System and method for atmospheric observations

Abstract

The present invention discloses a system and method for atmospheric observations. The system (100) comprises a sky observer (102) comprising a sky imager (104) to capture and communicate a sky image for analyzing cloud characteristics, a robotic sun photometer (106) to track and communicate the sun and measure aerosol properties, a pyranometer (108) with a shadowband (110) and a sensor unit (120) to measure environmental conditions and collect and communicate sensor data surrounding the sky observer (102). The system (100) comprises a communication network (114) communicatively connected to the sky observer (102) to collect aggregated atmospheric data comprising the sky image, aerosol properties, different components of solar irradiance and the sensor data and transmit the aggregated atmospheric data to a server (114). The server (114) receives and analyzes the aggregated atmospheric data to estimate impact of atmospheric conditions on solar energy production in a solar plant.

Description

[0001] English Description Description Technical Field [1] The field of invention generally relates to the field of meteorological monitoring and solar energy optimization. More specifically, it relates to a system and method for atmospheric observations such as cloud properties, movement and atmospheric turbidity from aerosols at sub-hourly to hourly intervals across solar power plants. Background Art [2] Solar Sky observation technology has emerged as a vital tool for monitoring atmospheric conditions, enabling advancements in meteorology and renewable energy applications. The use of ground-based sky observation systems, equipped with high-resolution imaging, has grown significantly in recent years, supporting real-time weather forecasting and solar energy optimization. These systems are deployed to analyze cloud properties, which directly influence solar radiation and weather patterns, thereby aiding in the efficient operation of solar power plants. [3] Currently, existing systems do not succeed in providing high-resolution, real-time comprehensive measurement and co-located study of cloud and aerosol properties for atmosphere observations. Traditional approaches, such as satellite-based observations and conventional ground-based methods, often fail to deliver spatial resolution and timeliness required, resulting in inaccurate predictions of solar radiation and weather patterns. [4] Other existing systems have tried to address this problem by enhancing ground- based observations. However, their scope was limited to low-frequency data collection and lacked the capability to capture dynamic changes in cloud and aerosol properties at sub-hourly intervals, leading to inefficiencies in solar plant operations and meteorological forecasting accuracy. English Description [5] Thus, in light of the above discussion, it is implied that there is need for a system and method for real-time observation and forecasting of atmosphere observations which is reliable and does not suffer from the problems discussed above. Object of Invention [6] The principal object of this invention is to provide a system and method to measure cloud and aerosol properties for atmospheric observations for applications requiring measurements of atmospheric conditions. [7] A further object of the invention is to provide a system and method for predicting atmospheric conditions at a desired location within a distance using a network of sky observers. This facilitates a scalable and distributed system for localized forecasting and detailed measurements. [8] Another object of the invention is to measure cloud and aerosol properties and predict cloud coverage at the desired location within a defied distance using a sky observer in the network. This benefits users by enabling precise cloud coverage forecasts for optimizing solar plant operations. [9] Another object of the invention is to capture high-resolution images of clouds to measure properties such as cloud cover, type, and movement.

[0010] Another object of the invention is to estimate cloud density, altitude, and velocity using machine learning models trained on a vast dataset of sky images from various locations and weather conditions.

[0011] Another object of the invention is to measure aerosol properties using a robotic sun photometer with more than one bandpass filter with the same or different wavelengths.

[0012] Another object of the invention is to aggregate data from all sky observers in a network for predictive modeling to forecast cloud population and solar irradiance changes at sub-hourly levels. Brief Description of Drawings

[0013] This invention is illustrated in the accompanying drawings, throughout which, like reference letters indicate corresponding parts in the various figures. English Description

[0014] The embodiments herein will be better understood from the following description with reference to the drawings, in which: Fig.1A

[0015] [Figure 1A] depicts a system for atmospheric observations, in accordance with an embodiment; Fig.1B

[0016] [Figure 1B] depicts a block diagram of a sky observers network of the system for atmospheric observations, in accordance with an embodiment; Fig.1C

[0017] [Figure 1C] depicts a block diagram of a server of the system for atmospheric observations, in accordance with an embodiment; Fig.2

[0018] [Figure 2] depicts an isometric view of a sky observer of the system for atmospheric observations, in accordance with an embodiment; Fig.3

[0019] [Figure 3] depicts a detailed view of a sky imager of the sky observer, in accordance with an embodiment; Fig.4

[0020] [Figure 4] depicts a front view of a robotic sun photometer, in accordance with an embodiment; Fig.5

[0021] [Figure 5] depicts a rear view of the robotic sun photometer mounted on a 2- DOF robotic arm with a controller, in accordance with an embodiment; Fig.6

[0022] [Figure 6] depicts a detailed view of a pyranometer with a motorized shadowband of the system for atmospheric observations, in accordance with an embodiment; English Description Fig.7

[0023] [Figure 7] depicts a detailed view of a sensor unit along with a processor for atmospheric observations, in accordance with an embodiment; Fig.8

[0024] [Figure 8] depicts a network of sky observers in an observation area, in accordance with an embodiment; Fig.9

[0025] [Figure 9] d depicts measurement of direction of cloud cover movement within the network, in accordance with an embodiment; Fig.10A

[0026] [Figure 10A] illustrates a time series plot of cloud fraction obtained at two different locations of the sky observers network, in accordance with an embodiment; Fig.10B

[0027] [Figure 10B] illustrates angle differences between cloud motion vectors found at the different locations of the sky observers network, in accordance with an embodiment; Fig.11A

[0028] [Figure 11A] illustrates a method for atmospheric observations, in accordance with an embodiment; and Fig.11B

[0029] [Figure 11B] illustrates the method for atmospheric observations, in accordance with an embodiment. Statement of Invention

[0030] The present invention discloses a system and method for atmospheric observations. The system comprises at least one sky observer comprising a sky imager to capture and communicate at least one sky image for analyzing cloud characteristics, a robotic sun photometer to track the sun and measure and English Description communicate aerosol properties, a pyranometer with a shadowband to measure and communicate solar irradiance data comprising at least one of global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI) and direct normal irradiance (DNI) and a sensor unit to capture and communicate environmental conditions and collect sensor data surrounding the sky observer. The system comprises a communication network communicatively connected to the sky observer to collect aggregated atmospheric data comprising the sky image, the aerosol properties, different components of solar irradiance data and the sensor data and transmit the aggregated atmospheric data to a server. The server to receive and analyze the aggregated atmospheric data to estimate impact of atmospheric conditions on solar energy production in a solar plant.

[0031] In an embodiment, the system comprises a network comprising a plurality of sky observers deployed in an observation area. Each sky observer comprising a sky imager to capture and communicate at least sky image for analyzing cloud characteristics, a robotic sun photometer to track the sun and measure and communicate aerosol properties and a pyranometer with a motorized shadowband to measure and communicate irradiance data comprising at least one of global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI) and direct normal irradiance (DNI), and a sensor unit to capture and communicate environmental conditions surrounding corresponding sky observer and collect sensor data. The system comprises a communication module to transmit aggregated atmospheric data comprising the sky image, the aerosol measurements, irradiance data and the sensor data to a processor. The processor pre-processes the aggregated atmospheric data and transmits the pre- processed aggregated data to a server.

