is used to acquire a corresponding segment vector based on each sequence segment, and the segment vector includes: a first subvector and a second subvector. The first subvector is obtained based on the power generation data of U consecutive moments in the sequence segment, and the second subvector is obtained based on the power generation power of the last N moments in the U moments, where U is greater than or equal to M and less than or equal to (M+N).

[0111] The third acquisition module 303 is used to acquire at least two target fragment vectors having a similarity greater than a similarity threshold from multiple fragment vectors based on the similarity between the search vector and the first subvector of each fragment vector, wherein the search vector is determined based on at least N future moments of meteorological data.

[0112] The first determination module 304 is configured to perform a weighted summation on the second subvectors of at least two target segment vectors to determine the generated power of the target device at each future moment.

[0113] Optional, see Figure 4 , the apparatus 300 may further include:

[0114] A second determination module 305 is used to determine, for each target segment vector, a weight of a second subvector of the target segment vector based on a similarity corresponding to the target segment vector and a sum of similarities corresponding to at least two target time series;

[0115] The weight of the second subvector of the target segment vector is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum value.

[0116] Optional, please continue to see Figure 4 , the apparatus 300 may further include:

[0117] A fourth acquisition module 306, configured to acquire a distance between the search vector and the first subvector of each fragment vector;

[0118] The third determination module 307 is used to determine the similarity between the search vector and the first sub-vector based on the distance, and the similarity is negatively correlated with the distance.

[0119] Optionally, the distance is a dynamic time warping distance or a shape-based distance.

[0120] Optional, such as Figure 4 As shown, the device 300 may also include:

[0121] The standardization processing module 308 is used to perform standardization processing on the power generation data sequence of the target device and the meteorological data at N future moments respectively.

[0122] In summary, the embodiment of the present application provides a power generation prediction device, which can obtain multiple sequence segments based on the power generation data of multiple consecutive historical moments of the target device and the meteorological data of N future moments, and obtain the corresponding segment vector based on each sequence segment. The first subvector of the segment vector is obtained based on the power generation data of U consecutive moments in the corresponding sequence segment, and the second subvector is obtained based on the power generation of the last N moments in the U moments, where U is greater than or equal to M. Then, the device can obtain at least two target segment vectors whose similarity is greater than the similarity threshold based on the similarity between the search vector and the first subvector in each segment vector, and then perform weighted summation on the second subvectors in the at least two target segment vectors to determine the power generation of the target device at each future moment. Since the first subvector can at least reflect the meteorological conditions of M consecutive moments, the search vector can at least reflect the meteorological conditions of N future moments, and the power generation mainly depends on the meteorological conditions, the method can achieve accurate prediction of the power generation at N future moments by obtaining the power generation under meteorological conditions similar to the meteorological conditions of at least N future moments based on the first subvector and the search vector. In this process, since there is no need to deeply study the physical characteristics and operation mechanism of the photovoltaic power generation system, the computational complexity of predicting the power generation is reduced. In addition, since the device can determine the power generation of the target device at each future moment based on the second subvector in at least two target fragment vectors, the prediction accuracy of the power generation can be further improved.

[0123] Figure 5 is a structural block diagram of a controller provided in an embodiment of the present application. The controller can execute the power generation prediction method provided in the above method embodiment. Figure 5 As shown, the controller 400 includes: a processor 401 and a memory 403. The processor 401 and the memory 403 are connected, such as through a bus 402. Optionally, the controller 400 may also include a transceiver 404. It should be noted that in actual applications, the transceiver 404 is not limited to one, and the structure of the controller 400 does not constitute a limitation on the embodiments of the present application.

[0124] Processor 401 may be a CPU (Central Processing Unit), a general purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 401 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0125] The bus 402 may include a path to transmit information between the above components. The bus 402 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 402 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0126] The memory 403 is used to store a computer program corresponding to the power generation prediction method of the above embodiment of the present application, and the computer program is controlled and executed by the processor 401. The processor 401 is used to execute the computer program stored in the memory 403 to implement the content shown in the above method embodiment.

[0127] The controller 400 includes but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers) and PMPs (portable multimedia players), as well as fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The controller 400 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0128] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the power generation prediction method provided in the above method embodiment is implemented. Figure 1 or Figure 2 The method shown.

