Methods for creating irradiance maps to predict output of solar farms and to orient solar trackers

Abstract

Systems and methods are described for using AI processing to translate sky images and other available data into irradiance maps to produce signal outputs for the control of single axis solar trackers and forecasted energy output of solar photovoltaic projects. The processes involve the collection of sky imagery and pyranometer data (local and / or remote), and segmentation processing of the information. The system collects and applies regional historical and geospatial data (geography, topography, etc.) relevant to the locale and the time of year to generate a dynamic shadow map and to provide quality predictive local and real time irradiance maps. The irradiance maps are then translated into control signals (digital and then analog) suitable for transmission and use to move and orient the single axis solar trackers within the solar collector field. The irradiance maps are also translated into forecasted energy output of the solar photovoltaic project.

Description

TITLEMETHODS FOR CREATING IRRADIANCE MAPS TO PREDICT OUTPUT OF SOLAR FARMS AND TO ORIENT SOLAR TRACKERSBACKGROUND OF THE INVENTION

[0001] The present invention relates generally to the efficient production of solar energy using predictive control of the orientation of single axis rotation solar panels. The invention can also be used as a standalone application to predict solar power output of an existing solar farm. The present invention relates more specifically to the use of artificial intelligence (Al) to translate sky images and other available data, both real time and historic, into irradiance maps to: (a) produce signal outputs for control of the motion and orientation of single axis solar trackers and associated solar panels; and (b) create look-ahead energy output forecast of solar photovoltaic (PV) projects.DESCRIPTION OF THE RELATED ART

[0002] Efforts have been made in the past to provide more efficient systems and methods for driving solar trackers to optimize production from solar collectors. Examples of such efforts include those described in U.S. Patent No.: 10,541,644, issued January 21, 2020, in the name of Arliaud et al., and U.S. Patent No.: 11,307,284, issued April 19, 2022, in the name of Arliaud et al., the full disclosures of which are incorporated herein by reference. Most of these prior efforts have fallen short of fully utilizing and processing the available information in a manner that truly optimizes power output.

[0003] It would be desirable, therefore, to have improved systems and methods that utilize artificial intelligence (Al) processing capabilities to more fully analyze and apply real time dataand historical information to accurately predict and direct solar tracker orientations to optimize solar energy production.SUMMARY OF THE INVENTION

[0004] The present invention provides systems and methods using artificial intelligence (Al) processing to translate sky images and other available data into irradiance maps to produce signal outputs for the control of the motion and orientation of single axis solar trackers and associated solar PV panels as well as predict the energy output of solar PV projects. The processes of the present invention involve the collection of sky imagery (local and remote) and pyranometer data (local and remote), and segmentation (SG) processing of this information within Al pattern recognition and prediction systems. In addition to the real time data collected, the system collects and applies historical and geospatial data relevant to the locale and the time of year. The present invention utilizes Al to quickly analyze real time sky data (digitally imaged and segmented) to combine and compare it with historical data with the goal of providing quality predictive local and real time irradiance maps (IR). The system further applies regional geospatial data (geography, topography, etc.) to generate a dynamic shadow map (SM) and a local, real time, and look-ahead irradiance map (IR). The irradiance map (IR) is then translated into control signals (digital and then analog) suitable for transmission and use within the particular type of solar collector field being monitored and controlled. The IR is also processed to create look-ahead energy output of the solar PV project.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Fig. 1 is a schematic block diagram showing the primary components of the system of the present invention implementing the methods of the present invention to optimize solar power collection.

[0006] Fig. 2 is a flowchart of the primary steps in the methods of the present invention to optimize solar power collection.

[0007] Fig. 3A is a raw sky image as may be used prior to processing to generate a shadow map according to the methods of the present invention.

[0008] Fig. 3B is the sky image of Fig. 3A processed with Al implemented segmentation (SG) as may be used to generate a shadow map according to the methods of the present invention.

[0009] Fig. 4 is an example of a global horizontal solar irradiance map of the type generated and used according to the methods of the present invention.

[0010] Fig. 5A is a raw cloud shadow image characterizing the process of generating a digital shadow map according to the methods of the present invention.

[0011] Fig. 5B is a digital shadow map of the type generated according to the methods of the present invention.

[0012] Figs. 6A - 6D are schematic perspective views of the four basic types of single axis solar tracker associated solar panels of the type operable in conjunction with the systems and methods of the present invention.DETAILED DESCRIPTION

[0013] Reference is made first to Fig. 1 which is a schematic block diagram showing the basic elements and components of the system of the present invention structured to implement the methods of the present invention to optimize solar power collection. The system is centered around artificial intelligence (Al) processor 100 where the relevant information and data are collected and the method steps of the present invention are carried out. Al processor 100 receives data from local sky camera 102 paired with local pyranometer 104, and optionally from one or more additional remote sky cameras 102n paired with remote pyranometers 104n. Additionally, Al processor 100 has access to historical data 106 such as from the NationalRenewable Energy Laboratory (NREL) and other publicly available sources. Further, Al processor 100 has access to geospatial data 108 such as satellite, mapping, local and remote weather data, and topographical imagery and data from a wide range of public and private sources. Finally, Al processor 100 has access to weather data 107 such as from the National Weather Service (NWS) and other publicly available sources.