[0032] The method for atmospheric monitoring comprises, capturing and communicating, by a sky imager in a sky observer, at least one sky image to analyze cloud characteristics, tracking the sun for measuring and communicating aerosol properties by a robotic sun photometer in the sky observer, capturing environmental conditions surrounding the sky observer and collecting and communicating sensor data by a sensor unit in the sky observer, capturing and communicating different components of solar irradiance by a pyranometer with a motorized shadowband, capturing environmental conditions surrounding the sky English Description observer and collecting and communicating sensor data by a sensor data in the sky observer, collecting and communicating aggregated atmospheric data comprising the sky image, the aerosol properties, different components of solar irradiance data and the sensor data through a communication network communicatively connected to the sky observer, transmitting the aggregated atmospheric data to a server through the communication network and analyzing, by the server, the aggregated atmospheric data to estimate the impact of atmospheric conditions on solar energy production in a solar plant. Detailed Description

[0033] The The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and / or detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.

[0034] The present invention discloses a system and method for observing atmospheric conditions through a network of sky observers, each equipped with a sky imager, a robotic sun photometer, a pyranometer with a motorized shadowband and a sensor unit to measure cloud and aerosol and irradiance characteristics. The system captures high-resolution sky images, aerosol data, irradiance data and environmental data and aggregates this information across the network and utilizes advanced predictive models to forecast cloud behavior and solar irradiance changes at frequent intervals. This approach supports applications in weather monitoring and solar energy optimization by delivering accurate, localized atmospheric predictions.

[0035] Figure 1A depicts a system 100 for atmospheric observations, in accordance with an embodiment. The system 100 comprises a sky observer 102, a communication network 114 and a server 116. The sky observer 102 further English Description comprises a sky imager 104, a robotic sun photometer 106 and a pyranometer 108 with a motorized shadow band 110.

[0036] In an embodiment, the sky observer 102 is a key component of the system for atmospheric observations. The sky observer 102 is deployed around a solar plant to collect high-resolution data on cloud dynamics, aerosols, and solar irradiance, supporting real-time forecasting and optimization for solar energy production.

[0037] In an embodiment, the sky imager 104 enables capturing high dynamic images (HDR) of the sky at 1-minute intervals for analyzing cloud characteristics such as cloud cover, cloud type, and cloud movement.

[0038] In an embodiment, the robotic sun photometer 106 is positioned adjacent to the sky imager 104 which enables sun tracking to measure aerosol optical depth with high resolution at five wavelengths (any wavelength between 345 - 1100 nm), each with a channel field of view (FOV) of less than 2°. The robotic sun photometer 106 operates with a temporal resolution of 1-minute intervals for solar disk tracking to measure direct solar irradiance at these five wavelengths, and 10-minute intervals for almucantar scans to measure diffuse solar irradiance. The robotic sun photometer 106 generates atmospheric data comprising aerosol optical depth (AOD) every minute (if the sun is unobstructed after confirmation by the sky imager 104), and every 10th minute, it produces single-scattering albedo (SSA), complex refractive index, and volume size distribution function.

[0039] In an embodiment, the pyranometer 108 is an EKO ML-01 sensor (Class C pyranometer) which is a compact, industrial-grade silicon pyranometer for high- quality solar irradiance measurements, specifically within the wavelength range of 400 to 1100 nanometers, with a sensitivity of 50 µV / W·m² (nominal), a response time of less than 1 ms (95% response), and an output range of 0-100 mV, operating reliably within a temperature range of -30°C to +70°C. This range of 400 to 1100 nanometers covers visible and near-infrared spectrum, making it suitable for capturing total solar radiation intensity relevant to applications like solar energy monitoring, meteorological studies, agricultural research, and environmental studies.

[0040] In an embodiment, the pyranometer 108 comprises the motorized shadow band 110 made of aluminum, with dimensions of 80 mm arc radius, 30 mm arc English Description width, and 3 mm arc thickness to measure global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI), and direct normal irradiance (DNI). The pyranometer 108 along with the motorized shadowband 110 and the sky imager 104 generates atmospheric data comprising direct-to-diffuse irradiance ratio (DDIR), cloud optical depth, and cloud albedo, with the variation of DDIR analyzed with respect to cloud type and cloud cover for detailed atmospheric studies.

[0041] In an embodiment, the sky observer 102 comprising the sky imager 104, the robotic sun photometer 106 and the pyranometer 108 comprising the motorized shadow band 110 is housed inside a weatherproof box 112 to ensure reliable operation in various environmental conditions.

[0042] In an embodiment, the communication network 114 facilitates transmission of collected atmospheric data such as high-resolution sky images from the sky imager 104, aerosol measurements from the robotic sun photometer 106, irradiance measurements from the pyranometer 108 along with the motorized shadowband 110 and sensor data from a sensor unit 120 (shown in Figure 1B) to the server 116 for further processing.

[0043] In an embodiment, the server 116 is communicatively attached to the sky observer 102 through the communication network 114 to receive the atmospheric data comprising the high-resolution sky images, the aerosol measurements and the irradiance data and sensor data for centralized analysis using forecasting models. The server 116 ensures reliable and efficient processing, enabling scalable and centralized forecasting of cloud population and solar irradiance changes.

[0044] Figure 1B depicts a block diagram of the sky observer 102 communicatively connected to the server 116 for aggregated data and forecasting, in accordance with an embodiment.

[0045] In an embodiment, the sky observer 102 comprises the sky imager 104, the robotic sun photometer 106, the pyranometer 108 with corresponding motorized shadow band 110, a sensor unit 120 and a communication module 122, all integrated to collect and transmit atmospheric data for real-time analysis. English Description

[0046] In an embodiment, the sensor unit 120 comprises one or more sensors to measure atmospheric and positional data surrounding the sky observer 102. Referring to Figure 2 illustrating an isometric view of the sky observer 102, the one or more sensors of the sensor unit 120 comprises an inertial measurement unit (IMU) sensor 220 to measure orientation of the sky observer 102, a geographical positioning system (GPS) sensor 226 to measure geographical location data of the sky observer 102 and a sensing module 222 to measure environmental conditions surrounding the sky observer 102. The sensing module 222 comprises a temperature sensor to measure temperature in atmosphere surrounding the sky observer 102, a pressure sensor to measure pressure in the atmosphere surrounding the sky observer 102 and a humidity sensor to measure humidity in the atmosphere surrounding the sky observer 102.

[0047] In an embodiment, the communication module 122 is an integrated hardware component within the sky observer 102 that enables wireless transmission of the atmospheric data to the server 116 from the communication module 122 of the sky observer 102. The communication module 122 ensures reliable, real-time transmission of the HDR images of the sky, the aerosol measurements, the irradiance data, GPS coordinates to the server 116, critical for accurate cloud motion forecasting. The communication module 122 can support high-speed, low latency data transfer and utilizes 4G long term evolution (LTE) technology for data transmission. The communication module 122 may also utilize other communication technologies such as 5G, and LoRA, among others. Referring to Figure 2, the sky observer 102 comprises a processor 218, and the communication module 122 is integrated into the processor 218 (shown in Figure 7).

[0048] In an embodiment, the sky observer 102 is enabled to preprocess the aggregated atmospheric data locally through the processor 218 (shown in Figure 2 and 7). The preprocessing comprises at least one of noise reduction, cropping, resizing, removal of lens glare, and clearing occlusions for HDR sky images, as well as calibration for AOD and irradiance measurements, before transmitting the aggregated data to the server 116 via the communication module 122, optimizing data transmission efficiency. English Description

[0049] Figure 1C depicts a block diagram of the server 116, in accordance with an embodiment. The server 116 aggregates atmospheric data from the sky observer 102, performs real-time optimization and runs forecasting models for energy forecasting to support solar energy applications.

