This thesis addresses the imperative need for sustainable energy solutions in the residential sector which is a significant contributor to global energy consumption and greenhouse gas emissions. Focusing on modeling and optimizing residential energy systems, the study explores the integration of photovoltaic (PV) panels, heat pumps, and batteries. The purpose is extending the applications of the Photovoltaic Material and Devices (PVMD) toolbox, offering a framework for future research and advancements. The objective is to enhance energy efficiency, reduce carbon footprints, and contribute to energy security by diversifying energy sources.
The research encompasses a comprehensive analysis of heat pump models, emphasizing their role in space heating, space cooling, and domestic hot water functions. Two specific models are chosen for Coefficient of Performance (COP) and Energy Efficiency Ratio (EER) calculations. These models, along with the utilization of the nPro tool, lay the foundation for integrating heat pump systems with PV production and battery storage in residential buildings.
The study successfully integrates the model of the battery’s performance and the overall grid-connected energy system. Employing a mathematical modeling approach, each component is systematically incorporated into the system, including the previously developed heat pump model. This integration, coupled with the Alternating Current (AC) output of the PVMD toolbox and the battery, establishes the groundwork for subsequent economic and performance analyses of the system.
The study systematically selects various locations with different environmental conditions such as Equivalent Sun Hours (ESH) and average ambient temperature to analyze the economic aspect by checking the Net Present Cost (NPC) and performance aspects by checking Self Consumption Ratio (SCR) and Self Sufficiency Ratio (SSR) of the integration model. The findings emphasize the economic viability of heat pump investments in cities with distinct heating and cooling demands. It has been demonstrated that in colder cities where heating demand is predominant, heat pumps are economically attractive, resulting having heat pumps in optimal scenarios that give the minimum NPC in Amsterdam and Lisbon. Additionally, although the individual components of the system may seem cost-ineffective, their value is derived more from integration in milder cities where heating demand is dominant, resulting having heat pump and PV integrated for the optimal scenario for Lisbon. However, cities dominated by cooling demand face challenges in achieving financially optimal designs because the operational savings for cooling cannot be accurately included such as Cairo and Dakar. The research underscores the importance of considering various system factors, including initial investment costs and electricity tariffs, to achieve financially optimal sizing. As the initial cost of the battery decreases, battery technology becomes economically appealing for Lisbon and Dakar. Moreover, changes in the tariff prove economically favorable for integrating the battery system in Lisbon, Cairo, and Dakar.
This work contributes valuable insights to the field of renewable energy, providing practical solutions for the transition towards cleaner and more efficient residential energy systems.

With advancements in the photovoltaics (PV) market, involving increased PV module efficiency and reduced costs, the logical progression is the integration of PV into various surfaces, including vehicles (VIPV). For a driving VIPV, the constant change in irradiance presents a significant challenge. The location, weather, and time are key factors that affect the irradiance prediction. An accurate prediction of solar energy generation is crucial for realizing the full potential of VIPV. In this study, the incident irradiance loss is referred to as shading loss.

This thesis is conducted in collaboration with Delft University of Technology (TU Delft) and the Netherlands Organisation for Applied Scientific Research (TNO). TNO has an existing shading loss model that accounts for the surrounding obstructions and weather effects using a fixed value, where the only variable input is the day of the year. On average, it assumes 27% shading losses yearly. This study has improved TNO’s VIPV shading assessment by incorporating more variables into a new model, thereby contributing to a better estimation of the VIPV energy yield. The inputs of the new model are based on key factors causing irradiance loss, where location data is obtained using a Digital Surface Model (DSM). The new contribution to the DSM-based approach is the Web Coverage Service (WCS), which retrieves height data directly through a URL, eliminating the need for the previous manual operations and building upon the model’s scalability. The improved model determines shading losses along a route using two varying shading indicators: Sky View Factor (SVF) and Shading Factor (SF). It can input various irradiance datasets, including or excluding weather effects, and precisely determines the Sun’s position based on the time and day of the year. The DSM-based approach is validated using a dynamic and static validation dataset, with a SVF≤1. The accurate representation of surrounding obstructions during validation contributes to a good-quality model. Both validation studies differ in inputs based on location, time, and weather conditions. The differences in measured and simulated irradiance values are larger in the dynamic dataset than in the static dataset. Deviations are caused by uncertainties and errors, such as local weather effects, sensor inaccuracies, traffic shading, and LiDAR discrepancies, which is the inability to simulate open or changing structures. Weather variability is the primary factor affecting model quality in this validation study, resulting in no fixed inaccuracy factor for model validity. The validated model can accurately represent the environment, but weather variability should be minimized when comparing measured and simulated irradiance. Therefore, averaging is chosen to compare results in this work as it removes unknown or untraceable data points.

