Photovoltaic equipment optimization direction prediction method and device based on data learning

By constructing a two-factor learning curve model, predicting the development trend of photovoltaic technology, the accuracy of the average electrification cost of photovoltaic equipment and the prediction of energy recovery period in the existing technology is solved, and a scientific reference for the optimization direction of photovoltaic equipment is provided.

CN118886594APending Publication Date: 2024-11-01SHANGHAI JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410871934.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-01
Publication Date
2024-11-01

AI Technical Summary

Technical Problem

The prior art is difficult to accurately predict the average electrification cost and energy recovery period of photovoltaic equipment, which affects the development reliability and equipment planning of photovoltaic technology.

Method used

Using a data learning method, a two-factor learning curve model is constructed, and the photovoltaic installation volume and raw material price are fitted, the learning coefficient and cost change rate are obtained, the development trend of photovoltaic technology is predicted, and the average is calculated based on photoelectric conversion efficiency and component life. Electrolysis cost and energy recovery period.

Benefits of technology

Accurate prediction of photovoltaic technology is achieved, reference for optimizing the direction of equipment, and improving the scientificity and accuracy of photovoltaic equipment planning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118886594A_ABST
    Figure CN118886594A_ABST
Patent Text Reader

Abstract

The invention relates to a photovoltaic equipment optimization direction prediction method and device based on data learning, and the method comprises the steps: obtaining photovoltaic data corresponding to different photovoltaic technologies, constructing a photovoltaic database, and enabling the photovoltaic data to comprise the photovoltaic module cost, the photovoltaic installation capacity and the raw material price; constructing a two-factor learning curve model, and performing model fitting according to the photovoltaic database to obtain prediction curves of different photovoltaic technologies; the two-factor learning curve model calculates the cost of the photovoltaic module according to the photovoltaic installed capacity, the raw material price and the cost of the photovoltaic module in the initial year, sets a learning coefficient for the photovoltaic installed capacity, and sets a raw material cost change rate for the raw material price; and predicting the photovoltaic efficiency according to the obtained prediction curves of different photovoltaic technologies, and determining the optimal photovoltaic technology, thereby adjusting the optimization direction of the photovoltaic equipment. Compared with the prior art, the development trend of various photovoltaic technologies can be predicted, and reference is provided for determining the optimization direction of photovoltaic equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic equipment optimization, and in particular to a method and device for predicting the optimization direction of photovoltaic equipment based on data learning. Background Art

[0002] Photovoltaic power generation has been widely used as a green and renewable energy source. Various photovoltaic technologies have developed rapidly: traditional solar cells represented by crystalline silicon account for a large proportion due to their rich materials, mature industry and high energy, and now account for more than 95% of the market share; at the same time, emerging photovoltaic technologies are also rapidly advancing the industrialization process, especially perovskite technology, which has strong commercial potential due to its low material cost, can produce economically feasible photovoltaic devices and scalable manufacturing.

[0003] Different photovoltaic equipment configurations, fuel costs, etc. will affect the average cost of electrification and energy payback period of photovoltaic technology. Predicting the average cost of electrification and energy payback period of photovoltaic technology can not only ensure the reliability of the development of photovoltaic technology, but also help with the rational planning of photovoltaic equipment. It has important practical significance and guiding significance for promoting the healthy and stable development of the photovoltaic power generation industry. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method and device for predicting the optimization direction of photovoltaic equipment based on data learning, to predict the development trend of various photovoltaic technologies, and to provide a reference for determining the optimization direction of photovoltaic equipment.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A photovoltaic device optimization direction prediction method based on data learning includes the following steps:

[0007] Obtain photovoltaic data corresponding to different photovoltaic technologies and build a photovoltaic database, wherein the photovoltaic data includes photovoltaic module costs, photovoltaic installed capacity, and raw material prices;

[0008] A two-factor learning curve model is constructed, and the model is fitted according to the photovoltaic database to obtain prediction curves of different photovoltaic technologies; the two-factor learning curve model calculates the photovoltaic module cost according to the photovoltaic installed capacity, raw material prices and the photovoltaic module cost in the initial year, and sets a learning coefficient for the photovoltaic installed capacity and a raw material cost change rate for the raw material price;

[0009] The photovoltaic efficiency is predicted based on the prediction curves of different photovoltaic technologies, and the optimal photovoltaic technology is determined, thereby adjusting the optimization direction of the photovoltaic equipment.

