Search

We present a temperature-dependent Suns-V OC technique and use it to examine the temperature- and illumination-dependent performance of various silicon solar cells including p-PERC, n-PERT, n-SHJ, and n-TOPCon. We investigate the temperature coefficients (TC) of their electrical parameters and the illumination-dependent TC of their open-circuit voltage (V OC ). At one-sun, the n-SHJ cell is seen to be the least temperature-sensitive, driven by its relatively superior TC Voc at intensities above 0.03 suns. As illumination decreases, the V OC of each cell becomes significantly more sensitive to temperature. To our knowledge, this is the first published study to utilize Suns-V OC (T) to provide a detailed comparison of temperature- and illumination-dependent performance of those solar cell technologies.

High contact resistance keeps black phosphorus (BP) from fully wielding its excellent material property. Using first-principles calculations, we analyze the interfacial binding behavior and the impact of binding on the other layers of a trilayer BP. We found that the interfacial charge density and charge transfer of Scandium (Sc)-contacted trilayer BP are 2.67 and 3.29 times greater than Au-contacted trilayer BP, respectively. Moreover, the interfacial tunneling barrier height and width of Sc-contacted trilayer BP are 0 eV and 1.851 Å, which are significantly smaller than that of 5.1 eV and 2.447 Å observed in Au-contacted trilayer BP. All these facts suggest a strong bonding and efficient carrier transmission between Sc contact and trilayer BP substrate. Therefore, we conclude that the Sc electrode can lead to a superior performance that is consistent with the experiment.

The importance of electrical trees and interfacial electrical track growth as aging processes in high voltage polymeric insulation systems is well established. It is argued that the majority of testing using needle-plane electrode systems with small insulation gaps is not representative of insulation in high voltage service conditions. In particular, the relationship between partial discharge magnitude, tree length and sample scale is considered. Track growth has been considered in laboratory samples with a 90 mm gap between electrodes. This arrangement allows the channels to grow independent of their proximity to the electrodes. The applied voltage and sample thickness, as well as the electrical field, are critical parameters to be considered. The growth of channels here was voltage limited. Smaller and lower voltage tests cannot replicate this behavior because in small samples tree channels are always in close proximity to, and dominated by, the electrode geometry.

In this study, we utilized low cost and easily available bubble wrap, homemade indicator solution (cabbage juice), and custom built mobile app based on color detection to find pH values of common liquid samples. We studied two common color definition models i.e. the Red, Green, Blue (RGB) and Hue, Saturation, and Light (HSL) to extract pH values of eight different liquids upon color change after reaction with cabbage juice. Our results show that the HSL based color definition model performs better than the RGB model for both calibration liquids (vinegar with pH=3 and bicarbonate soda pH=9) and eight target liquids with the pH range spanning from 2 to 13.

In this paper, we are going to explain the design process for our GVGAI game-learning agent, which is going to be submitted to the GVGAI competition’s learning track 2020. The agent relies on a local forward modeling approach, which uses predictions of future game-states to allow the application of simulation-based search algorithms. We first explain our process in identifying repeating tiles throughout a pixel-based state observation. Using the tile information, a local forward model is trained to predict the future state of each tile based on its current state and its surrounding tiles. We accompany this approach with a simple reward model, which determines the expected reward of a predicted state transition. The proposed approach has been tested using multiple games of the GVGAI framework. Results show that the approach seems to be especially feasible for learning how to play deterministic games. Except for one non-deterministic game, the agent performance is very similar to agents using the true forward model. Nevertheless, the prediction accuracy needs to be further improved to facilitate a better game-playing performance.

For general video game playing agents, the biggest challenge is adapting to the wide variety of situations they encounter and responding appropriately. Some success was recently achieved by modifying search-control parameters in agents on-line, during one play-through of a game. We propose adapting such methods for Rolling Horizon Evolutionary Algorithms, which have shown high performance in many different environments, and test the effect of on-line adaptation on the agent's win rate. On-line tuned agents are able to achieve results comparable to the state of the art, including first win rates in hard problems, while employing a more general and highly adaptive approach. We additionally include further insight into the algorithm itself, given by statistics gathered during the tuning process and highlight key parameter choices.

