Generative machine learning has emerged as a powerful tool for design representation and exploration. However, its application is often constrained by the need for large datasets of existing designs and the lack of interpretability about what features drive optimality. To address these challenges, we introduce a systematic framework for constructing training datasets tailored to generative models and demonstrate how these models can be leveraged for interpretable design. The novelty of this work is twofold: (i) we present a systematic framework for generating archetypes with internally homogeneous but mutually heterogeneous inputs that can be used to generate a training dataset, and (ii) we show how integrating generative models with Bayesian optimization can enhance the interpretability of the latent space of admissible designs. These findings are validated by using the framework to design a flow battery manifold, demonstrating that it effectively captures the space of feasible designs, including novel configurations while enabling efficient exploration. This work broadens the applicability of generative machine-learning models in system designs by enhancing quality and reliability.

In this paper, we present a physics-informed neural network (PINN) approach for predicting the performance of an all-vanadium redox flow battery, with its physics constraints enforced by a two-dimensional (2D) mathematical model. The 2D model, which includes 6 governing equations and 24 boundary conditions, provides a detailed representation of the electrochemical reactions, mass transport and hydrodynamics occurring inside the redox flow battery. To solve the 2D model with the PINN approach, a composite neural network is employed to approximate species concentration and potentials; the input and output are normalized according to prior knowledge of the battery system; the governing equations and boundary conditions are first scaled to an order of magnitude around 1, and then further balanced with a self-weighting method. Our numerical results show that the PINN is able to predict cell voltage correctly, but the prediction of potentials shows a constant-like shift. To fix the shift, the PINN is enhanced by further constrains derived from the current collector boundary. Finally, we show that the enhanced PINN can be even further improved if a small number of labeled data is available.

Scientists seek stable, high-energy batteries designed for the electric grid. Bringing new sources of renewable energy like wind and solar power onto the electric grid will require specially designed large batteries that can charge when the sun is shining and give energy at night. One type of battery is especially promising for this purpose: The flow battery. Flow batteries contain two tanks of electrically active chemicals that exchange charge and can have large volumes that hold a lot of energy. For researchers working on flow batteries, their chief concern involves finding target molecules that offer the ability to both store a lot of energy and remain stable for long periods of time.

When searching through theoretical lists of possible new materials for particular applications, such as batteries or other energy-related devices, there are often millions of potential materials that could be considered, and multiple criteria that need to be met and optimized at once. Now, researchers at MIT have found a way to dramatically streamline the discovery process, using a machine learning system. As a demonstration, the team arrived at a set of the eight most promising materials, out of nearly 3 million candidates, for an energy storage system called a flow battery. This culling process would have taken 50 years by conventional analytical methods, they say, but they accomplished it in five weeks. The findings are reported in the journal ACS Central Science, in a paper by MIT professor of chemical engineering Heather Kulik, Jon Paul Janet PhD '19, Sahasrajit Ramesh, and graduate student Chenru Duan.

Lots of experimental robots involve a little bit of cheating. Rather than containing all the necessary electronics and energy sources, they have tethers and wires that provide power and control without weighing the robot down or taking up too much internal space. This is especially true for soft-bodied robots, which typically pump air or fluids to drive their motion. Having to incorporate a power source, pumps, and a reservoir of gas or liquid would significantly increase the weight and complexity of the robot. A team from Cornell University has now demonstrated a clever twist that cuts down on the weight and density of all of this by figuring out how to get one of the materials to perform two functions.

Lots of experimental robots involve a little bit of cheating. Rather than containing all the necessary electronics and energy sources, they have tethers and wires that provide power and control without weighing the robot down or taking up too much internal space. This is especially true for soft-bodied robots, which typically pump air or fluids to drive their motion. Having to incorporate a power source, pumps, and a reservoir of gas or liquid would significantly increase the weight and complexity of the robot. A team from Cornell University has now demonstrated a clever twist that cuts down on the weight and density of all of this by figuring out how to get one of the materials to perform two functions.