Currently, the accuracy of modeling a photovoltaic (PV) array for fault diagnosis is still unsatisfactory due to the fact that the modeling accuracy is limited by the accuracy of extracted model parameters. In this paper, the modeling of a PV array based on multi-agent deep reinforcement learning (RL) using the residuals of I–V characteristics is proposed. The environment state based on the high dimensional residuals of I–V characteristics and the corresponding cooperative reward is presented for the RL agents. The actions of each agent considering the damping amplitude are designed. Then, the entire framework of modeling a PV array based on multi-agent deep RL is presented. The feasibility and accuracy of the proposed method are verified by the one-year measured data of a PV array. The experimental results show that the higher modeling accuracy of the next time step is obtained by the extracted model parameters using the proposed method, compared with that using the conventional meta-heuristic algorithms and the analytical method. The daily root mean square error (RMSE) is approximately 0.5015 A on the first day, and converges to 0.1448 A on the last day of training. The proposed multi-agent deep RL framework simplifies the design of states and rewards for extracting model parameters.

Thin films of ZnO:Ga2O3 (ZGO) were deposited by radio frequency (RF) sputtering at voltages of 0.5, 0.9 and 1.1 kV. The films were studied with respect to their suitability in flexible piezoelectric nanogenerators. The analysis of the spectroscopic and microscopic results showed that piezoelectric features were revealed for the films grown at all sputtering voltages, but the most favorable morphology in terms of low roughness was achieved at 1.1 kV. The effect of the sputtering voltage on the films crystallinity and lattice strain was studied. It was found that the increasing sputtering voltage promoted the films crystallization. Additionally, the presence of oxygen vacancies in the piezoelectric films was negligible as it is not a major factor affecting their performance. The electrical measurements of the Ag/ZnO:Ga2O3/Ag harvester on a flexible substrate in the low-frequency range showed a piezoelectric voltage of 414 mV, a current of 10.4 µA and an electric power output of 1.4 µW at a mass load of 100 g. These results were achieved by a simple architecture of a single piezoelectric layer with a relatively small size of 3 cm2 and small piezoelectric film thickness (600 nm) containing lead-free material. It was proven that the sputtered ZGO films are suitable for energy harvesting elements and their performance could be tuned by the sputtering voltage. Another possible application of the proposed device, excluding low-frequency vibrational harvesting, could be a pressure sensor or strain gauge, due to the good linearity of the electrical parameter dependences on the strain.

Thin films of BaSrTiO3 (BST) were deposited by radio frequency (RF) sputtering with sputtering voltage varied between 0.5 and 0.7 kV. BaSrTiO3 films are investigated regarding their suitability for use in flexible energy harvesting devices. Analysis of the spectra and micrographs showed the piezoelectric features and the polycrystalline nature for the films grown at higher sputtering voltage. Meanwhile, the presence of an amorphous phase with a lack of characteristic bonds was found out for BaSrTiO3 films grown at the lower sputtering voltage. The electrical measurements of Al/BaSrTiO3/Al harvester on plastic substrate in the low-frequency range revealed superior piezoelectric voltage and 40% exceeding power density for the samples prepared at 0.7 kV as compared to the other thin films deposited at 0.5 kV. These results are achieved by simple architecture of single layer harvester with small size (0.4 cm 2 ) and small film thickness (0.5 µm), containing lead-free material. It was proved that the sputtered BaSrTiO3 films are suitable for thin film piezoelectric nanogenerators and their performance could be easily tuned by the sputtering voltage. Possible applications of the proposed element are low-frequency vibrational harvesting and a strain gauge.

In this joint project the development of a robust low-cost monitoring system of air quality and wall humidity values is presented. Using open source software on a RaspberryPi 3 with some plug-in sensors for temperature, pressure, CO2 and humidity data sets every minute are stored in an universal data base. Data analysis gives good results in recognition of ventilation states inside the apartment, which can be used to give ventilation advices to the apartment users in time of renovation, to avoid bad air quality and mould groth. For protection of personal room user data a NTRU Public Key Cryptosystem is developed and implemented.

In our work we observe the light behavior before and after light-induced degradation of amorphous/microcrystalline (a-Si/μc-Si) tandem silicon-based solar cells. We show that during light-induced degradation the efficiency increases for illumination levels lower than standard test condition illumination. Additionally we examine the temperature coefficient of the open circuit voltage, short circuit current and fill factor. The temperature-coefficient of silicon-based thin film solar cells varies during the light-induced degradation.