Beyond previous research studies, this work focuses on two public transport case studies, given the systematic routing with a fixed timetable. The first case study includes a car and a bus, and the second, more extensive case study focuses solely on buses. The choice of inputs in both case studies centers on key factors influencing VIPV irradiance loss. In the Bus 40 study, simulations under clear sky conditions and a fixed time around solar noon resulted in shading losses of 22.7% for the bus and 27.7% for the car. Moreover, the simulated shading loss shows a 2.6% relative difference compared to the yearly average TNO value. For the case study considering 27 bus routes in Amsterdam, simulations are performed using averaged KNMI irradiance and all hours a day, where the annual average shading loss resulted in 24.8%. The SF and clear sky irradiance correlate in the first case study and the SVF and average KNMI irradiance correlate in the second case study. However, using clear sky conditions and uniform overcast conditions does not represent realistic weather conditions and leads to a bias in the correlation. Moreover, classified terrain types do not correlate with SVF, in contrast to the reversed approach by Araki et al.

Finally, the model can effectively characterize the shading losses on the route level. However, this methodology falls short when reaching finer resolution, such as when examining smaller route segments or partial shading. The determination of shading losses at the route level can be of further benefit to various stakeholders, such as bus operators who want to determine the optimal bus routes for maximum solar energy generation. In addition, this model can be helpful for policymakers, to predict the required charging infrastructure in a scenario when vehicles are equipped with solar panels. Moreover, this model improves shading loss estimation by accepting a range of datasets instead of TNO’s current approach. By inputting irradiance data for clear skies to complete overcast skies, the shading losses can be predicted in a safer range instead of over- or underestimation based on one shading value.

To make silicon thin film solar cells attractive to the market, their efficiencies should reach the efficiencies of the dominant crystalline silicon solar cell. Therefore the search for ways to improve the efficiency of silicon thin film solar cells continues. TCO front contact layers and back reflector layers play a significant role in the increase of thin film solar cell efficiencies. For the optical characterization of these TCO layers, commonly used methods are found to differ significantly in results regarding the extinction coefficient which makes their accuracy questionable.

This research consists of two main parts. For the first part, the commonly used optical characterization methods spectroscopic ellipsometry and spectrophotometry + data analysis in SCOUT are analyzed in accuracy and compared to newly introduced methods using spectrophotometry + data analysis in GenPro4 and photothermal deflection spectroscopy + data analysis in GenPro4 to build a guide on optical characterization of TCO materials. For the second part, the optical response determined using the software GenPro4 of a double junction silicon thin film solar cell for a novel bi-layer front contact design consisting of IOH and i-ZnO will be compared to the optical response for standard used AZO, and ITO single layers to find the best front contact design. Furthermore, the optical response of a double junction silicon thin film solar cell for a back reflector containing an i-ZnO layer on top of the silver back contact will be compared to the optical response for a back reflector containing an AZO layer on top of the silver back contact to find the best back reflector design.

From the results of the optical characterization methods, it is concluded that photothermal deflection spectroscopy + data analysis in GenPro4 is the most accurate method for determining the extinction coefficient of a TCO material. The results of the optical simulations for front contact TCO and back reflector TCO designs showed that the bi-layer can enhance the optical response of a double junction silicon thin film solar cell significantly compared to the AZO and ITO single layers. The i-ZnO TCO back reflector layer was found to induce less parasitic absorption and therefore a better optical response of the double junction silicon thin film solar cell.

Batteries play a crucial role in powering contemporary devices and systems, ranging from smartphones and electric vehicles to renewable energy storage. With the increasing demand for more efficient and reliable battery technologies, the need for accurate monitoring and assessment of battery condition and performance has also grown. Online estimators, that continuously analyse battery conditions in real-time, have emerged as valuable tools to meet these goals. This thesis focuses on constructing model-based online state estimators for real-time estimation of a battery's State of Health (SOH).

These estimators, necessitate a state space model for a lithium-ion cell. In the initial part of the thesis, a physics-based reduced-order model (ROM) of a lithium-ion cell is developed. This model accounts for major aging mechanisms such as Solid Electrolyte Interphase (SEI) layer formation, Loss of Active Material (LAM), and Lithium Plating (LIP). To incorporate temperature effects on cell parameters, a simplified lumped thermal model is integrated into the battery model. The model is subsequently transformed into a state space model using the Discrete Realization Algorithm (DRA) process.