[0129] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0130] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0131] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0132] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0133] In this application, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0134] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for predicting power generation, characterized in that: The method comprises: Based on the power generation data sequence of the target device and the meteorological data of N future moments, a plurality of sequence segments are obtained, wherein the power generation data sequence includes: power generation data of a plurality of consecutive historical moments, the power generation data includes: power generation power and meteorological data, and each of the sequence segments includes: the power generation data of (M+N) consecutive moments, the (M+N) moments including the historical moment; Based on each of the sequence segments, a corresponding segment vector is obtained, the segment vector comprising: a first subvector and a second subvector, the first subvector being obtained based on the power generation data at U consecutive moments in the sequence segment, and the second subvector being obtained based on the power generation power at the last N moments of the U consecutive moments, where U is greater than or equal to M and less than or equal to (M+N); Based on the similarity between a search vector and a first subvector of each of the segment vectors, obtaining at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors, wherein the search vector is determined based on the meteorological data at the N future moments; A weighted sum is performed on the second subvectors of the at least two target segment vectors to determine the generated power of the target device at each of the future moments.

2. The method according to claim 1, characterized in that Before performing weighted summation on the second subvectors of the at least two target segment vectors to obtain the generated power of the target device at each of the future moments, the method further includes: For each of the target segment vectors, determining a weight of a second subvector of the target segment vector based on a similarity corresponding to the target segment vector and a sum of similarities corresponding to the at least two target time series; The weight of the second subvector of the target segment vector is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum value.

3. The method according to claim 1, characterized in that Before acquiring at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors based on the similarity between the search vector and the first subvectors of each segment vector, the method further includes: Obtaining the distance between the search vector and the first subvector of each of the fragment vectors; Based on the distance, a similarity between the search vector and the first sub-vector is determined, wherein the similarity is negatively correlated with the distance.

4. The method according to claim 3, characterized in that: The distance is a dynamic time warping distance or a shape-based distance.

5. The method according to any one of claims 1 to 4, characterized in that: Before acquiring a plurality of sequence segments based on the power generation data sequence of the target device and the meteorological data at N future moments, the method further includes: The power generation data sequence of the target device and the meteorological data at N future moments are standardized respectively.

6. A power generation prediction device, characterized in that: The device comprises: A first acquisition module is used to acquire multiple sequence fragments based on the power generation data sequence of the target device and the meteorological data of N future moments, wherein the power generation data sequence includes: power generation data of multiple consecutive historical moments, the power generation data includes: power generation power and meteorological data, and each of the sequence fragments includes: the power generation data of (M+N) consecutive moments, and the (M+N) moments include the historical moment; A second acquisition module is used to acquire a corresponding segment vector based on each of the sequence segments, wherein the segment vector includes: a first subvector and a second subvector, wherein the first subvector is obtained based on the power generation data of U consecutive moments in the sequence segment, and the second subvector is obtained based on the power generation power of the last N moments among the U moments, where U is greater than or equal to M and less than or equal to (M+N); A third acquisition module is used to acquire at least two target segment vectors having a similarity greater than a similarity threshold from the multiple segment vectors based on a similarity between a search vector and a first subvector of each of the segment vectors, wherein the search vector is determined based on at least the meteorological data at the N future moments; The first determination module is used to perform a weighted summation on the second subvectors of the at least two target segment vectors to determine the generated power of the target device at each of the future moments.

7. The device according to claim 6, characterized in that The device also includes: a second determination module, configured to determine, for each of the target segment vectors, a weight of a second subvector of the target segment vector based on a similarity corresponding to the target segment vector and a sum of similarities corresponding to the at least two target time series; The weight of the second subvector of the target segment vector is positively correlated with the similarity corresponding to the target segment vector, and negatively correlated with the sum value.

8. The device according to claim 6, characterized in that The device also includes: A fourth acquisition module, configured to acquire a distance between the search vector and a first subvector of each of the fragment vectors; The third determination module is used for determining the distance and the similarity between the search vector and the first sub-vector, wherein the similarity is negatively correlated with the distance.

9. The device according to claim 8, characterized in that The device also includes: The standardization processing module is used to perform standardization processing on the power generation data sequence of the target device and the meteorological data at N future moments respectively.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images