[0050] In an embodiment, the server 116 comprises a cloud forecast model 130, an irradiance forecast model 132 and an energy forecast model 140. The irradiance forecast model 132 further comprises a convolutional neural network (CNN) model 134 and a time series network model (TSNM) 136 comprising a long short term memory (LSTM) model 138. The energy forecast model 140 comprises a machine learning (ML) model 142.

[0051] In an embodiment, the server 116 feeds the atmospheric data comprising the HDR sky images, the aerosol measurements, irradiance measurements and sensor data to the cloud forecast model 130. The cloud forecast model 130 predicts future cloud population characteristics comprising cloud cover, cloud type, cloud density, cloud altitude and cloud velocity using spatiotemporal cloud population transfer techniques. The cloud forecast model 130 calculates motion vectors from consecutive HDR sky images captured at 1-minute time intervals, assigns weights to the motion vectors for reliability and updates predictions in real-time. The cloud forecast model 130 outputs predicted cloud population images that map cloud density across a grid for future time steps (e.g., a dense cloud arriving within 5 to 15 minutes or longer intervals).

[0052] In an embodiment, the predicted cloud population images from the cloud forecast model 130 are fed to the irradiance forecast model 132 which comprises the CNN model 134. The irradiance forecast model 132 utilizes the CNN model 134 to extract spatial features such as cloud density and edges from the predicted cloud images and then flatten the cloud images into a vector.

[0053] In an embodiment, the vector is fed to the TSNM 136 comprising the LSTM model 138. The TSNM 136 utilizes the vector along with additional data comprising the aerosol measurements and irradiance data comprising past irradiance values to predict irradiance by capturing temporal patterns such as cloud movement over time, thereby generating initial irradiance forecasts. English Description

[0054] In an embodiment, these initial irradiance forecasts are refined by comparing them with ground-truth data from a ground-truth sky observer (not shown) which provides actual irradiance measurements for validation thereby minimizing errors to output finalized irradiance forecasts.

[0055] In an embodiment, these finalized irradiance forecasts are fed into the energy forecast model 140 comprising the ML model 142. The energy forecast model 140 utilizes these irradiance forecasts to estimate production of solar energy in solar plants. The energy forecast model 140 outputs a final energy forecast (e.g., in kWh over the next 24 hours) for solar energy applications.

[0056] Figure 2 depicts an isometric view of the sky observer 102, in accordance with an embodiment. The sky observer 102 comprises the sky imager 104, collimating tubes 210, bandpass filters 212, motors 214, a motor controller 216, a processor 218, the pyranometer 108 comprising a motor with shadow band 208, a fan 206, and a voltage regulator 224, all housed within the weatherproof box 112 to ensure reliable operation in various environmental conditions.

[0057] In an embodiment, the sky imager 104 comprises at least one camera 202 which is a high-resolution camera equipped with a 160° field-of-view fisheye lens. The camera 202 is a 12.3 MP Complementary Metal-Oxide-Semiconductor (CMOS) camera with a resolution of 4056 x 3040 pixels and featuring a 1 / 2.3- inch CMOS sensor optimized for low-light performance and dynamic range, capable of capturing the high dynamic range (HDR) images of the sky at 1- minute intervals. The camera 202 comprises lens which have an F2.4 aperture with a 6.5 mm focal length and fixed focus, suitable for wide-angle sky observation. The sky imager 104 operates on 12V, consuming 20W, making it energy-efficient for continuous use, and functions reliably within a temperature range of -20°C to 50°C. The camera 202 is equipped with at least one 128GB MicroSD Card for ample local image storage.

[0058] In an embodiment, the captured HDR images by the sky imager 104 undergo preprocessing by the processor 218. The pre-processing comprises cropping, resizing, removal of lens glare, and clearing any occlusions around the sky imager 104, before being uploaded to the server 116 for further analysis by the server 116. English Description

[0059] In an embodiment, the sky imager 104 may be positioned beneath a transparent dome 204, shown in Figure 3 to protect the camera 202 from the environment while maintaining optical clarity.

[0060] In an embodiment, the collimating tubes 210 comprises five narrow optical tubes that align incoming sunlight into parallel beams, ensuring only direct sunlight (within a field of view less than 2°) reaches the robotic sun photometer 106, thereby minimizing scattered light interference for accurate aerosol optical depth (AOD) measurements across multiple wavelengths.

[0061] In an embodiment, the bandpass filters 212 isolate specific wavelengths within a range of 450-970 nm (e.g., 450 nm, 500 nm, 675 nm, 870 nm, 970 nm) with a narrow bandwidth (e.g., 10 nm). The bandpass filters 212 enables spectral analysis of sunlight intensity to calculate aerosol optical depth (AOD) using the Beer-Lambert Law, providing data on atmospheric clarity and turbidity.

[0062] In an embodiment, the motors 214 comprises servo motors (shown in Figure 5), each equipped with 0.09° angular resolution, enabling two-axis tracking of the sun (azimuth and elevation) with high precision (e.g., within 0.1 degrees). The motors 214 adjusts orientation of the collimating tubes 210 every few seconds to maintain alignment with the sun, ensuring reliable AOD measurements.

[0063] In an embodiment, the motor controller 216 (shown in Figure 5) manages the motors 214 by receiving data corresponding to position of the sun from the processor 218to calculate the sun’s azimuth and zenith angles. The motor controller 216 sends control signals to the motors 214 to control their positioning while optimizing power usage.

[0064] In an embodiment, the atmospheric data comprising aerosol optical depth (AOD), the single-scattering albedo (SSA), the complex refractive index and the volume size distribution function, generated by the robotic sun photometer 106 is transmitted to the processor 218 for pre-processing.

[0065] In an embodiment, the processor 218 collectively pre-processes atmospheric data collected from the sky imager 104, the robotic sun photometer 106, the pyranometer 108 with corresponding motorized shadowband 110 and the sensor unit 120. The pre-processing comprises performing noise reduction for HDR sky images, calibration for AOD measurements, calculating sun position for motor English Description control, and transmitting the preprocessed atmospheric data to the server 116 via the communication module 122, ensuring efficient operation within power constraints of the sky observer 102.

[0066] In an embodiment, the pyranometer 108 comprises a shadow band mechanism that enables measurement of global, diffuse and direct irradiance for accurate solar energy forecasting. The motor with shadow band 208 comprises a servo motor with 0.09° angular resolution and the shadow band 110 (30 mm wide, 80 mm arc radius, 3 mm thickness) made of aluminum to selectively block direct sunlight while allowing diffuse skylight to reach the EKO ML-01 sensor of the pyranometer 108. The motor with shadow band 208 is controlled by the motor controller 216 and the processor 218, wherein the processor 218 signals the motor controller 216 to rotate the motor with shadow band 208 to align with the sun’s position.

[0067] In an embodiment, using the GPS sensor 226 and the motor with the shadow band 208, the EKO ML-01 sensor of the pyranometer 108 measures the diffuse horizontal irradiance (DHI) every few minutes (e.g., every 5 minutes). In case the shadow band is moved aside, the EKO ML-01 sensor measures the global horizontal irradiance (GHI) and then a direct normal irradiance (DNI) is calculated using the GHI and DHI. The motor with the shadow band 208 ensures precise alignment (e.g., within 0.1 degrees) with the sun’s trajectory, adjusting the position of the motor with shadow band 208 to provide continuous, real-time irradiance data. The real-time irradiance data is preprocessed by the processor 218 and transmitted to the server 116 for use in the irradiance forecast model 132 and validation by the ground-truth sky observer (not shown).