Access to a fast and easily copied forward model of a game is essential for model-based reinforcement learning and for algorithms such as Monte Carlo tree search, and is also beneficial as a source of unlimited experience data for model-free algorithms. Learning forward models is an interesting and important challenge in order to address problems where a model is not available. Building upon previous work on the Neural GPU, this paper introduces the Neural Game Engine, as a way to learn models directly from pixels. The learned models are able to generalize to different size game levels to the ones they were trained on without loss of accuracy. Results on 10 deterministic General Video Game AI games demonstrate competitive performance, with many of the games’ models being learned perfectly both in terms of pixel predictions and reward predictions. The pre-trained models are available through the OpenAI Gym interface and are available publicly for future research here: https://github.com/Bam4d/Neural-Game-Engine

This paper introduces a computer-game design tool which enables game designers to explore and develop game mechanics for arbitrary game systems. The tool is implemented as a plugin for the Godot game engine. It allows the designer to view an abstraction of a game’s states while in active development and to quickly view and explore which states are navigable from which other states. This information is used to rapidly explore, validate and improve the design of the game. The tool is most practical for game systems which are computer-explorable within roughly 2000 states. The tool is demonstrated by presenting how it was used to create a small, yet complete, commercial game.

In the realm of games research, Artificial General Intelligence algorithms often use score as main reward signal for learning or playing actions. However this has shown its severe limitations when the point rewards are very rare or absent until the end of the game. This paper proposes a new approach based on event logging: the game state triggers an event every time one of its features changes. These events are processed by an Event-value Function (EF) that assigns a value to a single action or a sequenceThe experiments have shown that such approach can mitigate the problem of scarce point rewards and improve the AI perfor-mance. Furthermore this represents a step forward in controlling the strategy adopted by the artificial agent, by describing a much richer and controllable behavioural space through the EF. Tuned EF are able to neatly synthesise the relevance of the events in the game. Agents using an EF show more robust when playing games with several opponents.

Electrical tree initiation in LDPE under positive DC stresses with controlled AC ripple is investigated. Needle-plane geometries were used and partial discharge (PD) signals recorded. Tree initiation occurred around +45 kV, and the initial growth within the first second was long compared to the more gradual processes seen with AC excitation. In each case, when the first tree image was captured, they were of the branch type and at least one order of magnitude longer than those seen under pure AC stresses. 5 kV peak AC superimposed onto +45 kV led to rapid sample failure. Unlike in pure AC cases, PD detection did not correlate well with the progress of electrical treeing (initiation, growth or breakdown) under pure DC stresses. Phase resolved PD analysis did however provide useful information in the presence of an AC ripple.

A study of chemical aging prior to electrical tree initiation has been undertaken using a high-resolution chemical analysis technique, AFM-IR (atomic force microscopy-infrared spectroscopy). This technique has provided chemical characterisation with a spatial resolution of around 50 nm. It is seen that the use of tests on this scale provide information and insight which was not previously observable using lower resolution techniques. Previously unobserved degradation products are observed in epoxy resin prior to tree initiation with ethers (C-O) and methyl groups (CH 3 ) appearing to form alongside carbonyls (C=O), each occurring in different locations around the needle tip. Such spatial resolution of aging chemistry will allow improved models of tree initiation and growth to be developed.

We present a “post-mortem” procedure for analyzing degradation of fielded photovoltaic modules. We use this procedure to investigate degradation mechanisms in silicon heterojunction (SHJ) modules. We study cell fragments cored from two modules, one operated in the field for a decade and one stored in the dark as a control. We find that both the front surface passivation and bulk have likely been degraded. Through comparisons of spectral photoluminescence emissions, we conclude that although an increase in density of the pre-existing types of radiative defects is possible, it seems that any new defects types are purely non-radiative.

In this work a novel device structure aiming to suppress the degradations under bending in flexible devices is demonstrated, named as a wing-shape TFT. Investigations of the electrical characteristics and reliabilities are supported by simulations, theoretical calculations, and structural discussions. First, through both electrical and mechanical stress simulation, the regions of degradation are confirmed to occur at the corner of the gate insulator. After long term mechanical stress, behaviors including an abnormal hump, serious threshold voltage (Vth) shifts, on-current increase, and off-current decrease are observed. By extending the active layer, also serving as a stress relief layer, removes the degradation regions away from the main channel. Therefore, the reliability of the devices is dramatically improved. The geometries of the wings in wing-shape TFTs are also in consideration. From results, with different wing lengths, a similar improvement of the reliability is shown, but the on current is enhanced when the wing length is extended. These phenomena are realized as a result of the difference in the effective width. The proposed wing-shape LTPS TFT has its potential and benefits for future flexible displays.

