Distributed photovoltaic actual measurement load data adaptive fitting method and device

Abstract

The invention provides a distributed photovoltaic actual measurement load data adaptive fitting method and device. According to four fitting schemes, centralized photovoltaic actual power data, 10kV distributed photovoltaic actual power data, low-voltage distributed photovoltaic template station actual power data and weighted average of the three actual power data are respectively utilized to carry out actual power data fitting at 96 time points in a day according to installed capacity; setting an error threshold value; fitting scheme switching control is carried out based on a self-adaptive PID algorithm, the error of the distributed photovoltaic actual power generation data at the current moment and the last moment is compared with the error threshold value, switching is carried out in sequence if the error threshold value is exceeded, and fitting of the distributed photovoltaic actual power generation data is continued according to the fitting scheme after switching; and calculating actually measured load data according to the distributed photovoltaic actual power data obtained by fitting. According to the method, the problem that the power grid load prediction accuracy is reduced due to distributed photovoltaic operation fluctuation can be solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of new energy electric power, and in particular relates to a distributed photovoltaic measured load data adaptive fitting method and device. Background Art

[0002] At this stage, new energy electricity has been rapidly developed. While the total installed capacity of wind power and photovoltaic power has increased rapidly, it has also brought huge challenges to the safe operation of the power grid. The accuracy of load forecasts in various power grids has been affected to varying degrees, especially in areas where low-voltage distributed photovoltaics have no access to actual power data. The quality of distributed photovoltaic actual power data directly affects the measured load data, which further affects the accuracy of power grid load forecasts. Summary of the invention

[0003] The present invention proposes a distributed photovoltaic measured load data adaptive fitting method and device to solve the problem of reduced grid load prediction accuracy caused by distributed photovoltaic operation fluctuations.

[0004] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0005] A distributed photovoltaic measured load data adaptive fitting method, comprising: using centralized photovoltaic actual power data, 10kV distributed photovoltaic actual power data, low-voltage distributed photovoltaic model station actual power data, and the weighted average of the three actual power data, setting four fitting schemes for fitting the actual power data at 96 points in a day according to the distributed photovoltaic installed capacity;

[0006] Setting an error threshold; performing fitting scheme switching control based on an adaptive PID algorithm, comparing the error between the distributed photovoltaic actual power data at the current time point and the previous time point obtained by fitting with the error threshold, and switching the four fitting schemes in order if the error threshold is exceeded, and continuing to fit the distributed photovoltaic actual power data according to the switched fitting scheme;

[0007] The measured load data is calculated based on the distributed photovoltaic actual power data obtained by fitting.

[0008] Furthermore, a fifth scheme is included, which uses the distributed photovoltaic short-term power prediction results to replace the distributed photovoltaic actual power data at 96 points in a day; the fifth scheme is added to the scheme switching sorting based on the adaptive PID algorithm as the last scheme in the sorting.

[0009] Furthermore, the fitting scheme using the actual power data of centralized photovoltaic power generation includes: selecting a centralized photovoltaic power station that is operating normally within the region and has direct optical fiber actual power generation, using the actual power generation of the centralized photovoltaic power station every 15 minutes, and fitting the total actual power generation of distributed photovoltaic power generation according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 集中式 )*P 集中式 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaics, P 集中式 、M 集中式 They are respectively the actual generated power and installed capacity of the centralized photovoltaic system.

[0010] Furthermore, the fitting scheme using the 10kV distributed photovoltaic actual power data includes: selecting a 10kV distributed photovoltaic power station that is operating normally within the region and has optical fiber direct actual power collection, using the actual power of the 10kV photovoltaic power station every 15 minutes, and fitting the total actual power of distributed photovoltaic according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 10kV )*P 10kV , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 10kV 、M 10kV They are respectively the actual power and installed capacity of the 10kV distributed photovoltaic.

[0011] Furthermore, the fitting scheme using the actual power data of the low-voltage distributed photovoltaic model station includes: selecting a low-voltage distributed photovoltaic power station that is operating normally and realizing actual power collection within the region as the low-voltage model station, using the actual power of this low-voltage model station every 15 minutes, and fitting the total actual power of distributed photovoltaic according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 样板站 )*P 样板站 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 样板站 、M 样板站 They are respectively the actual power generated and installed capacity of the low-voltage distributed photovoltaic model station.