[0068] In an embodiment, the fan 206 is a lower power cooling fan, mounted within the weatherproof box 112 to maintain optimal operating temperatures in the sky observer 102 and its components by facilitating air flow and dissipating heat generated during continuous operation, especially in high-temperature environments. The fan 206 operates intermittently based on readings from the temperature sensor to prevent overheating.

[0069] In an embodiment, the voltage regulator 224 is a critical component to maintain a stable voltage supply to the components of the sky observer 102, English Description ensuring consistent operation of the sky imager 104, the robotic sun photometer 106, the pyranometer 108 with corresponding motorized shadowband 110 and the sensor unit 120. The voltage regulator 224 prevents fluctuations that could affect the performance of sky observer 102.

[0070] Figure 4 depicts a front view of the robotic sun photometer 106, in accordance with an embodiment. The robotic sun photometer 106 further comprises detectors 402 and a high resolution analog to digital converter (ADC) 404.

[0071] In an embodiment, the detectors 402 comprising silicon photodiodes are positioned at ends of each collimating tube 210 to measure intensity of filtered sunlight at specific wavelengths such as 450 nm, 500 nm, 675 nm, 870 nm, 970 nm. Each detector 402 generates a photocurrent proportional to intensity of sunlight which is converted into a voltage signal using a transimpedance amplifier to provide raw data required to calculate the AOD with high sensitivity to low light levels.

[0072] In an embodiment, the high resolution ADC 404 is a 16 bit analog to digital converter which digitizes the voltage signal obtained from the detectors 402 and samples at a high frequency to capture subtle intensity variations critical for precise AOD measurements. This data is transmitted to the processor 218 for pre-processing and subsequent transmission to the server 116.

[0073] Figure 6 depicts a front view of the pyranometer 108 with corresponding motorized shadowband 110, in accordance with an embodiment. The sky imager 104 operatively communicates with the motorized shadowband 110. The sky imager 104 provides precise location of the sun, assisting the motorized shadowband 110 to cover the pyranometer 108 from direct irradiance of the sun with better accuracy. In such a condition the pyranometer 108 measures the DHI, thus calculating the DNI value from the measured GHI, DHI and sun’s position (azimuth ϕ and elevation θ).

[0074] In an embodiment, the sky imager 104 operatively communicates with the robotic sun photometer 106 such that the sky imager 104 provides precise location of the sun, assisting the robotic sun photometer 106 to point to sun with better accuracy. The sky imager 104 also provides an input to the robotic sun photometer 106 to filter any measurement in which the clouds are obstructing the English Description sun. The advantage of this configuration of the current invention is that the sky imager 104 and the robotic sun photometer 106 work together to improve the AOD measurements and provide cleaner measurements.

[0075] The sky imager 104 operatively communicates with the motorized shadowband 110. The sky imager 104 provides precise location of the sun, assisting the motorized shadowband to cover the pyranometer 108 from direct irradiance of the sun with better accuracy. In such a condition the pyranometer measures the DHI, thus calculating the DNI value from the measured GHI, DHI and sun’s position (azimuth ϕ and elevation θ) to align the robotic sun photometer 106 and the pyranometer 108.

[0076] In an embodiment, the sky imager 104 identifies the sun’s position (x,y) as a centroid of brightest spot in the HDR sky image, with image center at (cx,cy) and radius r. The fisheye lens maps sky hemisphere using equi-solid projection. The radial distance is calculated as:

[0078] The elevation and azimuth angles are computed as:

[0079] Elevation angle θ=2∙sin^(-1)^(r / (2∙f))∙180 / π where f is the focal length of the fisheye lens

[0080] Azimuth angle ^^ = arctan 2(^^ − ^^^^, ^^ − ^^^^)

[0081] These angles are converted to Cartesian coordinates in a Sky Imager Coordinate System (SICS) represented as:

[0083] In an embodiment, to control the robotic arm of both the shadow band 110 and the robotic sun photometer 106, computed azimuth angle and elevation angle are required in the robotic arm coordinate system (RACS). A transformation matrix is applied that rotates and shifts the sun’s position from the sky imager coordinate system to robotic arm coordinate system which the motors 214 for robotic sun photometer 106 uses to point at the sun for direct irradiance measurement by defining a transformation between the SICS and the RACS. The English Description relationship between the SICS and the RACS involves both rotation and translation. If R is a rotation matrix and T is a translation vector from the SICS and the RACS then R and the T are represented as: ^^11^^12^^13Rotation matrix: ℛ = [^^21^^22^^23] ^^31^^32^^33^^^^

[0085] Translation matrix: ^^ = [^^^^]

[0086] Transform coordinate from the SICS to RACS are represented as:

[0088] Further, the cartesian coordinated in the RACS is converted to spherical coordinates as:

[0089] Azimuth angle: ^^ = ^^^^^^^^^^^^2(^^^^^^^^^^, ^^^^^^^^^^)

[0090] Elevation angle: ^^ = ^^^^^^^^^^^^2(^^ 2^^^^^^^^, √^^^^^^^^^^ + ^^^2^^^^^^^

[0091] From the above computations, it is understood that the sky imager 104 assists the robotic sun photometer 106 to point it towards the sun eliminating the need for a sun sensor. The detection of both sun and sun and cloud position by the sky imager 104 helps the pyranometer 108 with the motorized shadowband 110 to accurately measure cloud optical depth (COD), which is an important feature that affects irradiance prediction by the irradiance forecast model 134. Further the pyranometer 108 with the motorized shadowband 110 provides information about illumination change that affects the sky imager 104 and the robotic sun photometer 106 radiation transfer feature helps them adjust with changing illumination and capture / measure accurately.

[0092] The In an embodiment, a plurality of sky observers 102 / 1- 102 / 8 is strategically positioned in the form of a network 800 and deployed in an observation area 802 around a solar plant as illustrated in Figure 8, in accordance with an embodiment. The network 800 of sky observers 102 / 1- 102 / 8 is positioned such that adjacent sky observers 102 / 1 – 102 / 8 comprise overlapping fields of view (FOV). The overlapping FOV enables correlation of cloud features between at least two points A and B corresponding to a pair of English Description adjacent sky observers 102 / 1, 102 / 2. The sensor unit 120 comprising the GPS sensor 226 provides location data to map spatial relationships between the sky observers 102 / 1-102 / 8 for tracking cloud movement across the observation area 802.

[0093] In an embodiment, the observation area 802 may be a 10 km x 10 km region surrounding the solar plant comprising eight sky observers 102 / 1 - 102 / 8 (as an example) positioned at key points (e.g., corners and midpoints of perimeter of the observation area 802) to maximize spatial data collection. Each sky observer 102 / 1-102 / 8 is equipped with corresponding sky imager 104, corresponding robotic sun photometer 106, corresponding pyranometer 108 with motorized shadowband 110, corresponding GPS sensor 226, and corresponding communication module 122, enabling the collection of HDR sky images, aerosol optical depth (AOD), irradiance measurements, and positional data, which are transmitted to the server 116 for aggregated processing.