[0014] The basic process steps of the method of the present invention (described in more detail below with respect to Fig. 2) involve generation of the irradiance map (IR) which occurs within Al processor 100. Initially, the raw data associated with the sky imagery undergoes segmentation processing (SG) 110. This is followed by generation of shadow map (SM) 112 and then with generation of a local and real time irradiance map (IR) 113. In some cases, the shadow map is not created, and the irradiance map is created directly if sky conditions permit. The output of processor 100 provides the control signal information that ultimately drives the solar trackers for the local field. Typically, this involves conversion or translation of the local and real time irradiance map 113 to control signals through digital signal controller 116. These control signals are translated into analog control signals through analog signal controller 118 that regulates and switches the power that drives the solar tracker orientation motors at the single axis solar trackers 120. On a parallel path within Al processor 100, the Al is able to produce forecasted pictures of the sky for the next time intervals. These pictures are then used to create forecasted look-ahead irradiance maps. Those irradiance maps are then fed, along with weather data, project structure (PS) information 114 (project layout, project technical details, and PV panel characteristics), into the Al processor that will determine the forecasted project energy output 115 for the next time intervals.

[0015] Fig. 2 is a flowchart of the primary steps in the methods of the present invention to optimize solar power collection. The broad overall operational method begins at Step 130 which initiates the single axis solar tracker control processing. Step 132 involves the collectionof sky imagery (local and optionally remote) and pyranometer data (local and optionally remote). In addition to the real time data collected as described above, the system collects and applies historical data (such as from NREL) relevant to the locale and the time of year at Step 134. Processing begins at Step 136 with segmentation (SG) processing within the Al pattern recognition systems. As with predictive weather forecasting, the present system utilizes Al to quickly analyze real time sky data (digitally imaged and segmented) and combine / compare it with historical data to provide quality predictive local and real time irradiance information to produce shape of the sky for the next time intervals at Step 138. In conjunction with Step 136, method Step 140 applies regional geospatial data (geography, topography, infrastructure, etc.) to generate a dynamic shadow map (SM) and a local and real time irradiance map (IR), again with Al processing. Once again, in some cases, shadow maps are not created and only irradiance maps are created.

[0016] Step 142 then translates the irradiance map (IR) into a digital control signal suitable for transmission to the solar collector field. This control signal is used to orient solar tracker such that it captures more light from the sky instead of rotating the face of the panels in a direction perpendicular to the sun. Step 144 translates the digital control signal to operational analog control of the solar trackers positioned on the collector panels in the field. The overall process continues (repeats) with control processing during a daylight cycle at conclusion Step 150.

[0017] Parallel with the field control signal path (Steps 142 & 144), the energy output forecast path follows from Step 140 and involves incorporating weather data and project structure (project layout and technical specifications including PV characteristics) at Step 146 into the irradiance maps generated. This then results, at Step 148, in the generation of the project energy output forecast that is a key element of the present invention.

[0018] Reference is next made to Fig. 3A which is a raw sky image as may be used prior to processing to generate a shadow map according to the methods of the present invention. Thebasic elements of the image are direct sun 10, radiant sky 12, and cloud cover 14. Each of the basic elements of the sky image may provide varying degrees of radiant energy which is why the sky image is typically paired with one or more pyranometers. A typical sky camera - pyranometer combination is disclosed in Patent Application Publication No.: US 2023 / 0160745 Al, published May 25, 2023, in the name of Blum et al., the full disclosure of which is incorporated herein by reference.

[0019] Fig. 3B is the sky image of Fig. 3A processed with Al implemented segmentation (SG) as may be used to generate a shadow map and an irradiance map according to the methods of the present invention. The digital pattern recognition processing embeds additional relevant data in the segmented image by isolating parts of the image that provide the same or similar radiant energy. While direct sun 10 and radiant clear sky 12 may be easily discerned and quantified, differences in the transmission and / or occultation properties of cloud cover require more complex (Al) processing to be appropriately segmented to generate a more accurate dynamic shadow and irradiance maps. In the example shown, proximity to direct sun 10 as well as discrete cloud densities give rise to variations in transmitted (radiant) energy by various sky segments designate 18a - 18e, as examples, in Fig. 3B.

[0020] Fig. 4 is an example of a global horizontal solar irradiance map of the type generated and used according to the methods of the present invention. The map in Fig. 4 provides a portion of the United States showing a monthly (March in this example) average daily total solar resource using 1998 - 2016 data (PSM v3) with 0.038-degree latitude by 0.038-degree longitude (nominally 4 km x 4 km) resolution. The color gradients are referenced on the index in Fig. 4 and represent kWh / m2 / Day. This map and the data associated with it are typical of the historical data provided by the National Renewable Energy Laboratory (NREL) and which are publicly available through the NREL website ( el.gov). Even though the Al in the presentinvention generates the global horizontal irradiance, it can also predict its two components, namely direct normal irradiance and diffuse irradiance.

