Skip to main content

​
Summary

Irradiance optimization is a tracker angle calculation method that maximizes total rather than simply minimizing . It operates after the base tracking angle is determined (true tracking or backtracking) and overrides that angle when an alternative would provide higher total POA irradiance. PlantPredict implements two modes: a built-in optimization algorithm (Version 11+) and integration with the ArrayTechnologies API. The built-in algorithm evaluates candidate angles while accounting for tracker movement penalties and hesitation factors. If wind stow is also enabled, wind stow takes priority and overrides the irradiance-optimized angle when wind thresholds are exceeded.

​
Inputs

NameSymbolUnitsDescription
Tracker Angleα\alphadegreesTracker angle from tracking algorithm (true tracking or backtracking)
Front POA IrradianceGPOA,front(α)G_{POA,front}(\alpha)W/m²Front-side plane-of-array irradiance as a function of tracker angle
Hesitation Factorη\eta—0 to 1 factor representing control system inertia
Rotation Speedω\omegadegrees/sTracker angular rotation speed
Time IntervalΔt\Delta tsecondsWeather data time step

​
Outputs

NameSymbolUnitsDescription
Optimized Tracker Angleαopt\alpha_{opt}degreesFinal tracker angle maximizing POA irradiance

​
Detailed Description

​
PlantPredict Built-In Optimization

PlantPredict’s built-in irradiance optimization algorithm evaluates multiple candidate tracker angles to identify the angle that maximizes total POA irradiance. The algorithm operates only when GHI is non-zero. For each integer degree αi\alpha_i between 0° and α\alpha (from 0° to α\alpha when α>0\alpha > 0, from α\alpha to 0° when α<0\alpha < 0), the algorithm computes GPOA,front(αi)G_{POA,front}(\alpha_i) using the configured model (Hay-Davies or Perez) and selects the angle that maximizes it: αideal=arg⁡max⁡αi∈[0,α] GPOA,front(αi)\alpha_{ideal} = \underset{\alpha_i \in [0, \alpha]}{\arg\max} \, G_{POA,front}(\alpha_i) Optimization is applied only if the ideal angle provides higher POA irradiance than the baseline tracking angle (GPOA,front(αideal)>GPOA,front(α)G_{POA,front}(\alpha_{ideal}) > G_{POA,front}(\alpha)). If not, the original angle is retained: αopt=α\alpha_{opt} = \alpha. Historically, irradiance optimization algorithms have over-predicted field gains because they idealize tracker mechanics. Real trackers do not move instantaneously, consume energy when moving, and face asymmetric risk—the penalty for being at the wrong angle during clear conditions far exceeds the gain from optimization in diffuse conditions. To address this, PlantPredict introduces two correction factors. Movement Penalty (μ\mu): Represents the fraction of the timestep the tracker spends rotating from its current position to the ideal position. During transit, the tracker is at intermediate angles rather than the optimal angle. Instead of integrating over all transit angles, the algorithm approximates this using the midpoint between start and end positions. μ=∣αideal−α∣ω⋅Δt\mu = \frac{|\alpha_{ideal} - \alpha|}{\omega \cdot \Delta t} Hesitation Factor (η\eta): An empirical term representing control system reluctance to adopt the idealized angle. This accounts for sensor uncertainty, unknown future weather conditions, energy consumed during movement, and the asymmetric risk of incorrect positioning. The adjusted hesitation factor is constrained so combined penalties do not exceed 1: ηadj={η,if μ+η≤11−μ,if μ+η>1\eta_{adj} = \begin{cases} \eta, & \text{if } \mu + \eta \leq 1 \\ 1 - \mu, & \text{if } \mu + \eta > 1 \end{cases} Corrected Angle Calculation: When optimization is applied, the optimized angle is a weighted average of three contributions: αopt=(1−μ−ηadj)⋅αideal+μ⋅αideal+α2+ηadj⋅α\alpha_{opt} = (1 - \mu - \eta_{adj}) \cdot \alpha_{ideal} + \mu \cdot \frac{\alpha_{ideal} + \alpha}{2} + \eta_{adj} \cdot \alpha
  • Ideal contribution: Time at the optimal angle, weighted by (1−μ−ηadj)(1 - \mu - \eta_{adj})
  • Traversal contribution: Time spent rotating, approximated at the midpoint angle, weighted by μ\mu
  • Hesitation contribution: Bias toward the original tracking angle due to control system conservatism, equal to the original angle weighted by ηadj\eta_{adj}
When both factors are zero, the algorithm uses pure idealized optimization. When tuned appropriately, the factors produce more realistic estimates matching field observations.

​
ArrayTechnologies External Optimization

When configured to use ArrayTechnologies mode, PlantPredict sends site parameters, tracker configuration, and weather data to the ArrayTechnologies API. The service performs proprietary optimization calculations and returns time-series of optimized tracker angles. PlantPredict applies the returned angles directly, bypassing internal tracking and optimization calculations. The ArrayTechnologies algorithms are proprietary and not documented here.

​
References

  • Kelly, N. A., & Gibson, T. L. (2011). Increasing the solar photovoltaic energy capture on sunny and cloudy days. Solar Energy, 85(1), 111–125.
  • Marion, W., & Dobos, A. (2013). Rotation Angle for the Optimum Tracking of One-Axis Trackers. NREL Technical Report NREL/TP-6A20-58891. https://doi.org/10.2172/1089596
  • Passow, K., Fusaro, D., Moseley, J., Shah, S., & Lee, K. (2022). Strategies to Optimize and Validate Tracking Performance of Single-Axis Trackers on Diffuse Sites. 2022 IEEE 49th Photovoltaic Specialists Conference (PVSC).
  • Rodríguez-Gallegos, C. D., Gandhi, O., Panda, S. K., & Reindl, T. (2020). On the PV Tracker Performance: Tracking the sun Versus Tracking the Best Orientation. IEEE Journal of Photovoltaics, 10(5), 1474–1480. https://doi.org/10.1109/JPHOTOV.2020.3006994