[0094] In an embodiment, the observation area 802 operates cohesively with each sky observer 102 communicating via the communication module 122 to share real-time data that supports the cloud forecast model 130 by tracking cloud movement. The cloud forecast model 130 utilizes spatiotemporal cloud population transfer and calculates motion vectors 804 / 1- 804 / 8 from analyzing consecutive HDR sky images (captured at 1-minute intervals) to predict cloud movement. For instance, if a cloud edge moves 10 pixels eastward between two images from sky observer 102 / 1, the cloud forecast model 130 calculates its velocity using the 160° field of view of the sky imager 104 and GPS-derived distances between the sky observers 102 / 1 – 102 / 8. The motion vectors 804 / 1 – 804 / 8 represents direction and speed of cloud movement between adjacent sky observers 102 / 1 – 102 / 8. The cloud forecast model 130 assigns corresponding weights w1- w8 to corresponding motion vectors 804 / 1 - 804 / 8 based on their consistency. For instance, the motion vectors 804 / 1, 804 / 2, and 804 / 3 showing stable patterns across corresponding sky observers 102 / 1, 102 / 2, and 102 / 3 may be given higher weights as compared to inconsistent or noisy motion vectors 804 / 4 of the sky observer 102 / 4, and hence may receive lower weights, enhancing reliability of cloud movement predictions processed by the server 116 in real-time. Further, the GPS sensors 226 in each sky observer 104 / 1 - 104 / 8 English Description provide precise location data such as latitude, longitude, altitude with 1-meter accuracy, enabling the server 116 to map spatial relationships between the sky observers 102 / 1 - 102 / 8 which is crucial for correlating the motion vectors 804 / 1 to 804 / 8 with geographical coordinates and predicting cloud distribution across the solar plant.

[0095] Figure 9 illustrates a schematic diagram representing two sky observers 102 / 1, 102 / 2 at points A and B respectively, with black dotted lines representing a field of view (FOV) 902 of corresponding sky imagers 104 / 1, 104 / 2 within which similar cloud population may be observed moving from the point A to point B, and white arrows indicating motion vectors 804 / 1, 804 / 2 of clouds corresponding to the sky observers 102 / 1, 102 / 2. The server 116 utilizes directional filtering with the cloud forecast model 130 within the FOV 902 to predict cloud movement from the point A to the point B, estimating time of arrival (ToA) and enabling cloud population forecasting for solar energy applications. θA and θB represent angular positions of a cloud relative to zenith at the points A and B respectively. An angular difference between the clouds within the FOV 902 is computed by the equation presented below: θ_i = θ_B - θ_A

[0096] In an embodiment, the weights w1 - w8 represented as wi are assigned to corresponding cloud fraction (CF) when θi ≤ θth, wherein total estimated cloud population at any point k after time τ is expressed as a weighted sum of cloud contributions from all points i satisfying the FOV threshold: where the directional alignment based weight w_i is defined as: wherein, θ_i∈[0,θ_th] represents an angular deviation from a direction of advection, and cos^(∙)̇ ensures that higher weights are assigned to directions more closely aligned with a path of cloud transfer toward the point k. The lag time English Description τ is estimated based on wind speed, spatial separation, and optical flow magnitude.

[0097] The Figure 10A illustrates a time series plot 1000 of cloud fraction on March 01, 2025, as historical data for validation, in accordance with an embodiment. The time series plot 1000 compares an actual cloud fraction 1004 predicted from the point B against a lag-adjusted cloud fraction 1006 predicted from the point A, demonstrating accuracy of the cloud forecast model 130 in predicting cloud movement across the observation area 802.

[0098] Figure 10B illustrates a graph 1002 representing an angular difference 1008 between the motion vectors 804 / 1, 804 / 2 at the points A and B respectively, in accordance with an embodiment. This graph 1002 validates the cloud forecast model 130 by demonstrating that smaller angular differences 1008 correlate with better cloud fraction predictions, supporting the directional filtering method utilized by the cloud forecast model 130 in the server 116.

[0099] Figure 11A illustrates a method 1100 for atmospheric observations, in accordance with an embodiment. The method 700 begins with deploying a network 800 comprising a plurality of sky observers 102 / 1 – 102 / 8, each sky observer 102 / 1 – 102 / 8 comprising the sky imager 104, the robotic sun photometer 106, the pyranometer 108 with the motorized shadow band 110 and the sensor unit 120 as depicted at step 1102. Subsequently, the method 1100 discloses capturing high dynamic range (HDR) images by the sky imager 104 at 1-minute intervals by a camera 202 with a 160° field-of-view fisheye lens, as depicted at step 1104. Thereafter, the method 1100 discloses measuring aerosol properties such as aerosol optical depth (AOD) at five wavelengths by the robotic sun photometer 106 at 1-minute time intervals, as depicted at step 1106. Subsequently, the method 700 discloses measuring irradiance data, comprising global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI), and direct normal irradiance (DNI) by the pyranometer 108 with the motorized shadow band 110, at intervals of a few minutes, as depicted at step 1108. Thereafter, the method 1100 discloses collecting sensor data comprising orientation from an inertial measurement unit (IMU) sensor 220, geographical location (latitude, longitude, altitude with 1-meter accuracy) from a GPS sensor 226 and environmental conditions (temperature, pressure, humidity) from a sensing English Description module 222, as depicted at step 1110. Subsequently, the method 1100 discloses preprocessing the collected atmospheric data locally in the sky observer 102 by a processor 218 by performing tasks such as noise reduction and image cropping for the HDR sky images, calibration for AOD and irradiance measurements, as depicted at step 1112.

[0100] Figure 11B illustrates the method 1100 for atmospheric observations, in accordance with an embodiment. The method 1100 discloses transmitting the preprocessed HDR sky images, aerosol measurements, irradiance data, and sensor data to a server 116 via a communication module 122 in the sky observer 102and determining spatial relationships between the sky observers 102 / 1 -102 / 8 using GPS-derived distances, calculating corresponding motion vectors 804 / 1 – 804 / 8 of clouds (direction and speed) from consecutive HDR sky images for each sky observer 102 / 1 – 102 / 8, assigning weights w1- w8 to the motion vectors 804 / 1 – 804 / 8 based on consistency, as depicted at step 1114. Thereafter, the method 1100 discloses analyzing the HDR sky images, the aerosol measurements, the irradiance data, and the sensor data by a cloud forecast model 130 to predict cloud properties, including cloud cover, cloud type, cloud density, cloud altitude, and cloud velocity, using spatiotemporal cloud population transfer techniques and directional filtering, as depicted at step 1116. Subsequently, the method 1100 discloses processing the predicted cloud population images by an irradiance forecast model 132, where a convolutional neural network (CNN) model 134 extracts spatial features (e.g., cloud density, edges) and a time series network model (TSNM) 136 comprising a long short- term memory (LSTM) model 138 captures temporal patterns (e.g., cloud movement over time) to produce initial irradiance forecasts and refining the initial irradiance forecasts by comparing them with ground-truth data to generate finalized irradiance data, as depicted at step 1118. Subsequently, the method 1100 discloses feeding the finalized irradiance forecasts into an energy forecast model 140 which uses a machine learning (ML) model 142 to estimate solar energy production and generate an energy forecast , as depicted at step 1122. Finally, the method 1100 concludes by determining and communicating the final energy forecasts to end-users for solar energy optimization and applications, as depicted at step 1124. English Description

[0101] The advantages of the current invention include the ability to provide high- resolution, synergistic real-time operations, enabling more accurate cloud and aerosol products compared to conventional use of separate instruments. The system of the current invention achieves this through real-time data collection at 1-minute intervals, improving the timeliness and precision of atmospheric monitoring for both meteorological and solar energy purposes.