[0012] Furthermore, the weighted average fitting scheme using three types of actual power data includes: selecting a centralized power station that operates normally and realizes actual power collection within the region, a 10kV distributed power station, and a low-voltage distributed photovoltaic model station to participate in the weighted average fitting calculation,

[0013] P=(G1×P 集中式 / M 集中式 +G2×P 10kV / M 10kV +G3×P 样板站 / M 样板站 ) / 3;

[0014] Among them, G1, G2, G3 represent weights; P 样板站、 M 样板站 For the actual launch and installation of low voltage sample station, P 集中式、 M 集中式 For centralized delivery and installation, P 10kV、 M 10kV The actual generation and installation of 10kV distributed photovoltaics.

[0015] Furthermore, a scheme for using distributed photovoltaic short-term power prediction results to replace the actual power data of distributed photovoltaics at 96 moments in a day includes: generating short-term power prediction data for 96 moments in a day through distributed photovoltaic short-term power prediction, and if the actual power data of distributed photovoltaics at 96 moments in a day is missing at any moment, the short-term power prediction data at the corresponding moment is used to replace the missing data.

[0016] Furthermore, a method for switching fitting schemes based on an adaptive PID algorithm includes: using a proportional control unit of a PID controller, with an error threshold set as a proportional value; when the proportion of the error fluctuation exceeds the error threshold, the proportional control unit switches the fitting scheme.

[0017] On the other hand, the present invention also proposes a distributed photovoltaic measured load data adaptive fitting device, comprising:

[0018] Solution module: Use the actual power data of centralized photovoltaic, the actual power data of 10kV distributed photovoltaic, the actual power data of low-voltage distributed photovoltaic model stations, and the weighted average of these three actual power data to set four fitting schemes for fitting the actual power data at 96 points in a day according to the installed capacity of distributed photovoltaic.

[0019] Threshold module: set the error threshold;

[0020] Adaptive module: Based on the adaptive PID algorithm, the fitting scheme switching control is performed, and the error between the distributed photovoltaic actual power data at the current time point and the previous time point obtained by fitting is compared with the error threshold. If the error threshold is exceeded, the four fitting schemes are switched in sequence, and the fitting of the distributed photovoltaic actual power data is continued according to the switched fitting scheme;

[0021] Measured load module: Calculates the measured load data based on the distributed photovoltaic actual power data obtained by fitting.

[0022] Furthermore, the scheme module also includes a fifth scheme, which uses distributed photovoltaic short-term power prediction results to replace the actual power data of distributed photovoltaic at 96 points in a day; the fifth scheme is added to the scheme switching sorting based on the adaptive PID algorithm as the last scheme in the sorting.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. When it is impossible to monitor the actual power data of distributed photovoltaic power generation in real time, the present invention proposes five distributed photovoltaic actual power generation fitting schemes to obtain stable measured load data through fitting.

[0025] 2. When the actual power data of the mapping source fluctuates violently, the present invention can use an adaptive PID control algorithm to select different fitting schemes and switch the mapping source to ensure that the distributed photovoltaic actual power generation does not undergo large jumps and is within an acceptable range.

[0026] 3. The solution of the present invention can improve the quality of measured load data by fitting under the condition of fluctuations in distributed photovoltaic operation, thereby improving the accuracy of grid load prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0029] The present invention is further described below in conjunction with specific embodiments and drawings.

[0030] In grid load monitoring, distributed photovoltaic power generation is an important component. However, its volatility during operation makes load monitoring difficult.

[0031] The design concept of the present invention is to provide stable distributed photovoltaic actual power data through data fitting. However, the actual power of the mapping source of data fitting may be affected by other factors, causing the fitted distributed photovoltaic actual power data to jump. Therefore, a more stable control method is also needed.

[0032] Embodiment 1:

[0033] The method proposed by the present invention is as follows Figure 1 The specific instructions are as follows:

[0034] First, the present invention proposes five fitting schemes:

[0035] The first solution is to use the centralized photovoltaic power generation to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaic power. The fitting solution is: select a centralized photovoltaic power station that is operating normally in the region and has direct optical fiber power generation, and use the actual power generation of the centralized photovoltaic power station every 15 minutes (i.e., 96 points in a day for 24 hours) to fit the total actual power generation of distributed photovoltaic power according to the proportion of installed capacity. The formula is P 分布式 =(M 分布式 / M 集中式 )*P 集中式 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 集中式 、M 集中式 They are respectively the actual generated power and installed capacity of the centralized photovoltaic system.