[0010] Furthermore, the expression of the two-factor learning curve model is:

[0011] LqCy ijt =lnS ij +β ij LqCy tj +γ ij oeLh tj

[0012] In the formula, i is the region name, t is the tth year, j is the photovoltaic technology route, R ijt is the module cost of photovoltaic technology j in region i in year t, S ij is the initial year cost of photovoltaic modules for photovoltaic technology j in region i, C tj is the installed capacity of photovoltaic technology j in year t, P tj is the raw material price of photovoltaic technology j in year t, β ij is the learning coefficient of photovoltaic technology j in region i, γ ij is the rate of change of raw material cost of PV technology j in region i;

[0013] The two-factor learning curve model is used to fit and solve the learning coefficient β ij and the rate of change γ ij , thus obtaining the prediction curve of photovoltaic technology.

[0014] Furthermore, a piecewise fitting method is used in the model fitting process, and the piecewise fitting method is specifically as follows:

[0015] If the slope change threshold of the fitting result is greater than the preset slope threshold, the fitting area is segmented according to the position where the slope exceeds the slope change threshold, and the corresponding learning coefficient β is fitted for different segmented areas respectively. ij and the rate of change γ ij .

[0016] Furthermore, the method further comprises: ij The learning rate corresponding to photovoltaic technology j is calculated to evaluate the photovoltaic efficiency change trend of photovoltaic technology j.

[0017] Furthermore, the learning rate is calculated as follows:

[0018]

[0019] Where, L ij is the learning rate of PV technology j in region i.

[0020] Furthermore, the method also includes predicting the future installed capacity based on the photovoltaic data, thereby predicting the photovoltaic equipment installation demand and adjusting the manufacturing scale of the photovoltaic equipment. The prediction calculation expression of the future installed capacity is:

[0021] C t+1 =C t *(1+rate)

[0022]

[0023] In the formula, C t+1 is the future installed capacity in year t+1, C t is the installed capacity in the tth year, rate is the annual growth rate, year is the number of statistical years, and ∑rate_h is the sum of the annual growth rates in the past.

[0024] Furthermore, the method further includes calculating an inter-regional impact parameter to evaluate the intercommunication level of photovoltaic devices between regions, and the calculation expression of the inter-regional impact parameter is:

[0025] C i(t+1) -C it =k it (C it -C i(t-1) )+(1-λ t )(C i'(t+1) -C i't )

[0026] In the formula, C i(t+1) is the installed capacity of region i in year t+1, C it is the installed capacity of region i in year t, C i(t-1) is the installed capacity of region i in year t-1, λ t is the inter-regional impact parameter in year t, C i'(t+1) is the installed capacity in year t+1 except for region i, C i't is the installed capacity in year t excluding region i.

[0027] Further, the photovoltaic efficiency is predicted according to the obtained prediction curves of different photovoltaic technologies: the average electrification cost and the energy payback period are calculated according to the obtained prediction curves of different photovoltaic technologies, and the calculation expression of the average electrification cost is:

[0028]

[0029] Where, LCOE is the average electrification cost;

[0030] The calculation expression of the energy payback period is:

[0031]

[0032] Where EPBT is the energy payback period.

[0033] Furthermore, the photovoltaic data in the photovoltaic database also includes photoelectric conversion efficiency and component life, and the calculation process of the average electrification cost and energy payback period is based on the photoelectric conversion efficiency and component life.

[0034] The present invention also provides a photovoltaic equipment optimization direction prediction device based on data learning, comprising a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of the method described above.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] (1) The present invention predicts the efficiency of various photovoltaic technologies currently produced. By acquiring photovoltaic data of various photovoltaic technologies, a dual-factor learning curve model is established. The dual-factor learning curve model sets a learning coefficient according to the photovoltaic installed capacity, which can reflect the development trend of photovoltaic technology; sets the raw material cost change rate according to the raw material price, and predicts the raw material cost to avoid the influence of the raw material cost on the photovoltaic technology learning coefficient. The learning coefficient obtained in this way can accurately reflect the development trend of the corresponding photovoltaic technology and more accurately predict the photovoltaic efficiency, thereby providing a reference for determining the optimization direction of photovoltaic equipment.

[0037] (2) The present invention is based on actual conditions and takes into account the impact of possible industrial changes that may occur during the photovoltaic technology solution process on the calculation results of the photovoltaic technology learning coefficient of this solution. The learning coefficient is segmented according to the slope of change, and the corresponding learning coefficients are fitted for different segmented areas to suit the actual industrial development.

[0038] (3) The present invention also solves and calculates the influencing parameters between regions, which reflects the mutual acceptance degree of photovoltaic technologies between different regions. It can quantify the impact of the mutual influence of photovoltaic policies in various regions on the current region, and judge the development trend of various photovoltaic technologies in a more scientific and practical way.

