Searcharxiv⌕ Search

arXiv subjects

Moritz Gold

Publications and source records attributed to Moritz Gold.

2 recordsLinked to original sources

Metabolic heat and moisture reuse with climate-adaptive control for energy-efficient black soldier fly bioconversion in modular container farms

Black soldier fly larvae (BSFL, Hermetia illucens) convert organic residues into feed ingredients and fertiliser. Climate control is a major contributor to the energy demand of container-based bioconversion. We developed a control-oriented dynamic model linking larval growth, substrate dynamics and container climate to heating, ventilation and air conditioning (HVAC). A two-level Model Predictive Control (MPC) strategy planned climate trajectories and tracked them through HVAC actuation. Four simulation studies examined controller selection, HVAC configuration, larval age pairing and inter-container coupling, using subsets of eight climate scenarios. Across the four scenarios tested for controller selection, dynamic MPC reduced specific energy consumption (SEC) by 31 to 61% and increased predicted dry biomass at harvest by up to 27.5% relative to single fixed setpoints. Under the tested controller settings, active cooling was required in the two hot scenarios, whereas dehumidification provided no additional operational benefit. Pairing larval ages four days apart within one container reduced SEC by up to 33% in the tested non-hot scenarios. In the Cooldry scenario, two age-separated containers connected by an actuated air duct produced 9.0% more biomass and achieved 3.9% lower SEC than two independent containers with mixed ages, despite using 4.6% more total HVAC energy. The simulations suggest that dynamic climate control and metabolic heat and moisture reuse can improve the energy intensity of BSFL bioconversion. The quantitative estimates and hardware recommendations require experimental validation before application to industrial facilities.

eess.SY↗

HistoNet: Predicting size histograms of object instances

We propose to predict histograms of object sizes in crowded scenes directly without any explicit object instance segmentation. What makes this task challenging is the high density of objects (of the same category), which makes instance identification hard. Instead of explicitly segmenting object instances, we show that directly learning histograms of object sizes improves accuracy while using drastically less parameters. This is very useful for application scenarios where explicit, pixel-accurate instance segmentation is not needed, but there lies interest in the overall distribution of instance sizes. Our core applications are in biology, where we estimate the size distribution of soldier fly larvae, and medicine, where we estimate the size distribution of cancer cells as an intermediate step to calculate the tumor cellularity score. Given an image with hundreds of small object instances, we output the total count and the size histogram. We also provide a new data set for this task, the FlyLarvae data set, which consists of 11,000 larvae instances labeled pixel-wise. Our method results in an overall improvement in the count and size distribution prediction as compared to state-of-the-art instance segmentation method Mask R-CNN.

cs.CV↗