Building upon the noisy outputs from the ROM models, a set of five estimators is formulated: State of Charge (SOC), Voltage, SEI loss, LAM loss, and LIP loss estimators. These estimators are constructed based on Kalman filters and collectively contribute to real-time prediction of a battery's SOH. Given their reliance on the model, the ROM model's output is employed as a reference to gauge the precision of the estimators.

To address the real life scenarios and to check the robustness of the estimators, a series of sub-questions were analysed:
Sub question1: How to estimate the SOH of a cell in real-time using adaptive control techniques ?
Sub question2: How fast can the estimator react to changing initial states of the cell?
Sub question3: Can the estimators be adapted to concurrently estimate both the states and the time varying cell parameters in case of an old cell? How fast can this be
achieved in real-time ?

Overall, this research contributes to the development of model-based online estimators, which are poised to have a significant impact on enhancing battery performance, prolonging lifespan, and facilitating the transition towards a more sustainable energy future.

Emerging PV technologies like Perovskite has lead to the development of Perovskite/Silicon (Pvk/c-Si) tandem device (Multi-junction device) and has now gained a lot of momentum and attention due to the fact that the tandem device can reach a Shockley-Quisser limit of about 44.1% efficiency. But to speed up its development, a lot of emphasis is made on conducting tandem solar cell device simulations to have get a good idea on the experiments that can be carried out. Solar cell simulations for tandem devices are carried out in advanced 2D/3D solar cell simulators like SETFOS or Synopsys TCAD Sentaurus. These softwares have large computational time and have complex user interface making simulations challenging. There are 1D solar cell simulators like ASA (Advanced Semiconductor Analysis) developed at TU Delft that has less computational time with an easy and intuitive interface. But ASA does not have the appropriate tunnelling models to simulate the tunnel recombination junction characteristics of perovskite silicon tandem devices.

To make ASA a complete software suite for simulating perovskite silicon tandem device, tunnelling models proposed by Ieong et al. is used for both direct and band to band tunnelling. These tunnelling models use basic inputs like band energy data, electrostatic potential data and electric field data for a particular device and outputs the tunnelling generation rate that can simply be added to the continuity equations for electrons or holes during the simulation of the device. A fully fledged algorithm was designed for the automation of the direct tunnelling and the band to band tunnelling process with limited input from the user while using the models proposed by Ieong et al. In simple terms, the algorithm takes in the required input and scans through the device for every interface to check for tunnel contributions and performs calculations when necessary. All necessary conditions are included inside the algorithms to ensure that the simulation is accurately solved.

The validation of the newly developed tunnelling algorithm was commenced for PN junction, Silicon hetero-junction and finally the Perovskite-Silicon tandem device. The JV curve for the above device simulations with and without incorporated tunnelling models was extracted and a comparison was made with literature results and simulation results from other software suites. The newly developed tunnelling algorithms was concluded to be accurate. The JV curve obtained from the ASA simulation showed good agreement with literature results and simulation results from other software suites with error percentages of less than 3% for open circuit voltage, short circuit current and fill factor. With proper tuning of the device layer thickness, texturing on multiple layers for the tandem device and enhancing the mobility of the non-absorber layers for the perovskite top cell, further improvement can be expected in the simulation. These methods can eventually lead to reaching the Shockley Quisser limit of about 35.7% efficiency for the perovskite silicon tandem solar cell even when general opto-electrical losses are taken into account.

In this thesis, a new photovoltaic fault detection and classification method is proposed. It combines the generation of a synthetic photovoltaic training database and the use of a machine learning model to detect and classify faults in small-scale residential PV systems. The database was generated in Matlab, and the machine learning modeling was done with the scikit-learn library for Python. From the modeled PV systems, solely power yield is used as an indicator, combined with system age and meteorological conditions. Using these features, four types of machine learning models are used to detect malfunctioning PV systems and classify short-circuit faults and open-circuit faults. This thesis also shows the benefit of a synthetic PV training base as opposed to alternative methods, with increasing performance due to control of database balance.

The result of this thesis is a method that can be used to construct a model for detection and classification of photovoltaic faults, specific to a single residential PV system. Malfunctioning systems can be detected with an accuracy of 80.4% using a random forest algorithm. For fault type classification, an F1-score of 0.759 was achieved, also using a random forest.

To further improve the light management and optical performance of solar cells, research into nano-textured interfaces has led to a need for a fast and accurate wave optics model. Currently, within the PVMD group at the TU Delft an optical solar cell simulation tool, GenPro4, is being developed to allow for the quick optimization and analysis of solar cell designs. However, the implemented wave optics model has limited accuracy and thus an alternative model is sought.

