Fast approximations of power flow results are beneficial in power system planning and live operation. In planning, millions of power flow calculations are necessary if multiple years, different control strategies or contingency policies are to be considered. In live operation, grid operators must assess if grid states comply with contingency requirements in a short time. In this paper, we compare regression and classification methods to either predict multi-variable results, e.g. bus voltage magnitudes and line loadings, or binary classifications of time steps to identify critical loading situations. We test the methods on three realistic power systems based on time series in 15 min and 5 min resolution of one year. We compare different machine learning models, such as multilayer perceptrons (MLPs), decision trees, k-nearest neighbours, gradient boosting, and evaluate the required training time and prediction times as well as the prediction errors. We additionally determine the amount of training data needed for each method and show results, including the approximation of untrained curtailment of generation. Regarding the compared methods, we identified the MLPs as most suitable for the task. The MLP-based models can predict critical situations with an accuracy of 97-98 % and a very low number of false negative predictions of 0.0-0.64 %.

The US Department of Energy (DOE) has announced it will deploy $21 million (£16m) of funding for fusion energy research projects. The finance will enable scientists to take advantage of new artificial intelligence (AI) and machine learning technologies to speed up progress in fusion energy research. Part of the funding will also be allocated to improve operations at the Office of Science fusion facilities by automating data analysis and enabling algorithms. In a statement, DOE said: "AI and machine learning will help us to accelerate progress in fusion and keep American scientists at the forefront of fusion research."

Oak Ridge National Laboratory is behind the development of a new type of artificial intelligence (AI) software called Peregrine, meant to improve the quality of functional parts being produced via powder bed 3D printers. Peregrine requires no "expensive characterization equipment," yet possesses the ability to evaluate parts during manufacturing. "Capturing that information creates a digital'clone' for each part, providing a trove of data from the raw material to the operational component," said Vincent Paquit, leader of advanced manufacturing data analytics research as part of ORNL's Imaging, Signals and Machine Learning group. "We then use that data to qualify the part and to inform future builds across multiple part geometries and with multiple materials, achieving new levels of automation and manufacturing quality assurance." Oak Ridge National Laboratory researcher Chase Joslin uses Peregrine software to monitor and analyze a component being 3D printed at the Manufacturing Demonstration Facility at ORNL (Image: Luke Scime, ORNL, U.S. Dept. of Energy) The software is based on a convolutional neural network that imitates the human brain, rapidly evaluating images from cameras during printing.

The U.S. Department of Energy and Microsoft on Tuesday announced a partnership to develop artificial-intelligence tools aimed at helping first-responders better react to fast-changing natural events, such as floods and wildfires. From a report: "There are just so many technologies where we can solve some of the toughest problems, in a moment where we're having an explosion of wildfires and floods and some really major natural disasters," said Cheryl Ingstad, director of the Energy Department's Artificial Intelligence and Technology Office. "And we think we can bring AI to bear here and help save lives." The First Five Consortium, a nod to the importance of the first five minutes in responding to a natural disaster, aims to build between 10 and 30 different AI-powered systems. Microsoft will provide technological resources, including its Azure cloud for AI model training and inference.

Today's robot-mounted batteries provide electrical power but at the expense of added mass that in turn requires added power to move and use. But a team of researchers from the University of Michigan have devised a clever solution that will enable tomorrow's batteries to provide power while negating their own weight -- it just needs a bit of Kevlar. Led by Nicholas Kotov, a professor of chemical engineering at U of Michigan, the team has developed a battery system that is strong enough to also serve as a structural support for the rest of the robot. "Robot designs are restricted by the need for batteries that often occupy 20% or more of the available space inside a robot, or account for a similar proportion of the robot's weight," Kotov told the University of Michigan News "No other structural battery reported is comparable, in terms of energy density, to today's state-of-the-art advanced lithium batteries. We improved our prior version of structural zinc batteries on 10 different measures, some of which are 100 times better, to make it happen," he continued.