[0036] The second solution is to use the 10kV distributed photovoltaic power generation to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaic power generation. The fitting solution is: select a 10kV distributed photovoltaic power station that is operating normally in the region and has direct optical fiber power generation, and use the actual power generation of the 10kV photovoltaic power station every 15 minutes (i.e., 96 points in 24 hours a day) to fit the total actual power generation of distributed photovoltaic power generation according to the proportion of installed capacity. The formula is P 分布式 =(M 分布式 / M 10kV )*P 10kV , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 10kV 、M 10kV They are respectively the actual power and installed capacity of the 10kV distributed photovoltaic.

[0037] The third solution is to use the actual power generated by the low-voltage distributed photovoltaic model station to fit the actual power data at 96 points a day according to the total installed capacity of distributed photovoltaics. The fitting solution is: select a low-voltage distributed photovoltaic power station that is operating normally and realizes actual power collection within the region as a low-voltage model station, and use the actual power generated by this low-voltage model station every 15 minutes to fit the total actual power generated by distributed photovoltaics according to the proportion of installed capacity. The formula is P 分布式 =(M 分布式 / M 样板站 )*P 样板站 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 样板站 、M 样板站They are respectively the actual power generated and installed capacity of the low-voltage distributed photovoltaic model station.

[0038] The fourth scheme is to use the weighted average of centralized photovoltaic, 10kV distributed photovoltaic, and low-voltage distributed photovoltaic model stations to fit the actual power data at 96 points a day according to the total installed capacity of distributed photovoltaic. The fitting scheme is: select a centralized power station, a 10kV distributed power station, and a low-voltage distributed photovoltaic model station that are operating normally and realize actual power collection within the region to participate in the weighted average fitting calculation. The calculation formula is as follows:

[0039] P=(G1×P 集中式 / M 集中式 +G2×P 10kV / M 10kV +G3×P 样板站 / M 样板站 ) / 3;

[0040] Among them, G1, G2, G3 represent weights; P 样板站、 M 样板站 For the actual launch and installation of low voltage sample station, P 集中式、 M 集中式 For centralized delivery and installation, P 10kV、 M 10kV The actual generation and installation of 10kV distributed photovoltaics.

[0041] In this embodiment, the weights of the centralized power station, 10kV distributed power station, and low-voltage distributed photovoltaic model station are set to be the same, so G1, G2, and G3 can all be set to 1, and the calculation formula for the fitted power is as follows:

[0042] P=(P 集中式 / M 集中式 +P 10tV / M 10kv +P 样板站 / M 样板站 ) / 3.

[0043] Solution 5 is to use the distributed photovoltaic short-term prediction results to replace the distributed photovoltaic 96-point actual power data. The solution is: short-term power prediction for distributed photovoltaics will be generated every day, from 00:15 to 24:00 in the morning, with a granularity of 15 minutes, so there will be short-term power prediction data at 96 time points every day. If the distributed photovoltaic actual power data is missing at a certain time point, the prediction data of the corresponding time point in the short-term power prediction data can be found accordingly, and the predicted data value can be used to replace the actual power data value at that time.

[0044] Although the above five schemes can all be used for distributed photovoltaic data fitting, each of them has certain defects.

[0045] The actual power generated by centralized photovoltaic power generation may be unstable due to power restrictions and other reasons. For example, the actual power generated is 0. If fitting scheme 1 is used, the actual power generated by distributed photovoltaic power generation obtained by fitting will jump, which will cause the measured load to jump and affect the load forecasting results.

[0046] 10kV distributed photovoltaic and low-voltage distributed photovoltaic model stations may have unstable system actual power due to unstable access, for example, the actual power of the system may be 0. If fitting schemes 2 and 3 are used, the actual power of the distributed photovoltaic obtained by fitting will also jump, which will cause the measured load to jump and affect the load forecast results.