In this thesis, we researched several rigorous wave optics simulation methods such as: finite element method, finite difference time domain, transfer matrix method, and rigorous coupled wave analysis. Based on three decision criteria (speed, accuracy, and compatibility) the choice of implementing rigorous coupled wave analysis (RCWA) was made. Besides the standard RCWA formulations, a few improvements were made. S-matrices were used to increase
speed and memory efficiency of the program. Furthermore, calculation of the local E-field allowed for the determination of absorption per material in our cell. Lastly, an angular intensity distribution, known as scatter matrix, was produced as output to allow for the full integration with GenPro4.

We validated our new model by comparing results of a nano-textured GICS solar cell with previously conducted finite element method (FEM) simulations. Over a large wavelength range our model showed good agreement with the FEM data. However, slight inaccuracies were observed at longer wavelengths (>1000 nm) as simulation of the metal back reflector became infeasible. Highly absorbing materials, such as metals, potentially required several hundred Fourier modes to be modeled accurately. However, simulations with such a high number of Fourier modes is not feasible as the computer resources (RAM) and computational time
needed grow exponentially. Furthermore, for very small textures, RCWA converges with the GenPro4 flat model as is expected for sufficiently small textures.

Nevertheless, integration of RCWA with GenPro4 was achieved. To demonstrate the ability to quickly and easily simulate real work cell designs, a case study of a perovskite/silicon tandem cell was conducted. In the case study, a combination of 16 interface mythologies was simulated and compared. With the optimal combination showing a potential improvement of 2% over the reference cell.

Solar photovoltaic (PV) technologies offer a renewable alternative to power generation that is based on fossil fuels; however, to meet the rising demand in electricity a substantial amount of surface area is required. This will inevitably lead to the installment of these systems on agricultural land, subsequently intensifying the land-use conflict and threatening food security. To alleviate this, the use of agrophotovoltaic (APV) systems is investigated, which enable the simultaneous production of food and renewable energy. More specifically, the aim of this thesis is to determine the optimal topology for a medium-scale and stationary bifacial APV array, which is simulated under the climate of Boston, USA (42.37˚N, 71.01˚W). Irradiance modelling is performed with Radiance’s raytracing algorithm in combination with the daylight coefficient approach of Daysim and the Perez All Weather sky model, which is then coupled to the crop and electrical yield models to determine the overall land productivity. The modelling approach used is robust, while offering flexible manipulation of the array’s deployment configuration, which is crucial for multivariable optimization. The integration of bifacial PV offers various synergistic effects mainly due to the amplified ground irradiance necessary for crop growth. Owing to the decreased PV density and high elevation from ground, rear irradiance homogeneity and bifacial gain are enhanced. Widening of the row spacing resulted in a logarithmic increase of the incident ground irradiation, while the overall energy yield portrayed a negative exponential trend, which is attributed to the use of bifacial modules. East-west (E-W) facing and vertically installed topologies are better suited for shade intolerant species, or permanent crops since they permit additional light penetration, especially during the winter months. In contrast, south-north (S-N) facing and latitude inclined arrays lead to intense and non-homogeneous shading that is unfavorable for growth during winter or for crops that cannot acclimate to shade; nonetheless, energy yield and land equivalent ratio (LER) are significantly enhanced. Unlike previous studies where only conventional modules were examined, here the potential of a customized one is inspected to assess whether blueberries can grow effectively under shade. By integrating such a module in an E-W hinged PV topology, which is associated with the most optimal shading patterns and schedule, the agrivoltaic performance is optimized. In comparison to the reference case, the land’s productivity is increased by 59% with a reduction in electrical yield by a third. Through this holistic approach that incorporates a multi-scale sensitivity analysis, it is possible to achieve a spherical understanding of the limitations and potential synergies associated with the dual use of land, ultimately encouraging the sustainability of the agricultural sector.

Transparent conductive oxides (TCOs) play important roles in information and energy technologies. In the photovoltaic (PV) community, they are normally required to provide sufficient lateral carrier transport towards metal electrodes at the illumination side. Thus, a trade­off between optical and electrical
properties is of critical importance in relevant device fabrications. So far, within the PVMD group, the investigation on TCOs has been mainly focused from an experimental perspective. In this thesis, a firstprinciples approach based on DFT software is investigated to obtain the self­consistent opto­-electrical properties of TCOs.