[0021] Fig. 5A is a raw cloud shadow image indicative of the process of generating a digital shadow map according to the methods of the present invention. The view of Fig. 5A, namely from above the clouds, characterizes the basic intended structure of a shadow map of the type utilized and / or generated by the process of the present invention. What the image shows are the dramatic differences that can exist between direct sky radiance on the ground surface (and therefore on solar collectors) and cloud covered radiance on the ground surface.

[0022] Fig. 5B is a digital shadow map of the type generated and used according to the methods of the present invention. The shadow map (SM) of Fig. 5B incorporates topographical elements 30 which can produce topography shadows 36 separate from cloud shadows 32. Irradiated (non-shadow) areas 34 are also depicted and digitally represented in the generated shadow map (SM) shown in Fig. 5B.

[0023] Figs. 6A - 6D are schematic perspective views of the four basic types of single axis solar tracker associated solar panels of the type operable in conjunction with the systems and methods of the present invention. Single axis solar trackers continue to be the most cost- effective means for optimizing solar collector production and when operational control relies on accurate and predictive irradiance information can provide improved power generation from a solar collection field. Fig. 6A discloses a single axis solar tracking system where the panel pivots on a horizontal axis-oriented north - south with the rotation of the panel generally tracking east to west by rotating the panel about the north - south axis. Fig. 6B discloses a single axis solar tracking system where the panel pivots on a horizontal axis-oriented east - west with the rotation of the panel generally tracking north to south by rotating the panel about the east - west axis.

[0024] Fig. 6C discloses a single axis solar tracking system where the panel pivots on an upright vertical axis directed through the center of the panel with the rotation generally tracking through all compass points with a fixed angle of inclination to the vertical. Fig. 6D discloses a single axis solar tracking system where the panel pivots on an inclined axis generally oriented north to south with the rotation generally tracking east - west at a fixed angle of inclination to the horizontal.

[0025] Further examples of single axis solar trackers appropriate for use with the systems and methods of the present invention include those described in U.S. Patent No.: 8,459,249, issued June 11, 2013, in the name of Corio, and U.S. Patent No.: 9,631,840, issued April 25, 2017, in the name of Corio, the full disclosures of which are incorporated herein by reference.

[0026] Although the present invention has been described in conjunction with a number of preferred embodiments, those skilled in the art will recognize modifications to these embodiments that still fall within the scope of the present invention. Because of the variety of different single axis solar trackers, the control signal output of the system will vary according to the electrical and electronic control parameters of the specific trackers and collectors. Those skilled in the art will recognize the process of translating the irradiance maps into the control signals used to direct the specific types of solar trackers being used.

Claims

CLAIMS1. A system for producing real time and predictive irradiance maps to forecast energy output and to optimize control of single axis solar trackers associated with a field of solar collectors, the system comprising: at least one locally positioned sky camera paired with an associated pyranometer for producing sky images; an artificial intelligence (Al) processor having communication access to the sky images and to a database of historical solar irradiance information and a database of geospatial information relevant to a location associated with the field of solar collectors, the Al processor operable to carry out segmentation processing of the sky images to produce one or more shadow maps and to translate the one or more shadow maps into one or more irradiance maps; a signal controller for translating the one or more irradiance maps into control signals appropriate for directing motions and orientations of the solar collectors by way of the single axis solar trackers.

2. The system of Claim 1 further comprising at least one remotely positioned sky camera paired with an associated pyranometer and wherein the Al processor further has communication access to the remote sky images.

3. The system of Claim 1 wherein the Al processor further has communication access to local and remote weather data.

4. The system of Claim 1 wherein the Al processor further has access to a project structure, the project structure comprising the configuration of single axis solar trackers associated with the field of solar collectors.

5. The system of Claim 1 wherein the signal controller comprises one or more digital signal controllers driving one or more analog signal controllers, which in turn drive the single axis solar trackers.

6. A method for producing real time and predictive irradiance maps to optimize control of single axis solar trackers associated with a field of solar collectors, the method comprising the steps of: collecting sky imaging and pyranometer data in at least one location adjacent the field of solar collectors; carrying out segmentation processing of the sky imaging and pyranometer data using Al pattern recognition processing; collecting and applying historical irradiance data associated with the location of the field of solar collectors to the sky imaging and pyranometer data; collecting and applying regional geospatial data to the sky imaging and pyranometer data and generating one or more shadow maps; translating the one or more shadow maps into one or more irradiance maps using Al processing; and translating the one or more irradiance maps into control signals suitable for directing the operation of the single axis solar trackers for moving and orienting the solar collectors within the field.

7. The method of Claim 6 further comprising the steps of: collecting sky imaging and pyranometer data in at least one location remote from the field of solar collectors.

8. The method of Claim 6 further comprising, subsequent to the step of carrying out segmentation processing, the step of producing shape of the sky for next time interval data.

9. The method of Claim 6 further comprising, concurrent with the step of translating the one or more irradiance maps, the step of incorporating weather data, project layout and technical specifications with the irradiance map data to produce project energy output forecasts.

10. The method of Claim 9 further comprising the step of repeating the control processing and energy output forecasting throughout a twenty four hour solar cycle.

Images