[0047] Plan 4 and Plan 5 are relatively stable, but these two plans require large amounts of computation, rely on multiple data sources, and are inconvenient to process, which are also factors that need to be considered.

[0048] Although Scheme 4 is relatively stable, if the required measured data is missing (i.e. the actual power data of 96 points in a day of centralized photovoltaic, 10kV distributed photovoltaic, and low-voltage distributed photovoltaic model stations are missing), the weighted average calculation cannot be performed. Therefore, the fitted distributed photovoltaic 96-point actual power data cannot be obtained, which affects the load forecasting results.

[0049] Although Scheme 5 is relatively stable, it uses short-term forecast data to replace actual data. If the data of other schemes that use actual power data for fitting are relatively stable, it is still preferred to use the stable actual power data fitting scheme.

[0050] Therefore, the present invention adds an adaptive PID control algorithm based on the above five solutions.

[0051] Based on the adaptive PID algorithm, the distributed photovoltaic actual power generation error is optimized and a threshold is set. The threshold can be set to a fixed value. For example, the threshold is set to 100,000 kilowatts. If the algorithm tracking finds that the error exceeds this value (that is, the actual power generation at this moment and the actual power generation at the previous moment jump by more than 100,000 kilowatts), the distributed photovoltaic actual power generation fitting scheme is switched through the controller.

[0052] The scheme switching is controlled by the adaptive PID algorithm, and the switching order is: Scheme 1--->Scheme 2--->Scheme 3--->Scheme 4--->Scheme 5. The sequential control ensures that the distributed photovoltaic power generation does not experience large jumps. For example, first use Scheme 1, and use the actual power generation of centralized photovoltaics to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaics; if the actual power generation of centralized photovoltaics may be unstable due to power restrictions and other reasons, switch to Scheme 2, and use the actual power generation of 10kV distributed photovoltaics to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaics; if the 10kV distributed photovoltaics have unstable access, resulting in unstable actual power generation of the system, switch to Scheme 3, and use the actual power generation of low-voltage distributed photovoltaic model stations to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaics; if the low-voltage distributed photovoltaic model station has unstable access, resulting in unstable actual power generation of the system, switch to Scheme 4, and use the weighted average of centralized photovoltaics, 10kV distributed photovoltaics, and low-voltage distributed photovoltaic model stations to fit the actual power data at 96 points in a day according to the total installed capacity of distributed photovoltaics; if the measured data used for calculation in Scheme 4 is missing, switch to Scheme 5, and use the short-term prediction results of distributed photovoltaics to replace the 96-point actual power data of distributed photovoltaics. During the use of Scheme 5, if the actual power data of the centralized photovoltaic system in Scheme 1 becomes stable again, switch back to Scheme 1.

[0053] After fitting the actual power data of distributed photovoltaics, the measured load of distributed photovoltaics is calculated using the following formula.

[0054] Measured load = net load = user demand load - actual distributed photovoltaic power generation.

[0055] The above is a solution for finally obtaining the measured load data of distributed photovoltaics based on the adaptive PID control algorithm, which can provide stable measured load data for the load prediction model that strongly relies on historical data, and is conducive to training the load prediction model using the measured data of historical loads.

[0056] Since load forecasting results are strongly correlated with historical loads, stable distributed photovoltaic power generation monitoring is conducive to improving the quality of measured load data, further improving load forecasting results, and providing strong data support for power grid technology formulation.

[0057] Embodiment 2:

[0058] In this embodiment 2, a distributed photovoltaic measured load data adaptive fitting device is proposed, including:

[0059] Solution module: Use the actual power data of centralized photovoltaic, 10kV distributed photovoltaic, and low-voltage distributed photovoltaic model station, and the weighted average setting of these three actual power data to fit the actual power data at 96 times a day according to the distributed photovoltaic installed capacity, and use the distributed photovoltaic short-term power prediction results to replace the distributed photovoltaic actual power data at 96 times a day.

[0060] Threshold module: set the error threshold;

[0061] Adaptive module: Based on the adaptive PID algorithm, the fitting scheme switching control is performed, and the error between the distributed photovoltaic actual power data at the current time point and the previous time point obtained by fitting is compared with the error threshold. If the error threshold is exceeded, the four fitting schemes are switched in sequence, and the fitting of the distributed photovoltaic actual power data is continued according to the switched fitting scheme;

[0062] Measured load module: Calculates the measured load data based on the distributed photovoltaic actual power data obtained by fitting.