[0102] The sky imager of the current invention aids in improving accuracy of AOD measurement of the robotic sun photometer by automatically detecting and removing cloud contaminated measurements. The sky imager also helps the pyranometer with the shadowband to measure accurate cloud optical depth. Likewise, AOD measurements aid in understanding irradiance change measurements by the pyranometer 108 with motorized shadowband 110. Further, the pyranometer provides information of illumination change that affects the sky imager 104 and the robotic sun photometer 106. This integrated configuration of the current invention is an advantage of the system and the components working in synergy is a critical feature which is not addressed by any of the conventional systems.

[0103] The current invention also offers the advantage of enhanced spatial resolution through the strategic placement of sky observers in a network. By positioning sky observers at key points and using GPS sensors for precise location mapping, the system captures fine-scale cloud movement and aerosol variations across the region. This high spatial resolution, combined with the directional filtering of the cloud forecast model and weighted motion vectors, improves the accuracy of cloud population transfer predictions, particularly for sub-hourly forecasting critical for solar plant operations.

[0104] Another advantage of the current invention is its robust and weatherproof design, which ensures reliable operation of the sky observers in various environmental conditions. Each sky observer is housed in a weatherproof box and equipped with a fan for temperature regulation and a voltage regulator for stable power supply. This durability enhances the system’s ability to provide continuous, high-quality atmospheric data, even in challenging climates, thereby supporting uninterrupted meteorological monitoring and solar energy forecasting. English Description

[0105] A further advantage is the system’s ability to perform local preprocessing of atmospheric data, reducing the computational load on the server and minimizing data transmission requirements. The single board computer in each sky observer preprocesses data by performing tasks such as noise reduction, image cropping for HDR sky images, and calibration for aerosol optical depth (AOD) and irradiance measurements. This local processing, combined with high module transmission via the communication module, ensures efficient and scalable data handling, making the system suitable for large-scale networks of sky observers.

[0106] An additional advantage is that the system of the current invention leverages advanced machine learning models to estimate the impact of cloud dynamics on solar energy generation, allowing for better planning and resource allocation in solar plant operations.

[0107] The system of the current invention provides the advantage of adaptability and scalability for diverse applications beyond solar energy and meteorology, such as environmental monitoring and climate research. The network of sky observers with their ability to measure aerosol optical depth, irradiance, and environmental conditions, can contribute to studies on air quality, atmospheric composition, and climate change impacts. The modular design of the sky observers and the use of machine learning models by the server allow the system to be adapted for additional forecasting tasks or expanded to larger observation networks, enhancing its utility across multiple scientific and industrial domains.

[0108] Applications of the current invention include meteorological monitoring and forecasting, where the system of the current invention provides detailed cloud and aerosol data to enhance weather predictions, and solar energy management, where the energy forecasts support optimization of solar plant operations by predicting sunlight availability based on cloud cover and density.

[0109] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and / or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended English Description within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the scope of the embodiments as described here.

Claims

English Claims Claims

1. A system (100) for atmospheric monitoring, comprising: at least one sky observer (102) comprising: a sky imager (104) to capture and communicate at least one sky image for analyzing cloud characteristics; a robotic sun photometer (106) to track the sun to measure and communicate aerosol properties; and a sensor unit (120) to capture and communicate environmental conditions and sensor data surrounding the sky observer (102); a communication network (114) communicatively connected to the sky observer (102) to collect aggregated atmospheric data comprising the sky image, the aerosol properties and the sensor data and transmit the aggregated atmospheric data to a server (114); and the server (114) to receive and analyze the aggregated atmospheric data to estimate impact of atmospheric conditions on solar energy production in a solar plant.

2. The system (100) as claimed in claim 1, wherein the sky imager (104) comprises: a transparent dome (204); a camera (202) positioned beneath the transparent dome (204) to capture high dynamic range (HDR) sky images at 1-minute intervals, wherein the camera (202) comprises a Complementary Metal-Oxide-Semiconductor (CMOS) camera for low-light performance and dynamic range; a fisheye lens comprising a 160° field-of-view (FOV) to enable wide-angle capture of the HDR sky images, wherein the fisheye lens comprises an F2.4 aperture, a 6.5 mm focal length and fixed focus; and a storage unit comprising at least one MicroSD card to store the captured HDR sky images.English Claims

3. The system (100) as claimed in claim 1, wherein the robotic sun photometer (106) generates the aerosol measurements comprising at least one of single-scattering albedo (SSA), complex refractive index and volume size distribution function with a temporal resolution of 1-minute intervals for tracking the sun to measure direct solar irradiance and diffuse solar irradiance, wherein the robotic sun photometer (106) comprises: motors (214) comprising 0.09° angular resolution to enable two-axis tracking of the azimuth and the elevation of the sun with a precision of within 0.1 degrees and adjust orientation of the robotic sun photometer (106) to maintain alignment with the sun; collimating tubes (210) to align incoming sunlight into parallel beams within a field of view (FOV) within a range of 1°- 2°; bandpass filters (212) to isolate specific wavelengths within a range of 450 nm - 970 nm with a narrow bandwidth of 10 nm to generate filtered sunlight; detectors (402) comprising silicon photodiodes, positioned at ends of the collimating tubes (210) to measure intensity of the filtered sunlight, each detector (4020 generating a photocurrent proportional to the intensity of the filtered sunlight, wherein the photocurrent is converted into a voltage signal; and a high resolution analog-to-digital converter (ADC) (404) to digitize the voltage signal.

4. The system (100) as claimed in claim 1, wherein the sky observer (102) comprises: a pyranometer (108) comprising a motorized shadowband (110), wherein the pyranometer (108) comprising the motorized shadowband (110) measures at least one of global horizontal irradiance (GHI), diffuse horizontal irradiance (DNI) and direct normal irradiance (DNI); an EKO ML-01 sensor disposed in the pyranometer (108) to measure solar irradiance within a wavelength range of 400 to 1100 nm, with a sensitivity of 50 µV / W·m², a response time of less than 1 ms, an output range of 0–100 mV; andEnglish Claims a motor with shadow band (208) to drive the motorized shadow band (110) with approximately 0.09° angular resolution to selectively block direct sunlight, wherein the motor with shadow band (208) aligns the shadow band (110) with trajectory of the sun within 0.1 degrees of precision.

5. The system (100) as claimed in claim 1, wherein the sensor unit (120) comprises: at least one inertial measurement unit (IMU) sensor (220) to measure orientation of the sky observer (102); at least one geographical positioning system (GPS) sensor (226) to measure geographical location data of the sky observer (102); and at least one sensing module (222) comprising: at least one temperature sensor to measure temperature of atmosphere surrounding the sky observer (102); at least one pressure sensor to measure pressure of the atmosphere; and at least one humidity sensor to measure humidity of the atmosphere.

6. The system (100) as claimed in claim 1, wherein the server (116) comprises: a cloud forecast model (130) to predict future cloud characteristics by analyzing the sky image and output predicted cloud population images using a spatiotemporal cloud population transfer technique; an irradiance forecast model (132) to generate a finalized irradiance forecast, the irradiance forecast model (132) comprising: a convolutional neural network (CNN) model (134) to extract spatial features from the predicted cloud population images and flatten the predicted cloud population images into a vector; and a time series network model (TSNM) (136) comprising a long short term memory (LSTM) model (138) to predict an initial irradiance forecast by capturing temporal patterns of cloud movement over time by utilizing the vector from the CNN model (134), the aerosol properties from the robotic sunEnglish Claims photometer (106) and irradiance data from a pyranometer (108) comprising a motorized shadowband (110) and output a finalized irradiance forecast; and an energy forecast model (140) to estimate the solar energy production in the solar plant, the energy forecast model (140) comprising a machine learning (ML) model (142) to process the finalized irradiance forecast.

