Deep learning's fitness for purpose:A transformation problem frame's perspective

Combined sewer overflows represent significant risks to human health as untreated water is discharged to the environment.Municipalities,such as the Metropolitan Sewer District of Greater Cincinnati(MSDGC),recently began collecting large amounts of water-related data and considering the adoption of deep learning(DL)solutions like recurrent neural network(RNN)for predicting overflow events.Clearly,assessing the DL's fitness for the purpose requires a systematic understanding of the problem context.In this study,we propose a requirements engineering framework that uses the problem frames to identify and structure the stakeholder concerns,analyses the physical situations in which the highquality data assumptions may not hold,and derives the software testing criteria in the form of metamorphic relations that incorporate both input transformations and output comparisons.Applying our framework to MSDGC's overflow prediction problem enables a principled way to evaluate different RNN solutions in meeting the requirements.

Day-ahead Scheduling of Multi-carrier Energy Systems with Multi-type Energy Storages and Wind Power

The integration of large-scale wind power brings challenges to the operation of integrated energy systems(IES).In this paper,a day-ahead scheduling model for IES with wind power and multi-type energy storage is proposed in a scenario-based stochastic programming framework.The structure of the IES consists of electricity,natural gas,and heating networks which are all included in the model.Operational constraints for IES incorporating multi-type energy storage devices are also considered.The constraints of the electricity network,natural gas network and heating network are formulated,and non-linear constraints are linearized.The calculation method for the correlation of wind speed between wind farms based on historical data is proposed.Uncertainties of correlated wind power were represented by creating multiple representative scenarios with different probabilities,and this was done using the Latin hyper-cube sampling(LHS)method.The stochastic scheduling model is formulated as a mixed integer linear programming(MILP)problem with the objective function of minimizing the total expected operation cost.Numerical results on a modified PJM 5-bus electricity system with a seven-node natural gas system and a six-node heating system validate the proposed model.The results demonstrate that multi-type energy storage devices can help reduce wind power curtailments and improve the operational flexibility of IES.

Virtual screening of inorganic materials synthesis parameters with deep learning

Virtual materials screening approaches have proliferated in the past decade,driven by rapid advances in first-principles computational techniques,and machine-learning algorithms.By comparison,computationally driven materials synthesis screening is still in its infancy,and is mired by the challenges of data sparsity and data scarcity:Synthesis routes exist in a sparse,highdimensional parameter space that is difficult to optimize over directly,and,for some materials of interest,only scarce volumes of literature-reported syntheses are available.In this article,we present a framework for suggesting quantitative synthesis parameters and potential driving factors for synthesis outcomes.We use a variational autoencoder to compress sparse synthesis representations into a lower dimensional space,which is found to improve the performance of machine-learning tasks.To realize this screening framework even in cases where there are few literature data,we devise a novel data augmentation methodology that incorporates literature synthesis data from related materials systems.We apply this variational autoencoder framework to generate potential SrTiO_(3) synthesis parameter sets,propose driving factors for brookite TiO_(2) formation,and identify correlations between alkali-ion intercalation and MnO_(2) polymorph selection.

Hybrid component and configuration model for combined-cycle units in unit commitment problem

This letter proposes a novel hybrid component and configuration model for combined-cycle gas turbines(CCGTs) participating in independent system operator(ISO) markets. The proposed model overcomes the inaccuracy issues in the current configuration-based model while retaining its simple and flexible bidding framework of configuration-based models. The physical limitations—such as minimum online/offline time and ramping rates—are modeled for each component separately, and the cost is calculated with the bidding curves from the configuration modes. This hybrid mode can represent the current dominant bidding model in the unit commitment problem of ISOs while treating the individual components in CCGTs accurately. The commitment status of the individual components is mapped to the unique configuration mode of the CCGTs. The transitions from one configuration mode to another are also modeled. No additional binary variables are added, and numerical case studies demonstrate the effectiveness of this model for CCGT units in the unit commitment problem.

BSIM--making the first international standard MOSFET model

BSIM （Berkeley Short-Channel IGFET） became the first international industry standard model for simulation of MOS integrated circuits in 1997. The cumulative sales of ICs that have been designed with the aid of BSIM and produced for computing, communication, consumer, and industrial applications is estimated to be around 400 billion US dollars. From 0.35μm CMOS to multi-gate FinFET, BSIM serves a wide range of technologies. Many China educated researchers have contributed to its success.