A PV module’s output depends on various factors. One such critical factor is the amount of direct irradiance collected by the modules from the Sun. The radiation received by the system varies throughout the day as the angle at which the Sun's rays fall on the modules changes. Maintaining a fixed amount of radiation on the modules throughout the day will increase the overall amount of solar energy collected by these systems.
A solar tracker follows the Sun’s trajectory and maintains the modules at a normal angle to the Sun's radiation all time during the day. The best way to track the Sun would be to follow its path from east to west and/or as it rises to its highest point in the sky.
The benefit of any tracking system is to absorb the solar energy for the longest period of the day, and with the most accurate alignment as the Sun shifts its position with the seasons.
Despite several advancements, the tracking techniques still face numerous hinderances. An optimized and controlled system is paramount to precisely point the position of the sun considering the regular changes in the daily altitude angle, seasonal latitude offset as well as changes occurring in azimuth angle of the sun [1].
The first means of categorizing a solar tracker is by its means of actuation. Solar trackers can be either manually, passively, or actively controlled.
• Manual trackers depend on personnel to gradually adjust the solar tracker throughout the day. This method is costly and not ideal.
• Passive solar trackers use a liquid with low boiling point. It evaporates and causes an imbalance to turn the tracker towards the direction of the Sun’s rays. They require minimal moving parts compared to active solar trackers.
• Active solar trackers rely on actuators such as electric motors or hydraulic cylinders to change the position of the solar tracker.
The second means to categorize the solar trackers is by their axis of rotation. The solar tracker may be a single or a dual axis tracker.
• A horizontal axis solar tracker will track the Sun as it rises and falls in the sky. It has only one degree of freedom and moves on a North to South axis of rotation. It is used in locations with low latitude and is widely used for commercial applications.
• A vertical axis solar tracker also has only one degree of freedom and can move on axis of rotation from East to West. However, they are better suited for higher latitudes.
• A dual-axis solar tracker has two degrees of freedom. These can track the Sun on both East to West axis and North to South axis. Although these can track the Sun’s rays more accurately than the previous two, the major disadvantage is the initial cost of setting it on a large scale.
Saxena et al., (1990) designed a model where the solar position is calculated as a function of time. A highly accurate angle measuring device, such as a digital shaft encoder, must be installed on the rotating axis to position the PV module to the calculated angle.
With the recent breakthrough in artificial intelligence, artificial neural networks (ANNs) have gained significant importance to actively control the solar trackers with utmost precision. Neural Networks (NNs) are made of processing elements called neurons (nodes) and are often treated as black boxes. The sole purpose of those nodes is to accept different inputs and produce desired outputs. They are trained on large sets of data to gain new patterns, associations, and functional dependence.
The solar modules depend on inputs like temperature and solar irradiance which exhibit highly nonlinear behavior and are subjected to perturbations (change in wind speed, humidity, the solar intensity, seasons).
As solar modules are designed to operate for 20 to 25 years, it is expected that their parameters will also fluctuate over that long period. Therefore, it is recommended to apply artificial intelligence control techniques which are also non-linear in nature to effectively control the solar trackers (Chekired, 2011).
How to train your Neural Networks
In a non-linear control system, different logics are in place to train the NNs. Among those, the ones which are easy to implement with higher efficiency are Perturb and Observe (P&O), and Fuzzy Logic.
• The Perturb and Observe (P&O) method compares the last calculated output power P(k) with its past value P(k-1). The difference between these two values is observed to check if the output power of the system is increasing or decreased. The goal is to continue with the iterations till the error is zero. If the power increases, the maximum power point (MPP) will be approached, the same operation continues for further perturbation, else it moves in the reverse direction.
• Fuzzy logic has recently gained superior importance in designing and implementing nonlinear controllers because of its simplicity, ease of design and implementation. The control of knowledge-based systems using linguistics variables that do not have precise values is of concern, and this allows the use of traditional human experience in designing the system (Cavallo et al., 1996). The basis of this approach is the basis for human communication and fuzzy system can be created to match any set of input and output data. It has tolerance for imprecise information.
A comparative study of these two algorithms has been done using MATLAB. Out of these two algorithms, MATLAB Simulink results have shown that fuzzy controllers perform rapidly, with better accuracy than their P&O counterpart
Recent trends involve designing solar trackers using fuzzy logic as the training algorithm to achieve higher performance. In the block model below, the NN identifier is placed in parallel with the solar tracker. Both are provided with the same input u(k). The error value between the ST and the NN is used as a training signal. It is fed to the fuzzy logic algorithm, and the output is used to train the identifier. The training of the NN is dependent on determining the correct weights to the nodes. These weights are applied to both the hidden and output layers.
Illumine-i has extensive experience in projects involving solar trackers. In a recent Texas project, the system was designed to optimize solar energy collection utilizing advanced tracking technology to automatically adjust the position of solar panels, enabling them to follow the sun's path throughout the day.
Our designers drafted the setup and mounting details to closely resemble the actual workings of the dual-axis solar tracking system. We also analyzed the depth and footing to ensure that the dimensions of the foundation align with the specifications outlined in the installation manual. Post installation, the controller started within 5 minutes and the host control unit downloaded the GPS data automatically.
Researchers have proven that solar trackers that use fuzzy logic controllers exhibit significant performance boost. As solar energy continues to lead the transition toward clean energy, these advancements ensure that solar trackers can meet growing global energy needs more efficiently.
