Private sensors and crowdsourced rainfall data: accuracy and effectiveness for monitoring severe events and mapping pluvial flooding in urban areas

The increasing number of extreme rainfall events, combined with the high population density and the imperviousness of the land surface, makes urban areas particularly vulnerable to pluvial flooding. In order to design and manage cities to be able to deal with this issue, the reconstruction of weather phenomena is essential. Among the most interesting data sources which show great potential are the observational networks of private sensors managed by citizens (crowdsourcing). The number of these personal weather stations is consistently increasing, and the spatial distribution roughly follows population density. Precisely for this reason, they perfectly suit this detailed study on the modelling of pluvial flood in urban environments. The uncertainty associated with these measurements of precipitation is still a matter of research. In order to characterise the accuracy and precision of the crowdsourced data, we carried out exploratory data analyses. A comparison between Netatmo hourly precipitation amounts and observations of the same quantity from weather stations managed by national weather services is presented. The crowdsourced stations have very good skills in rain detection but tend to underestimate the reference value. In detail, the accuracy and precision of crowd- sourced data change as precipitation increases, improving the spread going to the extreme values. Then, the ability of this kind of observation to improve the prediction of pluvial flooding is tested. To this aim, the simplified raster-based inundation model incorporated in the Saferplaces web platform is used for simulating pluvial flooding. Different precipitation fields have been produced and tested as input in the model. Two different case studies are analysed over the most densely populated Norwegian city: Oslo. The crowdsourced weather station observations, bias-corrected (i.e. increased by 25%), showed very good skills in detecting flooded areas.

Deep Learning Architectures for time of arrival detection in Acoustic Emissions Monitoring

Acoustic Emission (AE) monitoring can be used to detect the presence of damage as well as determine its location in Structural Health Monitoring (SHM) applications. Information on the time difference of the signal generated by the damage event arriving at different sensors is essential in performing localization. This makes the time of arrival (ToA) an important piece of information to retrieve from the AE signal. Generally, this is determined using statistical methods such as the Akaike Information Criterion (AIC) which is particularly prone to errors in the presence of noise. And given that the structures of interest are surrounded with harsh environments, a way to accurately estimate the arrival time in such noisy scenarios is of particular interest. In this work, two new methods are presented to estimate the arrival times of AE signals which are based on Machine Learning. Inspired by great results in the field, two models are presented which are Deep Learning models - a subset of machine learning. They are based on Convolutional Neural Network (CNN) and Capsule Neural Network (CapsNet). The primary advantage of such models is that they do not require the user to pre-define selected features but only require raw data to be given and the models establish non-linear relationships between the inputs and outputs. The performance of the models is evaluated using AE signals generated by a custom ray-tracing algorithm by propagating them on an aluminium plate and compared to AIC. It was found that the relative error in estimation on the test set was < 5% for the models compared to around 45% of AIC. The testing process was further continued by preparing an experimental setup and acquiring real AE signals to test on. Similar performances were observed where the two models not only outperform AIC by more than a magnitude in their average errors but also they were shown to be a lot more robust as compared to AIC which fails in the presence of noise.

Development of wearable electrochemical sensors for the determination of biological analytes

Wearable biosensors are attracting interest due to their potential to provide continuous, real-time physiological information via dynamic, non-invasive measurements of biochemical markers in biofluids, such as interstitial fluid (ISF). One notable example of their applications is for glycemic monitoring in diabetic patients, which is typically carried out either by direct measurement of blood glucose via finger pricking or by wearable sensors that can continuously monitor glucose in ISF by sampling it from below the skin with a microneedle. In this context, the development of a new and minimally invasive multisensing tattoo-based platform for the monitoring of glucose and other analytes in ISF extracted through reverse iontophoresis in proposed by the GLUCOMFORT project. This elaborate describes the in-vitro development of flexible electrochemical sensors based on inkjet-printed PEDOT:PSS and metal inks that are capable of determining glucose and chloride at biologically relevant concentrations, making them good candidates for application in the GLUCOMFORT platform. In order to make PEDOT:PSS sensitive to glucose at micromolar concentrations, a biocompatible functionalization based on immobilized glucose oxidase and electrodeposited platinum was developed. This functionalization was successfully applied to bulk and flexible amperometric devices, the design of which was also optimized. Using the same strategy, flexible organic electrochemical transistors (OECTs) for glucose sensing were also made and successfully tested. For the sensing of chloride ions, an organic charge-modulated field-effect transistor (OCMFET) featuring a silver/silver chloride modified floating gate electrode was developed and tested.

Monitoring of macro- and microlitter in the Venice lagoon and coastal area

Marine litter and plastics are a significant and growing marine contaminant that has become a global problem. Macrolitter is subject to fragmentation and degradation due to physical, chemical and biological processes, leading to the formation of micro-litter, the so-called microplastics. The purpose of this research is to assess marine litter pollution by using remote sensing tools to identify areas of macrolitter accumulation and to evaluate the concentrations of microplastics in different environmental matrices: water, sediment and biota (i.e. mussels and fish) and to contribute to the European project MAELSTROM (Smart technology for MArinE Litter SusTainable RemOval and Management). The aim is to monitor the presence of macro- and microlitter at two sites of the Venice coastal area: an abandoned mussel farm at sea and a lagoon site near the artificial Island of Sacca Fisola; The results showed that both study areas are characterised by high amounts of marine litter, but the type of observed litter is different. In fact, in the mussel farm area, most of the litter is linked to aquaculture activities (ropes, nets, mooring blocks and floating buoys). In the Venice lagoon site, the litter comes more from urban activities and from the city of Venice (car tyres, crates, wrecks, etc.). Microplastics is present in both sites and in all the analysed matrices. Generally, higher microplastics concentrations were found at Sacca Fisola (i.e., in surface waters, mussels and fish). Moreover, some differences were also observed in shapes and colours comparing the two sites. At Sacca Fisola, white irregular fragments predominate in water samples, blue filaments in sediment and mussels, and transparent irregular fragments in fish. At the Mussel Farm, blue filaments predominate in water, sediment and mussels, while flat black fragments predominate in fish. These differences are related to the different types of macrolitter that characterised the two areas.