A comprehensive study is carried out to understand the fundamentals of DFT software, as this is crucial to get reliable results from it. Within the PVMD group, the DFT method is applied on the indium­oxide (IO) host material. It was found that referable results for the (partial) density of states and band
structure could be obtained for this structure using the PBE exchange­correlation (XC) functional. The dielectric function could be obtained by combining the PBE and HSE06 XC functional through the PHSmethod.

Based on a preliminary validation of IO, two case studies are carried out. In these case studies, it is investigated if DFT can be used to compare the opto­-electrical properties of different doping types and ratios. This is done for post­transition metals (Sn), transition metals (W and Mo) and anionic doping
(F). By observing the partial density of states of each element, it was found that significant hybridization of dopant states with the CBM of the IO host states occurs for the cases of Sn­ and F­doping.
Such disturbance in the host conduction band may lead to detrimental influences to the opto­-electrical properties of corresponding TCOs. However, in the cases of W­ and Mo­doped TCOs at commonly used doping concentrations, no hybridization between dopant states and the host conduction band was observed. Furthermore, physical parameters of different TCOs at different doping levels are extracted and compared, such as band gap, effective electron mass, work function and dielectric functions. These results may provide supportive and indicative information for the experimental work.

This thesis work has successfully introduced DFT calculation into TCO investigations within our research group. Although the preliminary results were not sufficiently accurate to predict the opto­-electrical properties of the TCOs in a quantitative way, the qualitative trend can still be used as guidance and support for explaining experimental results. However, many challenges still remain, especially for determining some optical properties like the band gap and dielectric function. Further research is still needed to improve the proposed method.

The installation of photovoltaic systems in the urban environment is becoming increasingly common. These photovoltaic systems are affected by a large fraction of reflected irradiance as well as partial shading. In a similar way these issues also have an important effect on bifacial photovoltaic power plants. Irradiance models that are able to take these issues into account are therefore gaining interest. For an accurate economic assessment of a PV system and the calculation of delivered power it is becoming very important to be able to accurately determine the irradiance on a PV system. Therefore there is a rising demand for accurate irradiance models that are able to take into account the effects that arise in complex geometric scenarios. The research in this thesis report aims to develop such a model which is able calculate the irradiance incident on bifacial modules and PV arrays in complex landscapes by combining the concepts of ray tracing and 3D view factors. The implementation of the proposed model decouples the irradiance simulation into independent blocks that allow to efficiently calculate the spectrally-resolved irradiance incident on a PV module. This decoupling makes the proposed model suitable for simulation of tandem devices in the near future. The developed simulation has been experimentally validated using measurements from the PVMD monitoring station and also in comparison with other sophisticated irradiance simulation models. Experiments with a large fraction of specular reflected irradiance showed a good match between the measured and the simulated irradiance. Validation over a longer time period was also performed for 3 different sensors on the PVMD monitoring station roof. The results show an overall low mean bias error between the measured and simulated irradiance for the time period between mid-August 2020 to mid-October 2020. The proposed model is able to model the irradiance impinging on a PV system in a complex urban environment using a decoupled structure and its performance is comparable to sophisticated existing ray tracing models. The performed simulations show a good match with the measurements for different situations. Further improvements are the implementation of higher order reflections, computational optimization and more extensive validation.

Renewable energy sources (RES) such as Solar and Wind energy rely on the availability of natural resources like sunlight in the case of Solar and wind speed in the case of Wind energy generation which is variable in nature. There are periods where there is excess energy production than needed and periods of energy shortage where not enough energy is produced to meet the demand. To mitigate this mismatch, a short term solution is to use batteries in order to store energy at times where energy production is more than the energy demand. This stored energy would be later used at times where energy production is low and meet the energy demand. However, the current battery technology is still novel for this application making it uneconomical when compared to current energy infrastructure of using power plants. The current battery market is held by Li-ion batteries which uses lithium as a raw material which is a rare earth material. In 2009, a battery cell utilizing Si as its anode and air as its cathode was discovered. As this system relies on two of the most abundant elements in the earth's crust which is silicon and oxygen and has much higher theoretical energy density than Li-ion batteries, it has become a growing area of research and development. Battery models are created to simulate battery operations based on empirical formulas and electrochemical reactions taking place in the battery. Development of these models are very critical as they provide results and optimum condition evaluations much faster than physical testing with minimal resources. A battery model for the alkaline Si-air battery which utilizes KOH as the cell electrolyte is developed in Simscape (MATLAB) as part of this thesis. The modelling parameters are also subjected various physical conditions such as varying electrolyte concentration and change in electrode materials and the variation is investigated for model validation to study whether changing physical conditions of the Si-air cell has an effect on the modelling cell parameters. It is supported with experimental results obtained from discharging a fabricated Si-air cell. It was concluded that there are cell parameters which are dependent only on the state of charge (SOC) of the cell and one cell parameter that is a function of both the SOC as well as the discharge profile of the cell. The fabricated Si-air cell gives higher open-circuit potential (OCP) values than what was reported constant 1.4 V in literature which is speculated to be due to the usage of a 99% Aluminum and 1% Silicon (Al:Si) back contact layer. Average OCPs ranging from 1.5 V to 1.45 V which varies due to change in electrolyte (KOH) concentration is achieved. The MATLAB battery block is calibrated to be integrated with energy system models as a Si-air battery.