7. The system (100) as claimed in claim 1, wherein the sky imager (104) determines an azimuth angle and an elevation angle of the sun acquired by the robotic sun photometer (106) by processing the sky image to: identify a position (x,y) of the sun as a centroid of a brightest spot in the sky image captured by a fisheye lens with a focal length f, the sky image having a center at (cx,cy) and a radial distance r; calculate the radial distance r from the center of the sky imager to the pointthe focal length of the fisheye lens and the azimuth angle (^^) by ^^ =arctan 2(^^ − ^^^^, ^^ − ^^^^);convert spherical coordinates to cartesian coordinates in a sky imager coordinate system (SICS) as:determine a transformation from the SICS to a robotic arm coordinate system (RACS) by computing a rotation matrix R as: R= ^^11^^12^^13= [^^21^^22^^23 ] ;^^31^^32^^33compute a translation matrix ^^ as: ^^ =;transform coordinates from the SICS to the RACS by: ℘^^^^^^^^ =∙ ℘^^^^^^^^ + ^^; andEnglish Claims compute the azimuth angle (^^) and the elevation angle (^^) in the RACS as: ^^ =^^^^^^^^^^^^2(^^^^^^^^^^, ^^^^^^^^^^) and ^^ = ^^^^^^^^^^^^2(^^ 2^^^^^^^^, √^^^^^^^^^^ + ^^^2^^^^^^^, wherein the computed azimuth angle and the elevation angle are communicated to the robotic sun photometer (106) and the pyranometer (108) comprising a motorized shadowband (110) for accurate positioning of sun photometer (106) and the pyranometer (108) with the motorized shadowband (110).

8. A system (100) for atmospheric observations comprising: a network (800) of a plurality of sky observers (102 / 1 – 102 / 8) to monitor an observation area (802), each sky observer (102 / 1 – 102 / 8) comprising: a sky imager (104) to capture and communicate at least one sky image for analyzing cloud characteristics; a robotic sun photometer (106) to track the sun to measure and communicate aerosol properties; and a pyranometer (108) comprising a motorized shadowband (110) to measure and communicate irradiance data comprising at least one of global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI) and direct normal irradiance (DNI); a sensor unit (120) to measure environmental conditions surrounding a corresponding sky observer (102) to collect and communicate sensor data; a communication module (122) to transmit aggregated atmospheric data comprising the sky image, the aerosol measurements, the irradiance data and the sensor data to a processor (218); and the processor (218) to pre-process the aggregated atmospheric data and transmit the pre-processed aggregated data to a server (116).

9. The system as claimed in claim 8, wherein the network (800) of sky observers (102 / 1- 102 / 8) is strategically positioned such that adjacent sky observers (102 / 1 – 102 / 8) comprise overlapping fields of view (FOV), wherein the overlapping FOV enables correlation of cloud characteristics between at least two points (A, B) corresponding to a pair of adjacent sky observers (102 / 1,English Claims 102 / 2), and wherein the sensor unit (120) comprises a GPS sensor (226) provides location data to map spatial relationships between the sky observers (102 / 1-102 / 8) for tracking cloud movement across the observation area (802).

10. The system as claimed in claim 8, wherein the sky imager (104) comprises at least one camera (202) to capture high dynamic range (HDR) sky images at 1-minute intervals, wherein the processor (218) in each sky observer (102 / 1-102 / 8) in the network (800) calculates corresponding motion vectors (804 / 1 – 804 / 8) by performing cross-correlation of the cloud characteristics between consecutive HDR sky images captured at 1-minute intervals, and assign weights (w1 – w8) to the motion vectors (804 / 1 – 804 / 8) based on one or more factors comprising image quality and consistency of cloud movement, and wherein weighted motion vectors (804 / 1- 804 / 8) represent direction and speed of cloud movement within the observation area (802).

11. The system as claimed in claim 10, wherein the processor (218) computes a cloud fraction (CF) from the HDR sky images, the CF representing proportion of sky covered by clouds, and adjusts the weights (w1 – w8) assigned to the motion vectors (804 / 1 – 804 / 8) based on the CF to account for variations in cloud density, and wherein the adjusted weighted motion vectors (804 / 1 – 804 / 8) are transmitted to the server (116) for predicting future cloud characteristics.

12. The system as claimed in claim 11, wherein the server (116) comprises a cloud forecast model (130), wherein the pre-processed aggregated atmospheric data and adjusted weighted motion vectors (804 / 1 – 804 / 8), the CF and geographical location data from a GPS sensor (226) are transmitted to the server (116), and wherein the cloud forecast model (130) aggregates the adjusted weighted motion vectors (804 / 1 – 804 / 8) from the network (800) of the sky observers (102 / 1–102 / 8), correlates the cloud characteristics across overlapping fields of view (FOV) between adjacent sky observers (104 / 1 – 104 / 8), and predict the future cloud characteristics across at least one of the observation area (802), the future cloud characteristics comprising at least one of cloud cover, cloud density and cloud velocity for time steps within a range of 5 to 15 minutes.English Claims

13. A method (1100) for atmospheric monitoring, comprising: capturing and communicating, by a sky imager (104) disposed in a sky observer (102), at least one sky image to analyze cloud characteristics; tracking the sun for measuring and communicating aerosol properties, by a robotic sun photometer (106) disposed in the sky observer (102); measuring environmental conditions surrounding the sky observer (102) for collecting and communicating sensor data, by a sensor unit (120) disposed in the sky observer (102); collecting and communicating aggregated atmospheric data comprising the sky image, the aerosol properties and the sensor data through a communication network (112) communicatively connected to the sky observer (102); transmitting the aggregated atmospheric data to a server (116) through the communication network (114); and analyzing, by the server (116), the aggregated atmospheric data to estimate the impact of atmospheric conditions on solar energy production in a solar plant.

14. The method (1100) as claimed in claim 13, comprises: capturing high dynamic range (HDR) sky images at 1-minute intervals by a camera (202) of the sky imager (104), the camera (202) being a 12.3 MP Complementary Metal-Oxide-Semiconductor (CMOS) camera for low-light performance and dynamic range; enabling wide-angle capture of the HDR sky images by a fisheye lens with a 160° field-of-view (FOV), the fisheye lens comprising an F2.4 aperture, a 6.5 mm focal length, and fixed focus; and storing the captured HDR sky images locally in the sky observer (102) on a MicroSD card.

15. The method (1100) as claimed in claim 13, wherein tracking the sun and measuring the aerosol properties by the robotic sun photometer (106) comprises:English Claims generating the at least one of single-scattering albedo (SSA), complex refractive index and volume size distribution function with a temporal resolution of 1-minute intervals to measure at least one of direct solar irradiance and diffuse solar irradiance; enabling two-axis tracking of the azimuth and the elevation of the sun with a precision of within 0.1 degrees by motors (214) with 0.09° angular resolution to adjust the orientation of the robotic sun photometer (106) and maintain alignment with the sun; aligning incoming sunlight into parallel beams within a field of view (FOV) of range within 1°- 2° by collimating tubes (210); isolating specific wavelengths of 450 nm, 500 nm, 550 nm, 630 nm, and 970 nm with a narrow bandwidth of 10 nm by bandpass filters (212) to generate filtered sunlight; measuring intensity of the filtered sunlight using detectors (402) comprising silicon photodiodes positioned at the ends of the collimating tubes (210), each detector (402) generating a photocurrent proportional to the intensity of the filtered sunlight; converting the photocurrent into a voltage signal; and digitizing the voltage signal using a high-resolution analog-to-digital converter (ADC) (404).