Combing through historical seismic data with a machine learning model, US researchers have unearthed distinct statistical features that marked the formative stage of slow-slip ruptures in the Earth's crust months before tremor or GPS data detected a slip in the tectonic plates. Given the similarity between slow-slip events and classic earthquakes, they suggest, in a paper in the journal Nature Communications, that these distinct signatures may help geophysicists understand the timing of the faster quakes as well. "The… model found that, close to the end of the slow slip cycle, a snapshot of the data is imprinted with fundamental information regarding the upcoming failure of the system," says lead author Claudia Hulbert, from the Los Alamos National Laboratory. "Our results suggest that slow-slip rupture may well be predictable, and because slow slip events have a lot in common with earthquakes, slow-slip events may provide an easier way to study the fundamental physics of earth rupture."

A breakthrough with biomorphic batteries could allow robots to store up to 72-times more energy through a system similar to biological fat reserves. Researchers at the University of Michigan – funded by the US Department of Defense – developed a new rechargeable zinc battery that integrates into the structure of a robot in order to free up space and reduce weight that conventional lithium-ion batteries create. "Robot designs are restricted by the need for batteries that often occupy 20 per cent or more of the available space inside a robot, or account for a similar proportion of the robot's weight," said Nicholas Kotov, a professor of engineering who led the research. "We don't have a single sac of fat, which would be bulky and require a lot of costly energy transfer. Distributed energy storage, which is the biological way, is the way to go for highly efficient biomorphic devices."

Predictive business process monitoring (PBPM) techniques predict future process behaviour based on historical event log data to improve operational business processes. Concerning the next activity prediction, recent PBPM techniques use state-of-the-art deep neural networks (DNNs) to learn predictive models for producing more accurate predictions in running process instances. Even though organisations measure process performance by key performance indicators (KPIs), the DNN's learning procedure is not directly affected by them. Therefore, the resulting next most likely activity predictions can be less beneficial in practice. Prescriptive business process monitoring (PrBPM) approaches assess predictions regarding their impact on the process performance (typically measured by KPIs) to prevent undesired process activities by raising alarms or recommending actions. However, none of these approaches recommends actual process activities as actions that are optimised according to a given KPI. We present a PrBPM technique that transforms the next most likely activities into the next best actions regarding a given KPI. Thereby, our technique uses business process simulation to ensure the control-flow conformance of the recommended actions. Based on our evaluation with two real-life event logs, we show that our technique's next best actions can outperform next activity predictions regarding the optimisation of a KPI and the distance from the actual process instances.

In identifying infected patients in a population, group testing is an effective method to reduce the number of tests and correct the test errors. In the group testing procedure, tests are performed on pools of specimens collected from patients, where the number of pools is lower than that of patients. The performance of group testing heavily depends on the design of pools and algorithms that are used in inferring the infected patients from the test outcomes. In this paper, an adaptive design method of pools based on the predictive distribution is proposed in the framework of Bayesian inference. The proposed method executed using the belief propagation algorithm results in more accurate identification of the infected patients, as compared to the group testing performed on random pools determined in advance.

Bayesian geoacoustic inversion problems are conventionally solved by Markov chain Monte Carlo methods or its variants, which are computationally expensive. This paper extends the classic Bayesian geoacoustic inversion framework using the mixture density network (MDN), which provides a much more efficient way to solve geoacoustic inversion problems in Bayesian inference framework. Some important geoacoustic statistics of Bayesian geoacoustic inversion are derived from the multidimensional posterior probability density (PPD) using the MDN theory. These statistics make it convenient to train the network directly on the whole parameter space and get the multidimensional PPD of model parameters. The network is trained on a simulated dataset of surface-wave dispersion curves with shear-wave velocities as labels. The results show that the network gives reliable predictions and has good generalization performance on unseen data. Once trained, the network can rapidly (within seconds) give a fully probabilistic solution which is comparable to Monte Carlo methods. It provides an promissing approach for real-time inversion.