The future of our energy supply cannot continue to depend on the use of exhaustible fossil fuels. The successful transition to a society powered by renewable energy sources is one of the major challenges our current generation faces. Renewable systems will only fully establish themselves if high conversion efficiencies can be obtained at a reasonable cost, in order to compete with conventional, carbon-based sources. With this in mind, photovoltaic solar energy can be a promising solution.
This study focuses on the investigation of the potential efficiency of a thin-film tandem solar cell, combining amorphous silicon and microcrystalline silicon for a high performance device. With the use of flexible, light substrates, inexpensive modules can be manufactured in a roll-to-roll configuration to produce high efficiency products for a large scale implementation in society. The semiconductor modelling software ASA was used to optimise the structure of the tandem, identifying the absorber layer thicknesses that maximise the efficiency. Moreover, the impact of possible optical variations in the design of the device on the performance were analysed.
To start, model parameters of single-junction amorphous silicon and microcrystalline silicon cells were calibrated to recreate the performance of such devices based on experimental data. By conducting a sensitivity analysis of the model parameters, this was effectively achieved. The single-junction models were combined in a tandem structure to forecast the operation of the multi-junction device.
It was found that the performance of the tandem was highly sensitive to the choice of refractive indices of the layers at the junction between the two subcells. A high mismatch
between the refractive indices of these layers and the absorber layers of the subcells results in an increased reflection of light to the top cell. This is beneficial for the current production of this subcell, but is too detrimental for the operation of the bottom cell. By bringing the refractive indices closer together, an efficiency of 13.0% was predicted. The inclusion of an encapsulation at the front of the cell boosts this efficiency to 13.7%, at a top/bottom absorber layer thickness combination of 160 nm/0.6 µm.
The possibilities of enhancing the efficiency with an intermediate reflector were examined. The optical conditions for the ideal light distribution with such a reflector were
computed to provide estimations of the potential increase in performance. For the best performing device, an increase of the initial efficiency of 13.0% to 13.2% was foreseen, at slightly thinner top cells and similar bottom cell thicknesses.
All in all, the predicted efficiencies of the amorphous silicon/microcrystalline silicon tandem solar cell evidences the potential of these devices for the production of cheap,
highly efficient modules. The developed analysis of the ideal intermediate reflector can be generalised to other multi-junction solar cells to establish the potential efficiency increase such a layer brings along. This can help in the decision of whether the increase in performance justifies the added manufacturing complexity and costs.

Photovoltaic (PV) system yield prediction models are an important topic in the field of PV solar energy. An accurate prediction model could not only be used for optimising the PV system design, but is also expected to realise the yield potential of innovative PV technologies. In the next generation of PV technologies, one of the most promising concepts is the tandem solar cells. In recent years, these cells have impressed the solar industry by their rapid growth of the maximum power conversion efficiency (PCE). Since tandem solar cells are still at the lab phase, their yield potential under realistic conditions are an interesting field of study as well. However, the existing yield prediction models are not yet available for these tandem cells. In order to fill this gap, a prediction model developed in the Photovoltaic Materials and Devices (PVMD) group of TU Delft, called the PVMD Toolbox, has been adapted to be compatible with the tandem solar cells. In this thesis project, the version 3 of the Toolbox was developed. First the PVMD Toolbox was improved. Except monofacial c-Si cells and bifacial c-Si cell, it is now also available for tandem cells. One of the most promising tandem concepts, the perovskite/c-Si tandem, was taken as a reference tandem configuration integrated in the Toolbox. The Toolbox was also modified to take the influence of solar spectra into consideration and analyses cell performances individually. Secondly, the accuracy of the Toolbox was validated. A new figure of merit called 'relative total deviation' is put forward in this report, which calculates the ratio between the sum of deviations that each simulation introduces, and the total measured energy yield. When the electrical parameters were taken from the own measurements, the relative total deviation of simulation results was only 8.6%. Finally, the Toolbox was applied to two case studies. In the first study, it was found that the annual energy yield (AEY) of perovskite tandems can be increased up to 0.4% by optimising perovskite thickness for a certain location. The second case study compared the AEY of the tandem module and conventional c-Si module. Perovskite tandem module showed a high AEY, around 36% higher than that of the c-Si module. However, perovskite tandem had a lower specific yield as they are more sensitive to the spectral variation