[0039] (4) The present invention constructs a complete database of various photovoltaic technology routes, collects and integrates various parameters of the entire life cycle of various photovoltaic technologies, such as PCE (photovoltaic conversion efficiency), LTs (module life), etc., and calculates the LCOE (average cost of electrification) and EPBT (energy payback period) corresponding to various photovoltaic technologies based on this, providing an effective means to achieve performance prediction of photovoltaic technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1A schematic flow chart of a photovoltaic device optimization direction prediction method based on data learning provided in an embodiment of the present invention;

[0041] Figure 2 A schematic diagram of a flexible perovskite yield-cost curve provided in an embodiment of the present invention;

[0042] Figure 3 A schematic diagram of a logarithmic learning curve of a rigid perovskite provided in an embodiment of the present invention;

[0043] Figure 4 A schematic diagram of a logarithmic learning curve of region A provided in an embodiment of the present invention;

[0044] Figure 5 Schematic diagram of a logarithmic learning curve of region B provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0046] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0048] Example 1

[0049] like Figure 1 As shown, this embodiment provides a photovoltaic device optimization direction prediction method based on data learning, comprising the following steps:

[0050] S1: Obtain photovoltaic data corresponding to different photovoltaic technologies and build a photovoltaic database. The photovoltaic data includes photovoltaic module costs, photovoltaic installed capacity, and raw material prices;

[0051] S2: Construct a two-factor learning curve model, and fit the model according to the photovoltaic database to obtain the prediction curves of different photovoltaic technologies; the two-factor learning curve model calculates the photovoltaic module cost according to the photovoltaic installed capacity, raw material prices and the photovoltaic module cost in the initial year, and sets the learning coefficient for the photovoltaic installed capacity and the raw material cost change rate for the raw material price;

[0052] S3: Predict photovoltaic efficiency based on the prediction curves of different photovoltaic technologies, determine the optimal photovoltaic technology, and thus adjust the optimization direction of the photovoltaic equipment.

[0053] The following is a detailed description of each of the above steps:

[0054] The process of acquiring photovoltaic data in step S1 is specifically as follows:

[0055] Collect various parameters related to photovoltaic technology, including PCE (photovoltaic conversion efficiency), LTS (module life), cost, installed capacity, maintenance costs, etc., adopt multiple data storage methods such as relational database, metadata database, distributed file storage, etc., and build a complete photovoltaic database through mining public databases on the Internet and standardized testing of samples.

[0056] Optionally, search for a large number of papers and patents related to photovoltaic technology, extract the viewpoints and innovations described therein, and use hash storage data structure to classify and integrate the collected literature, so that users can combine the literature content to set various possible future scenarios in subsequent research.

[0057] Optionally, for data storage, cloud computing architecture can be used to distribute and store massive amounts of different types of data.

[0058] For data display, you can use R language to write code to effectively visualize and intuitively display the data in the database, allowing users to find the data they need more conveniently and quickly.

[0059] In step S2, the process of constructing the two-factor learning curve model is specifically as follows:

[0060] For the traditional photovoltaic technology route, considering that it already has quite complete historical data, the corresponding two-factor learning curve model is expressed as:

[0061] LqCy ijt =lnS ij +β ij LqCy tj +γ ij oeLh tj

[0062] In the formula, i is the region name, t is the tth year, j is the photovoltaic technology route, Rijt is the module cost of photovoltaic technology j in region i in year t, S ij is the initial year cost of photovoltaic modules for photovoltaic technology j in region i, C tj is the installed capacity of photovoltaic technology j in year t, P tj is the raw material price of photovoltaic technology j in year t, β ij is the learning coefficient of photovoltaic technology j in region i, γ ij is the rate of change of raw material cost of PV technology j in region i;

[0063] like Figure 2 As shown, it is the yield-cost curve of flexible perovskite, which is not logarithmic. Figure 3 As shown, it is the linear fitting result after logarithmic processing of the data, and the logarithmic learning curve of the rigid perovskite is obtained.

[0064] The two-factor learning curve model is used to obtain the learning coefficient β by log-linear fitting. ij and the rate of change γ ij , thus obtaining the prediction curve of photovoltaic technology.

[0065] Considering that the photovoltaic industry will usher in certain industrial changes in the process of development, the corresponding learning coefficient will also change to a certain extent after the change. Therefore, the present invention adopts a segmented fitting method. For the section where the slope changes significantly, a new learning coefficient will be selected for fitting. That is, if the slope change threshold of the fitting result is greater than the preset slope threshold, the fitting area is segmented according to the position where the slope exceeds the slope change threshold, and the corresponding learning coefficient β is fitted for different segmented areas. ij and the rate of change γ ij .