16. The method (1100) as claimed in claim 13, comprising: measuring, by a pyranometer (108) comprising a motorized shadowband (110) in the sky observer (102), irradiance data comprising at least one of global horizontal irradiance (GHI), diffuse horizontal irradiance (DHI) and direct normal irradiance (DNI), by an EKO ML-01 sensor; selectively blocking direct sunlight using a motorized shadow band (110) driven by a motor with shadow band (208) having 0.09° angular resolution, wherein the motor with shadow band (208) aligns the shadow band (110) with the trajectory of the sun within 0.1 degrees of precision; and generating the irradiance data comprising at least one of direct-to-diffuse irradiance ratio (DDIR), cloud optical depth and cloud albedo.English Claims

17. The method (1100) as claimed in claim 13, wherein measuring environmental conditions and collecting the sensor data by the sensor unit (120) comprises: measuring orientation of the sky observer (102) by at least one inertial measurement unit (IMU) sensor (220); measuring geographical location data of the sky observer (102) by at least one geographical positioning system (GPS) sensor (226); measuring temperature, pressure, and humidity of the atmosphere surrounding the sky observer (102) by at least one sensing module (222) comprising: a temperature sensor to measure temperature surrounding the sky observer (102); a pressure sensor to measure pressure surrounding the sky observer (102); and a humidity sensor to measure humidity surrounding the sky observer (102).

18. The method (1100) as claimed in claim 13, wherein analyzing the aggregated atmospheric data by the server (116) comprises: predicting, by a cloud forecast model (130), future cloud characteristics from the sky image using a spatiotemporal cloud population transfer technique and outputting predicted cloud population images from the sky image; generating, by an irradiance forecast model (132), a finalized irradiance forecast by: extracting spatial features from the predicted cloud population images and flattening the predicted cloud population images into a vector by a convolutional neural network (CNN) model (134); and predicting an initial irradiance forecast by capturing temporal patterns of cloud movement over time by a time series network model (TSNM) (136) comprising a long short-term memory (LSTM) model (138), utilizing the vector, the aerosol properties, and irradiance data from a pyranometer (108), and output a finalized irradiance forecast; andEnglish Claims estimating, by an energy forecast model (140), the solar energy production in the solar plant by processing the finalized irradiance forecast by a machine learning (ML) model (142).

19. The system (100) as claimed in claim 13, wherein the sky imager (104) determines an azimuth angle and an elevation angle of the sun by processing the sky image by: identifying a position (x,y) of the sun as a centroid of a brightest spot in the sky image captured by a fisheye lens with a focal length f, the sky image having a center at (cx,cy) and a radial distance r; calculating the radial distance r from the center of the sky imager to the point (x,y) as r= √(^^ − ^^^^)2 + (^^ − ^^^^)2 ;the focal length of the fisheye lens and the azimuth angle (^^) by ^^ = arctan 2(^^ − ^^^^, ^^ − ^^^^);converting spherical coordinates to cartesian coordinates in a sky imager coordinate system (SICS) as:determining a transformation from the SICS to a robotic arm coordinate system ^^12^^13(RACS) by computing a rotation matrix ℛ as: ℛ =^^22^^23 ] ;^^32^^33computing a translation matrix ^^ as: ^^ =transforming coordinates from the SICS to the RACS by: ℘^^^^^^^^ =∙ ℘^^^^^^^^ + ^^; andcomputing the azimuth angle (^^) and the elevation angle (^^) in the RACS as: ^^ =^^^^^^^^^^^^2(^^^^^^^^^^, ^^^^^^^^^^) and ^^ = ^^^^^^^^^^^^2(^^^^^^^^^^, √^^ 2^^^^^^^^ + ^^^2^^^^^^^.

20. The method (1100) as claimed in claim 13, comprising: deploying a network (800) of sky observers (102 / 1–102 / 8) in an observation area (802);English Claims capturing, by the sky imager (104) in each sky observer (102 / 1–102 / 8), the sky image to analyze the cloud characteristics; tracking the sun and measuring the aerosol properties, by the robotic sun photometer (106) in each sky observer (102 / 1–102 / 8); measuring irradiance data, by a pyranometer (108) in each sky observer (102 / 1–102 / 8); measuring environmental conditions surrounding each sky observer (102 / 1– 102 / 8) and collecting the sensor data, by the sensor unit (120) in each sky observer (102 / 1–102 / 8); transmitting, by a communication module (122) in each sky observer (102 / 1– 102 / 8), the aggregated atmospheric data comprising the sky image, the aerosol properties, the irradiance data, and the sensor data to a processor (218); and preprocessing, by the processor (218), the aggregated atmospheric data and transmitting the pre-processed aggregated atmospheric data to the server (116).

21. The method (1100) as claimed in claim 20, wherein deploying the network (800) of sky observers (102 / 1–102 / 8) comprises positioning adjacent sky observers (102 / 1-102 / 8) to comprise overlapping fields of view (FOV), the overlapping FOV enabling correlation of cloud characteristics between at least two points (A, B) corresponding to a pair of adjacent sky observers (102 / 1 and 102 / 2), and wherein measuring the environmental conditions comprises using a geographical positioning system (GPS) sensor (226) in the sensor unit (120) of each sky observer (102 / 1 – 102 / 8) to provide location data for mapping spatial relationships between the sky observers (102 / 1–102 / 8), facilitating tracking of cloud movement across the observation area (802).

22. The method (1100) as claimed in claim 20, comprising, capturing high dynamic range (HDR) sky images at 1-minute intervals using a camera (202) in the sky imager (104), and wherein preprocessing by the processor (218) in each sky observer (102 / 1–102 / 8) comprises:English Claims calculating corresponding motion vectors (804 / 1–804 / 8) by performing cross- correlation of the cloud characteristics between consecutive HDR sky images captured at 1-minute intervals; and assigning weights (w1–w8) to the motion vectors (804 / 1–804 / 8) based on one or more factors comprising image quality and consistency of cloud movement, wherein the weighted motion vectors (804 / 1–804 / 8) represent direction and speed of the cloud movement within the observation area (802).

23. The method (1100) as claimed in claim 22, wherein the preprocessing by the processor (218) comprises: computing a cloud fraction (CF) from the sky image, the CF representing proportion of sky covered by clouds; adjusting the weights (w1–w8) assigned to corresponding motion vectors (804 / 1–804 / 8) based on the CF to account for variations in cloud density; and transmitting the adjusted weighted motion vectors (804 / 1–804 / 8) to the server (116) for predicting the cloud characteristics.

24. The method as claimed in claim 23, wherein analyzing the aggregated atmospheric data by the server (116) comprises: transmitting the pre-processed aggregated atmospheric data, comprising the adjusted weighted motion vectors (804 / 1–804 / 8), the cloud fraction (CF), and geographical location data from a GPS sensor (226) from the processor (218) to the server (116); aggregating, by a cloud forecast model (130) on the server (116), the adjusted weighted motion vectors (804 / 1- 804 / 8) from the network (800) of sky observers (102 / 1–102 / 8) using a spatiotemporal cloud population transfer technique; correlating the cloud characteristics across overlapping fields of view (FOV) between adjacent sky observers (102 / 1–102 / 8); and predicting the cloud population characteristics across at least one of the observation area (802), the cloud characteristics comprising at least one of cloud cover, cloud density, and cloud velocity for time steps within a range of 5 to 15 minutes or longer time intervals.

Images