Browsing by Author "Du Plessis, Armand"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
- ItemThe Influence of Dust Soiling on the Performance of Photovoltaic Modules in the Semi-Arid Areas of South Africa(Stellenbosch : Stellenbosch University, 2017-03) Du Plessis, Armand; Strauss, Johann M.; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: With various commercial photovoltaic (PV) power plants in South Africa located in the semiarid areas of the Northern Cape, this thesis provides field generated data for PV modules subjected to this environment. The effects of dust soiling, concerning the performance of PV modules, are analysed and the results obtained serve as a quantitative and mechanistic understanding for PV system engineers and investors. In attempt to determine whether cleaning PV modules is a relevant option, experimental dust mitigation methods are investigated. These methods include the application of a hydrophobic anti-soiling coating, as well as the execution of biweekly and long term (six months) cleaning routines, consisting of water based (wet) and dry cleaning methods. Results of these mitigation methods are compared to that of modules exposed indefinitely. The research objectives are successfully investigated by means of the design and commissioning of a PV research facility. The facility consists of 16 stationary mounted polycrystalline (pc-Si ) modules, analysed for the six month period of May to October 2016. A single axis tracker (SAT) system, is also designed and implemented. This provides the required experimental platform for the investigation of dust soiling on four tracking pc-Si modules, during a three month period of mid-August to November 2016. Raw data validation is established with comprehensive weather monitoring (ambient temperature, wind speed, wind direction, rainfall, pressure, and humidity), plane of array irradiance (GPOA) and PV module back sheet temperatures recorded, in accordance with the IEC 61724 standard. A MasterController, an intelligent data logging and communications device, is also designed and implemented, which is responsible for the gathering of the meteorological on-site data, measured at one minute log intervals. Also, as specified by the IEC 61724 standard, an intelligent device capable of extracting I-V curves, from individual PV modules at a 10 minute interval is utilised. PV module power output is derived from the measured I-V curves, validated with a singlediode curve fitting routine. A comparative study between various modules is analysed with a performance ratio (PR), defined as the temperature and irradiance corrected performance factor of a PV module. A clearness ratio (CR) is also used to further quantify dust soiling for the stationary modules, which compares the PR of modules to that of two reference modules, cleaned biweekly. For the six month stationary module analysis, results conclude a maximum recorded reduction in CR of 2.7 %. A maximum ideal PR difference of ~1:9 % is recorded for both the coated and uncoated sets of long term exposed modules, compared to the short term cleaned modules. This maximum deviation in performance is recorded after a 75 day absence of rainfall. The analysis does suggest that a rainfall of about 6 mm, every four to six weeks, is substantial to maintain the CR of unclean stationary modules, within 1 % of the cleaned reference modules. Results further indicate little to no deviation in performance between dry cleaned stationary modules and a set of water (distilled) cleaned modules. Regarding the SAT modules, a maximum ideal PR difference of 5.5 % is recorded for a coated, as compared to an uncoated module. The applied self-cleaning capability of the SAT system did not yield any conclusive remarks regarding this dust mitigation method. It is concluded that the hydrophobic coating for both topologies, stationary and tracking, promoted dust soiling. Finally, the research also suggests that SAT modules, which adopt a horisontal resting position during night time, are more vulnerable to dust soiling than stationary modules.
- ItemShort-term power output forecasting for large multi-megawatt photovoltaic systems with an aggregated low-level forecasting methodology(Stellenbosch : Stellenbosch University, 2021-03) Du Plessis, Armand; Rix, Arnold J.; Strauss, Johann M.; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: Power delivered from utility-scale Photovoltaic (PV) systems is characteristically intermittent, due to a dependence on atmospheric variables. To manage this uncertainty of an intermittent PV power supply, researchers traditionally adopt a macro-level forecasting approach, where a single model is trained to emulate the behaviour of the entire PV system. However, as commercial PV systems continue to expand in size, there is a growing uncertainty regarding the ability of these macro-level models to capture the non-uniform, low-level power output dynamics of large multi-megawatt PV systems. In response to this knowledge gap, a novel aggregated low-level forecasting methodology is proposed. With state-of-the-art deep learning (DL) implementations of Feedforward Neural Network (FFNN), Long Short-Term Memory Recurrent Neural Network (LSTM-RNN) and Gated Recurrent Unit Recurrent Neural Network (GRU-RNN) models, the proposed methodology is compared to the conventional macro-level forecasting approach. With data obtained from a commercial 75 MW PV system, multi-step 1 - 6 h ahead forecasts are delivered for a realworld scenario. Forecast models are trained for each of the 84 inverters, which collectively serve as the aggregated low-level forecasting solution. However, given the high computational expense of training multiple forecast models, a unique and scalable inverter-clustering approach towards model development is presented. The discrepancies in literature concerning biased model development are also addressed, with a heuristic process of systematic hyperparameter optimisation proposed, which serves to guide future forecasting practitioners. Concerning the results, this research successfully demonstrates the application of the proposed methodology. From the day-time-only forecast results, the aggregated inverter-level FFNN model shows the largest improvement, with a Mean Absolute Percentage Error (MAPE) of between 0.04 % - 0.4 % lower in comparison to the FFNN macro-level forecasts. This translates to an overall 30 kW - 300 kW improvement in forecasting accuracy. The aggregated GRU-RNN inverter-level model forecasts deliver a smaller overall MAPE performance increase, ranging between 0.03 % - 0.1 %. This is a 20 kW - 75 kW improvement. However, compared to all the DL forecast models applied, the low-level GRU-RNN model forecasts deliver the highest overall forecasting accuracy, with MAPE values ranging between 5.8 % - 8 % (day-time-only forecasts). From the 95 % Bootstrap con dence intervals, no improvements regarding the uncertainty analysis are observed for the aggregated low-level forecasting methodology. Finally, with this research it is concluded that researchers who have and continue to propose DL-based forecast model solutions for smaller multi-megawatt PV systems, can be con dent in the application of these models as macro-level solutions.