[0066] like Figure 4 and Figure 5 As shown in the figure, the photovoltaic logarithmic learning curves of the two regions are shown. Each data point corresponds to a year. It can be seen from the figure that there are turning points with large changes in slope, which may be due to industrial changes. Therefore, the photovoltaic logarithmic learning curves are fitted in two sections to achieve more accurate prediction.

[0067] Then according to the learning coefficient β obtained by fitting ij The learning rate corresponding to photovoltaic technology j is calculated to evaluate the photovoltaic efficiency change trend of photovoltaic technology j.

[0068] The learning rate calculation expression is:

[0069]

[0070] Where, L ijis the learning rate of photovoltaic technology j in region i. The slope of the photovoltaic technology route is calculated, and the visualization of the existing data is completed.

[0071] In this embodiment, the prediction end year is 2025. In the model construction, this embodiment assumes that each type of photovoltaic technology route has a certain share instead of being directly eliminated by the market.

[0072] As a preferred implementation, this embodiment also predicts the future installed capacity, thereby predicting the demand for photovoltaic equipment installation and adjusting the manufacturing scale of photovoltaic equipment.

[0073] Without considering the impact of regional policies on photovoltaic technology, directly use:

[0074] C t+1 =C t *(1+rate)

[0075]

[0076] In the formula, C t+1 is the future installed capacity in year t+1, C t is the installed capacity in the tth year, rate is the annual growth rate, year is the number of statistical years, and ∑rate_h is the sum of the annual growth rates in the past.

[0077] When considering the impact of regional policies on various photovoltaic technologies, this embodiment introduces an inter-regional impact parameter λ to characterize the scenario conditions under the influence of regional policies:

[0078] C i(t+1) -C it =k it (C it -C i(t-1) )+(1-λ t )(C i'(t+1) -C i't )

[0079] In the formula, C i(t+1) is the installed capacity of region i in year t+1, C it is the installed capacity of region i in year t, C i(t-1) is the installed capacity of region i in year t-1, λ t is the inter-regional impact parameter in year t, C i'(t+1) is the installed capacity in year t+1 except for region i, C i't is the installed capacity in year t excluding region i.

[0080] The value of λ is a value between 0 and 1 input by the user based on the specific scenario selected. The higher the value of λ is, the greater the obstacle of the corresponding regional policy to the development of the photovoltaic industry.

[0081] For the traditional photovoltaic technology route, considering that its technological development process has tended to be flat, this embodiment will directly use its learning rate at the beginning of the prediction year as its future learning rate and raw material cost and its cost change rate for prediction, and apply a two-factor model to predict its future component cost.

[0082] For the emerging photovoltaic technology routes, considering that there is not enough data for this program to directly obtain a complete, scientific and accurate learning curve model, this program will adopt the scenario preset by the user according to their own research needs, and give the learning rate obtained from the theoretical analysis based on the analysis of the preset scenario, and substitute it into the formula for calculation. At the same time, considering that the emerging photovoltaic route does not have a mature industrialization compared to the traditional photovoltaic technology route, this program will preset a time point as the time point for the industrialization of the technology. Correspondingly, with this point as the dividing point, the learning rate will change to a certain extent, and the learning rate at this preset point and thereafter will also be completed by the user in the interactive interface.

[0083] After obtaining the complete learning curve, this solution will use LCOE (average electrification cost) and EPBT (energy payback period) as the new vertical axis to draw two more curves:

[0084]

[0085] Where LCOE is the average electrification cost;

[0086]

[0087] Where EPBT is the energy payback period.

[0088] By conducting in-depth analysis of the intersection points in each graph, users can clearly observe the optimization direction of photovoltaic technology.

[0089] This embodiment also provides a photovoltaic equipment optimization direction prediction device based on data learning, including a memory and a processor, the memory stores a computer program, and the processor calls the computer program to execute the steps of the photovoltaic equipment optimization direction prediction method based on data learning as described above.

[0090] It should be noted that the specific content and beneficial effects of the device of the present application can be found in the above-mentioned method embodiment, which will not be repeated here.

