English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Many model based scientific and engineering methodologies, such as systemidentification, sensitivity analysis, optimization and control, require a largenumber of model evaluations. In particular, model based real-time control ofurban water infrastructures and online flood alarm systems require fastprediction of the network response at different actuation and/or parametervalues. General purpose urban drainage simulators are too slow for thisapplication. Fast surrogate models, so-called emulators, provide a solution tothis efficiency demand. Emulators are attractive, because they sacrificeunneeded accuracy in favor of speed. However, they have to be fine-tuned topredict the system behavior satisfactorily. Also, some emulators fail toextrapolate the system behavior beyond the training set. Although, there aremany strategies for developing emulators, up until now the selection of theemulation strategy remains subjective. In this paper, we therefore compare theperformance of two families of emulators for open channel flows in the contextof urban drainage simulators. We compare emulators that explicitly useknowledge of the simulator's equations, i.e. mechanistic emulators based onGaussian Processes, with purely data-driven emulators using matrixfactorization. Our results suggest that in many urban applications, naivedata-driven emulation outperforms mechanistic emulation. Nevertheless, wediscuss scenarios in which we think that mechanistic emulation might befavorable for i) extrapolation in time and ii) dealing with sparse and unevenlysampled data. We also provide many references to advances in the field ofMachine Learning that have not yet permeated into the Bayesian environmentalscience community.

Appraisal of data-driven and mechanistic emulators of nonlinear hydrodynamic urban drainage simulators