In this paper, a 100-nm gate length InAs high electron mobility transistor (HEMT) with non-alloyed Ti/Pt/Au ohmic contacts and mesa sidewall channel etch was investigated for high-speed and low-power logic applications. The device exhibited a low subthreshold swing (SS) of 63.3 mV/decade, a drain induced barrier lowering value of 23.3 mV/V, an ION/IOFF ratio of 1.34 × 104, a Gm,max/SS ratio of 27.6, a current gain cut-off frequency of 439 GHz with a gate delay time of 0.36 ps, and an off-state gate leakage current of less than 1.6 × 10-5 A/mm at VDS = 0.5 V. These results demonstrated that the presence of non-annealed ohmic contacts with mesa sidewall etch process allowed the fabrication of InAs HEMTs with excellent electrical characteristics for high-speed and low-power logic applications.

In this paper, we discover an operation method that can effectively decrease the forming voltage in resistance random access memory (RRAM). Forming voltage can be reduced by either increasing the rising time of the forming-waveform or by increasing the temperature in the forming process. However, the resulting electronic RRAM characteristics after each of these methods differ. While increasing the rising time causes greater damage to the switching layer due to longer accumulation of charge, increasing temperature in the forming process does not. The high temperature-formed RRAM excels in retention and endurance tests, proving an effective means to decrease forming voltage.

Game design is a time-consuming endeavor and, in contrast to other work in game development, has no tools aiding the designer in speeding up the process. In this paper, the initial work towards developing such a tool is presented. It focuses on the extraction of learning curves from single-player, deterministic, turn-based (puzzle) games. Learning curves in such games can be directly associated with their gameplay quality. The extraction is done by training a number of varying size, feed-forward neural networks on the game transition model. The main contribution of the work is the introduction of learning curve based puzzle game design concepts and their practical examples along with a quantitative method for learning curve estimation.

The Games Fusion Project was an experimental learning experience that mirrored commercial realities and processes in order to wrap the core game-making skills in a fusion of wider industry-relevant competencies. This provided both contextual knowledge and exposed participants to failure and reflection, foregrounding the importance of iterative learning loops, foundational in creative digital industry design processes as well as (experiential) learning paradigms. This project has explored how the relevance of game design for education lies not only in the creative, communication and project management competencies exposed in the game design process but also how this can be abstracted to inform project curriculum development more widely.

There are few digital tools to help designers create game mechanics. A general language to express game mechanics is necessary for rapid game design iteration. The first iteration of a mechanics-focused language, together with its interfacing tool, are introduced in this paper. The language is restricted to two-dimensional, turn-based, tile-based, deterministic, complete-information games. The tool is compared to the existing alternatives for game mechanics prototyping and shown to be capable of succinctly implementing a range of well-known game mechanics.

Accurate modelling of electrical tree growth is dependent upon a physical and chemical understanding of the insulation and degradation processes on the scale of the tree branches. While imaging techniques for the physical elements of tree growth have improved in recent years, those for the chemical regime are lagging behind. In this paper AFM-IR (Atomic Force Microscopy - Infrared Spectroscopy) is applied to a non-conducting tree channel grown in epoxy resin. This provides chemical analysis with a spatial resolution of 50 nm. The distinct chemistries from within buried channels, exposed channels and the epoxy bulk are revealed for the first time.

Sustainable management of urban wastewater networks is a global problem where sewer pipes are plagued by numerous harsh conditions causing concrete attack. The end result is premature pipe failure that comes with high financial, public health and environmental costs. Currently, water utilities do not have an in-situ, online monitoring capability for their concrete wastewater assets as conventional electrical sensors fail within a few weeks of installation due to the chemically harsh gaseous environment. Here we outline the development of robust fibre Bragg grating (FBG) sensors to optically measure concrete corrosion rates in wastewater environments over long time periods.