The use of current injection and light exposure is shown to distinguish the impact of degradation in the contact, and intrinsic regions of a-Si:H solar cells, respectively. The drop in the maximum power conversion capability of the cell after light exposure is a consequence of an increase of dangling bonds in the intrinsic layer of the cell due to the Staebler-Wronksi effect. This has a detrimental effect on short circuit current, open circuit voltage, and fill factor. On the other hand, injected current increases the open circuit voltage and greatly reduces the fill factor without affecting the short circuit current, which is attributed to an increase of defects in the p-layer. A clear distinction from both degradation mechanisms is observed from evolution of the ideality factor m of the main junction, and the ideality factor n of the ZnO/a-Si:H(p) interface. A back-to-back diode model of a solar cell that considers the effect of non-ideal contacts, where a high value of n represents an increased tunneling transport at the contact interface, can effectively address the separation of damage in the two regions.

The impact of boron doping on the p-layer of thin film silicon solar cells is assessed by measuring the effective Schottky barrier height of ZnO/a-Si:H and ZnO/μc-Si:H heterojunctions. A deviation from ideal diode characteristics is revealed by an increase of ideality factor with doping concentration. Higher current densities and lower effective Schottky barriers are evaluated for higher doping levels, resulting in increasingly Ohmic behaviour. This is attributed to an enhancement of tunneling through a thinner depletion region, as supported by computer simulations. Extracted barriers are in the range of 0.7–1 eV for the heterojunctions with rectifying behaviour.

Doped zinc oxide films are of high interest in thin film solar cell technology for application as transparent conducting oxide. Rapid and detailed characterisation of ZnO thin film properties is required for quality control and optimisation of the deposited films. In the present work, a new model of dielectric functions based on the effective medium approximation (EMA) is developed and is applied for characterisation of polycrystalline boron doped zinc oxide (ZnO:B) films, deposited by low pressure chemical vapour deposition (LPCVD) technique onto glass substrates. The model takes into account that polycrystalline ZnO is considered to consist of crystal grains surrounded by depletion layers. Using this model and Fourier Transform Infrared Spectroscopy (FTIR) performed in reflection configuration over a wide mid-infrared spectral region (from 2 μm up to 25 μm), the properties of depletion layer and the bulk of the grains in ZnO can be rapidly characterised in detail, and the volume fraction of the depletion layer can be extracted. The results are in good agreement with previously presented theories of electron transport in polycrystalline materials. Using electrical measurements like conductivity and Hall techniques in addition to the optically determined parameters, predominant electron scattering mechanisms in polycrystalline films for different doping levels are identified. The measurements show the impact of the doping level on depletion layer of the crystallites. It is shown, furthermore, that under a water vapour rich environment the volume fraction of the depletion layer may increase up to 5 times and more, while the mobility of the charge carriers in the depletion layer drops drastically from about 31 cm2V−1s−1 to about 8 cm2V−1s−1. This indicates that water vapour exposure causes an increase of the potential barrier in the grain boundary depletion layer, limiting the electron transport across the grain boundaries to a classical thermionic emission mechanism.

The influence of the surface morphology of the front contact layer on the properties of amorphous silicon based solar cells is presented. Based on a textured LPCVD substrate, which consists of pyramidal grains emerging out of the surface plane, the surface topology is modified by plasma etching. By varying the gas composition in the plasma at a fixed total flow, we were able to create surface morphologies whose features show deviations from the pyramidal shape, without significantly changing the surface roughness. For the plasma etched samples JSC decreases in comparison to the as-deposited ZnO. The decrease in current is not exclusively caused by a degraded red response but also due to a lowered quantum efficiency in the whole visible spectral range. This is attributed to a degraded light incoupling at the ZnO/ a-Si interface due to the altered pyramidal shape. It could be shown experimentally that the shape of the surface morphology is important for light incoupling as well as light trapping.

Transparent conductive oxides are often used as transparent front electrodes for thin film solar cells. Besides their conductivity and transparency in the absorption range of the solar cell also the light scattering ability is important for light management. The TCOs are textured, e.g. as grown LPCVD ZnO:B or texture-etched ZnO:Al, and depending on their morphology they are differently adequate for their application as front electrode in solar cells. An evaluation method based on angular resolved light scattering (ARS) is presented in this paper. Measurement results and evaluation of different types of textured doped TCOs like reactively MF sputtered ZnO:Al (Zn:Al target), RF sputtered ZnO:Al (ZnO:Al2O3 target) and LPCVD ZnO:B are shown. A correlation between ARS and short-circuit density of a-Si:H/μc-Si:H p-i-n solar cells was found for LPCVD ZnO:B and reactively or RF sputtered and etched ZnO:Al until a saturation current was reached.