[0091] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A photovoltaic equipment optimization direction prediction method based on data learning, characterized in that: The following steps are involved: Obtain photovoltaic data corresponding to different photovoltaic technologies and build a photovoltaic database, wherein the photovoltaic data includes photovoltaic module costs, photovoltaic installed capacity, and raw material prices; A two-factor learning curve model is constructed, and the model is fitted according to the photovoltaic database to obtain prediction curves of different photovoltaic technologies; the two-factor learning curve model calculates the photovoltaic module cost according to the photovoltaic installed capacity, raw material prices and the photovoltaic module cost in the initial year, and sets a learning coefficient for the photovoltaic installed capacity and a raw material cost change rate for the raw material price; The photovoltaic efficiency is predicted based on the prediction curves of different photovoltaic technologies, and the optimal photovoltaic technology is determined, thereby adjusting the optimization direction of the photovoltaic equipment.

2. A photovoltaic equipment optimization direction prediction method based on data learning according to claim 1, characterized in that: The expression of the two-factor learning curve model is: lnR ijt =lnS ij +b ij lnC tj +g ij lnP tj In the formula, i is the region name, t is the tth year, j is the photovoltaic technology route, R ijt is the module cost of photovoltaic technology j in region i in year t, S ij is the initial year cost of photovoltaic modules for photovoltaic technology j in region i, C tj is the installed capacity of photovoltaic technology j in year t, P tj is the raw material price of photovoltaic technology j in year t, β ij is the learning coefficient of photovoltaic technology j in region i, γ ij is the rate of change of raw material cost of PV technology j in region i; The two-factor learning curve model is used to fit and solve the learning coefficient β ij and the rate of change γ ij , thus obtaining the prediction curve of photovoltaic technology.

3. A photovoltaic equipment optimization direction prediction method based on data learning according to claim 2, characterized in that: The model fitting process adopts a piecewise fitting method, which is specifically as follows: If the slope change threshold of the fitting result is greater than the preset slope threshold, the fitting area is segmented according to the position where the slope exceeds the slope change threshold, and the corresponding learning coefficient β is fitted for different segmented areas respectively. ij and the rate of change γ ij .

4. The photovoltaic equipment optimization direction prediction method based on data learning according to claim 2 is characterized in that: The method also includes a learning coefficient β obtained by fitting. ij The learning rate corresponding to photovoltaic technology j is calculated to evaluate the photovoltaic efficiency change trend of photovoltaic technology j.

5. A photovoltaic equipment optimization direction prediction method based on data learning according to claim 4, characterized in that: The learning rate calculation expression is: Where, L ij is the learning rate of PV technology j in region i.

6. The photovoltaic equipment optimization direction prediction method based on data learning according to claim 1 is characterized in that: The method further includes predicting the future installed capacity based on the photovoltaic data, thereby predicting the photovoltaic equipment installation demand and adjusting the manufacturing scale of the photovoltaic equipment. The prediction calculation expression of the future installed capacity is: C t+1 =C t *(1+rate) In the formula, C t+1 is the future installed capacity in year t+1, C t is the installed capacity in the tth year, rate is the annual growth rate, year is the number of statistical years, and ∑rate_h is the sum of the annual growth rates in the past.

7. The photovoltaic equipment optimization direction prediction method based on data learning according to claim 1 is characterized in that: The method further includes calculating an inter-regional impact parameter to evaluate the intercommunication level of photovoltaic devices between regions. The calculation expression of the inter-regional impact parameter is: C i(t+1) -C it =k it (C it -C i(t-1) )+(1-λ t )(C i'(t+1) -C i't ) In the formula, C i(t+1) is the installed capacity of region i in year t+1, C it is the installed capacity of region i in year t, C i(t-1) is the installed capacity of region i in year t-1, λ t is the inter-regional impact parameter in year t, C i'(t+1) is the installed capacity in year t+1 except for region i, C i't is the installed capacity in year t excluding region i.

8. The photovoltaic equipment optimization direction prediction method based on data learning according to claim 1 is characterized in that: The photovoltaic efficiency is predicted according to the prediction curves of different photovoltaic technologies obtained. Specifically, the average electrification cost and energy payback period are calculated according to the prediction curves of different photovoltaic technologies obtained. The calculation expression of the average electrification cost is: Where LCOE is the average electrification cost; The calculation expression of the energy payback period is: Where EPBT is the energy payback period.

9. The photovoltaic equipment optimization direction prediction method based on data learning according to claim 8, characterized in that: The photovoltaic data in the photovoltaic database also includes photoelectric conversion efficiency and component life. The calculation process of the average electrification cost and energy payback period is based on the photoelectric conversion efficiency and component life.

10. A photovoltaic equipment optimization direction prediction device based on data learning, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of any one of the methods according to claims 1 to 9.