[0063] The solution modules include:

[0064] Scheme 1 unit: Select a centralized photovoltaic power station that is operating normally in the region and has direct optical fiber power generation. Use the actual power generation of the centralized photovoltaic power station every 15 minutes and fit the total actual power generation of distributed photovoltaic power according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 集中式 )*P 集中式 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 集中式 、M 集中式 They are respectively the actual generated power and installed capacity of the centralized photovoltaic system.

[0065] Scheme 2 Unit: Select a 10kV distributed photovoltaic power station that is operating normally in the region and has direct optical fiber power generation. Use the actual power generation of the 10kV photovoltaic power station every 15 minutes and fit the total actual power generation of distributed photovoltaic power according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 10kV )*P 10kV , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 10kV 、M 10kV They are respectively the actual power and installed capacity of the 10kV distributed photovoltaic.

[0066] Unit 3: The fitting scheme using the actual power data of low-voltage distributed photovoltaic model stations includes: selecting a low-voltage distributed photovoltaic power station that operates normally within the region and realizes actual power collection as a low-voltage model station, using the actual power of this low-voltage model station every 15 minutes, and fitting the total actual power of distributed photovoltaics according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 样板站 )*P 样板站 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaic power generation, P 样板站 、M 样板站 They are respectively the actual power generated and installed capacity of the low-voltage distributed photovoltaic model station.

[0067] Unit 4: The weighted average fitting scheme using three types of actual power data includes: selecting a centralized power station that operates normally and realizes actual power collection within the region, a 10kV distributed power station, and a low-voltage distributed photovoltaic model station to participate in the weighted average fitting calculation,

[0068] P=(G1×P 集中式 / M 集中式 +G2×P 10kV / M 10kV +G3×P 样板站 / M 样板站 ) / 3;

[0069] Among them, G1, G2, G3 represent weights; P 样板站、 M 样板站 For the actual launch and installation of low voltage sample station, P 集中式、 M 集中式 For centralized delivery and installation, P 10kV、 M 10kV The actual generation and installation of 10kV distributed photovoltaics.

[0070] Unit 5 of Scheme: The scheme of using the distributed photovoltaic short-term power prediction results to replace the distributed photovoltaic actual power data at 96 moments in a day includes: generating short-term power prediction data at 96 moments in a day through distributed photovoltaic short-term power prediction, and if the actual power data of distributed photovoltaic at 96 moments in a day is missing at any moment, the short-term power prediction data at the corresponding moment is used to replace the missing data.

[0071] The adaptive module includes: a proportional control unit using a PID controller, and an error threshold is set as a proportional value; when the proportion of the error fluctuation exceeds the error threshold, the proportional control unit switches the fitting scheme.

[0072] The device described in this embodiment 2 can implement the distributed photovoltaic measured load data adaptive fitting method described in embodiment 1, and achieve the same technical effect as embodiment 1.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A distributed photovoltaic measured load data adaptive fitting method, characterized in that: include: Using the actual power data of centralized photovoltaic, the actual power data of 10kV distributed photovoltaic, the actual power data of low-voltage distributed photovoltaic model stations, and the weighted average of these three actual power data, four fitting schemes are set to fit the actual power data at 96 points in a day according to the installed capacity of distributed photovoltaic. Setting an error threshold; performing fitting scheme switching control based on an adaptive PID algorithm, comparing the error between the distributed photovoltaic actual power data at the current time point and the previous time point obtained by fitting with the error threshold, and switching the four fitting schemes in order if the error threshold is exceeded, and continuing to fit the distributed photovoltaic actual power data according to the switched fitting scheme; The measured load data is calculated based on the distributed photovoltaic actual power data obtained by fitting.

2. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: It also includes a fifth scheme, which uses the distributed photovoltaic short-term power prediction results to replace the distributed photovoltaic actual power data at 96 points in a day; the fifth scheme is added to the scheme switching sorting based on the adaptive PID algorithm as the last scheme in the sorting.

3. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: The fitting scheme using the actual power data of centralized photovoltaic power generation includes: selecting a centralized photovoltaic power station that is operating normally within the region and has direct optical fiber actual power generation, using the actual power generation of the centralized photovoltaic power station every 15 minutes, and fitting the total actual power generation of distributed photovoltaic power generation according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 集中式 )*P 集中式 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaics, P 集中式 、M 集中式 They are respectively the actual generated power and installed capacity of the centralized photovoltaic system.

4. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: The fitting scheme using 10kV distributed photovoltaic actual power data includes: selecting a 10kV distributed photovoltaic power station that is operating normally in the region and has optical fiber direct actual power, using the actual power of the 10kV photovoltaic power station every 15 minutes, and fitting the total actual power of distributed photovoltaic according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 10kV )*P 10kV , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaics, P 10kV 、M 10kV They are respectively the actual power and installed capacity of the 10kV distributed photovoltaic.

5. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: The fitting scheme using the actual power data of low-voltage distributed photovoltaic model stations includes: selecting a low-voltage distributed photovoltaic power station that operates normally within the region and realizes actual power collection as a low-voltage model station, using the actual power of this low-voltage model station every 15 minutes, and fitting the total actual power of distributed photovoltaics according to the installed capacity ratio. The formula is P 分布式 =(M 分布式 / M 样板站 )*P 样板站 , where P 分布式 、M 分布式 are the actual power and installed capacity of distributed photovoltaics, P 样板站 、M 样板站 They are respectively the actual power generated and installed capacity of the low-voltage distributed photovoltaic model station.

6. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: The weighted average fitting scheme using three types of actual power data includes: selecting a centralized power station that operates normally and realizes actual power collection within the region, a 10kV distributed power station, and a low-voltage distributed photovoltaic model station to participate in the weighted average fitting calculation. P=(G1×P 集中式 / M 集中式 +G2×P 10kV / M 10kV +G3×P 样板站 / M 样板站 ) / 3; Among them, G1, G2, G3 represent weights; P 样板站、 M 样板站 For the actual launch and installation of low voltage sample station, P 集中式、 M 集中式 For centralized delivery and installation, P 10kV、 M 10kV The actual generation and installation of 10kV distributed photovoltaics.

7. The distributed photovoltaic measured load data adaptive fitting method according to claim 2 is characterized in that: The scheme of using the distributed photovoltaic short-term power prediction results to replace the distributed photovoltaic actual power data at 96 moments in a day includes: generating short-term power prediction data at 96 moments in a day through distributed photovoltaic short-term power prediction, and if the actual power data of the distributed photovoltaic at 96 moments in a day is missing at any moment, the short-term power prediction data at the corresponding moment is used to replace the missing data.

8. The distributed photovoltaic measured load data adaptive fitting method according to claim 1 is characterized in that: The method for switching the fitting scheme based on the adaptive PID algorithm includes: using a proportional control unit of a PID controller, and setting an error threshold to a proportional value; when the proportion of the error fluctuation exceeds the error threshold, the proportional control unit switches the fitting scheme.

9. A distributed photovoltaic measured load data adaptive fitting device, characterized in that: include: Solution module: Use the actual power data of centralized photovoltaic, the actual power data of 10kV distributed photovoltaic, the actual power data of low-voltage distributed photovoltaic model stations, and the weighted average of these three actual power data to set four fitting schemes for fitting the actual power data at 96 points in a day according to the installed capacity of distributed photovoltaic. Threshold module: set the error threshold; Adaptive module: Based on the adaptive PID algorithm, the fitting scheme switching control is performed, and the error between the distributed photovoltaic actual power data at the current time point and the previous time point obtained by fitting is compared with the error threshold. If the error threshold is exceeded, the four fitting schemes are switched in sequence, and the fitting of the distributed photovoltaic actual power data is continued according to the switched fitting scheme; Measured load module: Calculates the measured load data based on the distributed photovoltaic actual power data obtained by fitting.

10. The distributed photovoltaic measured load data adaptive fitting device according to claim 9, characterized in that: The scheme module also includes a fifth scheme, which uses distributed photovoltaic short-term power prediction results to replace the actual power data of distributed photovoltaic at 96 points in a day; the fifth scheme is added to the scheme switching sorting based on the adaptive PID algorithm as the last scheme in the sorting.

Images