Polyhedral Solar Sensor Neural Network Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing polyhedral solar sensors lack accuracy in measuring direct, diffuse, and reflected solar radiation due to reliance on analytical algorithms that do not account for various influencing parameters, and fail to estimate radiation reflected from the ground effectively.
Innovation Solution
A polyhedral solar sensor system utilizing a predictive neural network and analytical algorithms, with multiple photodetectors sensitive to different spectral bands, to accurately measure global solar radiation, including direct, diffuse, and reflected components, by self-tuning and leveraging sensor geometry and orientation data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If analytical algorithms are used to calculate direct and diffuse solar radiation components, then the device complexity is reduced, but the measurement precision deteriorates due to inability to account for various influencing parameters
Solution Approach 1:
A neural network is introduced as an intermediary between the photodetector measurements and the final radiation component calculation. The neural network processes the raw sensor data and accounts for various influencing parameters (ground reflectance, skyline shape, atmospheric conditions) that traditional analytical algorithms cannot handle, thereby improving measurement precision without significantly increasing device complexity
Solution Approach 2:
The system changes from using fixed analytical algorithms to using a trainable neural network that can adapt its parameters based on training data. This allows the system to learn and account for varying influencing parameters such as ground reflectance and skyline shape, improving measurement precision while maintaining reasonable complexity through efficient neural network architectures
2Device complexity
If traditional polyhedral solar sensors are used, then the device simplicity is maintained, but the ability to estimate ground reflected radiation is lost
Solution Approach 1:
The polyhedral solar sensor is enhanced with a neural network that enables it to perform multiple functions: measuring direct solar radiation, diffuse solar radiation, and ground-reflected radiation. This multi-functionality is achieved without fundamentally changing the sensor structure, maintaining device simplicity while significantly improving adaptability and versatility in radiation component estimation
3Measurement precision
If more photodetectors and complex processing are used to improve measurement accuracy, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The neural network is trained offline using training data that includes various environmental conditions. Once trained, the network can independently process new measurements and account for influencing parameters without requiring additional sensors or complex real-time processing hardware. This self-service capability improves measurement precision while avoiding increases in device complexity
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides precise and economical measurements of solar radiation components, reducing errors through calibration and geometric projection models, and accounts for ground reflectance and skyline shape, enhancing accuracy and robustness.
Implementation Method 1
a plurality of photodetectors (131, 132, 133) sensitive to different spectral bands of solar radiation
Data Source
Figure 1~3
AI summary
A system for measuring solar radiation by means of a polyhedral solar sensor comprises a group of sensors arranged according to faces of an upwardly inclined polyhedron and a second group of sensors vertically arranged. The sensors are able to detect solar radiation on three different bands, namely infrared, visible and ultraviolet. The processor integrated in the polyhedral solar sensor operates according to a predictive neural network model in which the sensors are the input nodes of the network, with the neural network also guided by an analytical algorithm suitable for implementing a geometric projection model that takes into consideration the orientation of the sensors, to measure the global solar radiation, direct and diffused, and advantageously also the shape of the skyline and the solar radiation reflected from the ground.