Photovoltaic Technologies in recent years have gained immense attention owing to reduced costs and increasing efficiencies. While decades of research in Photovoltaic Thermal (PVT) technologies, producing thermal and electrical energy simultaneously have brought these costs further down, improving a system’s overall performance. PVT collectors consist of PV modules with a thermal absorber bonded/attached underneath them. Excessive solar radiation that is not converted into electricity by the panels is released as heat, causing their temperatures to rise. Rising PV panel temperatures have an adverse effect on their efficiency, particularly for building integrated photovoltaics, that generally lack sufficient ventilation for this heat release. Thermal absorbers for PVT systems, are designed specifically for absorbing the excess heat generated by PV modules. In practise, helping the panels perform better by effectively removing the heat present behind them, with the help of a heat transfer fluid.

This project has been a collaborative effort between the TU Delft, and Exasun BV, a solar panel manufacturer located in the Netherlands. At Exasun, the project has also benefited by being a part of a larger consortium project, the PVT inSHaPe, currently underway at the Solar Energy Application Centre (SEAC), in Eindhoven. PVT inSHaPe aims to realise zero energy buildings by integrating PVT systems with heat pumps alongside effective thermal storage. As a manufacturer, Exasun BV specialises in state of the art building integrated photovoltaic systems (BIPV). BIPV systems aim to integrate photovoltaic technologies seamlessly into building facades. In doing so, they forego essential thermal ventilation required for maintaining lower panel temperatures. Thus, a novel BIPVT design was developed in-house at Exasun, for extracting the excess heat of panels, and utilising it to match the domestic hot water and space heating demand for Dutch households.

A simple thermal model for concentrating PV-Thermal collectors, currently under development at the TU Delft was validated alongside widely used steady-state and quasi-dynamic thermal models. Individually calculated thermal efficiencies from the models were juxtaposed, with the simple thermal model recording an error of 1.65% against the steady-state model, and an error of 9.21 % against the quasi-dynamic model. Once validated, the model was used further for system characterisation and performance evaluations of the design.

Various technology concepts have been tested extensively. However, further feasibility, reliability and optimisation studies need to be performed, in order to test the novel, cost-effective, and relatively maintenance free design in mind. The performance of PVT systems rely on high irradiance levels from the sun, and module temperatures. As space heating demands are higher during winter months, even after heat pump integration, the system is not effective enough to match the entirety of the load demand, and must be coupled with an auxiliary (electrical) heater, that can be powered by the PV system. Presumptive performance analysis carried out for a simulated household, revealed an average thermal efficiency of 10 % for the design, while recording a combined efficiency of 36 %. The stand-alone system was able to match over 55 % of the domestic hot water demand. When combined with a heat pump, the system is able to meet roughly 80 % of the hot water demand, while it is able to match almost 40 % of the complete thermal demand for a household.

In the process of replacing carbon based fuel with renewable energy, into the energy generation cycle of the world, mobility is considered to be one of the important sectors in which enormous progress can be implemented. Business tycoons are introducing electric vehicles which fascinates people. However, there is yet a stabilized solution to be found for charging electric vehicles with green energy on the basis of large scale implementations. In the Netherlands, there are more bikes than the total population and with the number of e-bikes increasing every year it would be brilliant to charge the e-bikes with decentralized renewable energy than with the conventional way of purchasing energy from the grid. In this thesis such a charging facility called the Solar Mobility Hub is studied thoroughly and validated for performance.

By validation, it means the system was tested for the number of e-bikes it can charge throughout the year as an autonomous system without loss of load. Instead of going into complete modelling a different approach was pursued. E-bike ride programs were conducted for a few months with the help of volunteers from which technical data necessary to develop exclusive full year load profiles were obtained. Thereby, generating two kind of load profiles - work to home (commute mode) and bike sharing (sharing mode). It was found that the solar mobility hub was capable of charging 19 e-bikes as a 100 % reliable autonomous charging infrastructure for commute mode and 15 e-bikes for sharing mode. Furthermore, advanced simulations were performed with system modifications like change of PV array tilt and increased battery bank. Finally, a performance analysis on the system was conducted and the reasons for the variation in performance were analysed. Furthermore, improvements and suggestions necessary for making such a BIPV commercially successful are provided.

Among the advantages of crystalline silicon hetero-junction (SHJ) there is the possibility to use a thin c-Si wafer that is bendable, but still has a power conversion efficiency higher than commercially avail- able thin film (TF) modules. Nowadays, all the photovoltaic applications on curved surfaces use TF technologies, however their low power conversion efficiency requires a large area, which is not always available. Therefore, the entry of SHJ technology in the environment-integrated applications will de- crease the required PV area being a competitor of TF wherever the surface available is scarce and the the shape curved and not complex.

A structure that would benefit the advent of SHJ modules is the infotainment spot, which is a solar- powered device that provides useful information and entertainment to the user through an interactive screen. Such a device was realized for the first time by Weeda et al. [1] in TUDelft, but the CIGS module that powered the device made the structure bulky, unattractive and unwieldy. In this thesis the experience of Kaneka Corp., which holds the world’s record of highest conversion efficiency for c-Si hetero-junction, was used to design the PV module that powers the stand-alone system of a new infotainment spot.

In order to calculate the irradiation on the surface of the infotainment spot the Optical Model developed by Santbergen et al. [2] was used because allows flexibility in choosing the design and the location separately. However, this model does not include the surrounding, thus in this thesis three new ap- proaches to include the influence of the surrounding environment were proposed and validated through experimental measurements. Moreover, two of these new approaches have the possibility to study the irradiance also in indoor places. After that, its power was simulated knowing the irradiance values, the temperature of the PV module, computed with a thermal-energy balance, and other meteorological data. Lastly, a load profile was developed to simulate the system performance throughout the year. All these information served to find the best pair of number of SHJ cells to interconnect and battery capacity and basically it is a trade-off between system reliability and system size. The study indicated that in Osaka the infotainment spot would need 42 series connected half-cut solar cells that power a 12 푉 system with a storage capacity of 150 퐴ℎ in order to make the system completely autonomous and operable the whole year. The thesis ended with a working prototype of the Solar-Powered Infotainment Spot 2.0.

PowerWindow, a built-in Luminescent Solar Concentrator (LSC), is composed of a window glass with thulium doped coating and a CIGS PV-cell strip attached to one glass edge. LSCs can absorb part of the incoming sunlight and re-direct it to the edges, where PV-cells can produce electricity. Light transport efficiency is then determined by several characteristics of the glass waveguide and thulium as luminescent particles. The main objectives of this work were to find the most relevant characteristics of a LSC affecting its optical efficiency, and to optimize such parameters for a high performance.

In the first part of this work, LightTools, a ray-tracing software, was tested to model a LSC based on red-dyes. Simulation results were compared with experimental measurements using a sample of PMMA doped with red-dye (Lumogen F 305). The optical efficiency of the red-dye model was 4.1%, and LSC sample had 4.5% of efficiency (measured in a previous thesis work [Overbeek, 2015]), meaning 0.4% of marginal error. Optical losses were identified and quantified; from them, specular transmission resulted in 25.5% losses, and 0.1% error compared to measurements. In this respect, LightTools was proven to be a useful tool for modelling LSCs and quantify accurately the optical efficiency. With these results, the ray-tracing method proposed was extended for thulium based LSCs.

Two thulium LSC models were simulated in LightTools: A glass doped with thulium particles (model 1), and pure glass with thulium coating (model 2, PowerWindow design). From the tested variables, the mean free path (MFP) length, related to the luminophore concentration, is one of the main parameters affecting the optical efficiency and transmission losses. Moreover, the MFP ratio showed high relevance in the optical efficiency since it describes how large is the path length for photons in the emission range compared to blue (absorbed) photons.
In both models, it was shown that large MFP ratio (ℛ = 10), high quantum yield (QY= 1), small area (25 cm²), and 5 mm of glass thickness provide optimum results. With these parameters, the optical efficiency of model 1 was 11%, and model 2 about 13%. However, a large scale model 2 of 1 m² had maximum optical efficiency below 1% with light transmission above 80%, and concentration factor of 1. Considering the optical efficiency of this large scale LSC model, and high performance CIGS PV-cells, the energy yield was estimated as only 1.4 kWh per year in the Netherlands. LSC performance might be improved by adding thin film layers in the glass, such as anti-reflection coating and selective filters to allow visible light and reflect NIR photons into the glass.