Being prepared for the changes in the industry of the future means that industrial companies must anticipate changes and adapt to new tools and approaches, while respecting environmental constraints.
The engineering school projects offer a practical and innovative learning platform for partner companies, start-ups or large groups to accompany this world in transition.
Automation, recycling, IOT, BIM. An outlook into three industrial innovation projects developed by students from the class of 2022 as part of their final year of the Master’s in Engineering programme at ESILV.
Processing of cigarette butts into a renewable resource
Smoky, a start-up launched by two former ESILV students, works around the recovery, decontamination, and treatment of used cigarette butts. Its two founders, Alexandre PERRET and Stefan PETROVIC asked us to study and design some applications using their depolluted cigarette butts as raw material.
Following a review of state of the art, we focused on developing only two of the applications we could find, with a spectrum of industrial sectors involved ranging from additive manufacturing to the textile industry to battery anodes.
Our initial goal was to provide our partners with two final prototypes of materials derived from cigarette butts and their respective production processes. The first was a cotton-like textile fiber, and the second was a standardized plastic sheet.
Due to the lack of time and equipment required to deliver these two prototypes, we had to concentrate only on the development of the plates that seemed to us to be the most versatile application of the two retained.
Our first experiments allowed us to produce three plates having different densities and thicknesses. Following these experiments, we agreed with our partner to adapt the deliverable and provide them with a comparative study of the mechanical properties, the overall production cost, and the carbon footprint generated between a cellulose acetate sheet of recycled origin and another in the first life cycle.
After choosing an experimental protocol according to the ISO 3167 standard, we designed and produced a thermoforming mold intended for the production of flat specimens for our mechanical tests. Despite our previous success in plate processing, we couldn’t obtain samples of sufficient quality to give us usable results.
Our partners are now using our progress to support their upcoming fundraising.
Progress of an automated construction site using 5D BIM and IoT
The CAD.42 project aims to measure the progress of a construction site in real-time using IoT and a 5D BIM model.
By proposing a Cloud solution, the group’s idea is to be able to update the schedule of a construction project automatically, thus allowing to anticipate delays and problems encountered on site.
In this context, our team worked on the recognition of two contextual scenes, the loading of a concrete skip and the unloading of a truck, using data recovered in the field (photos, geolocation of the tracker, weight, date, etc.).
The information is based on the predictions of site objects provided by Microsoft Custom Vision. We performed several multi-criteria approaches using Python and compared their performance using confusion matrices and metrics.
In addition, a Python script was written to simulate the arrival of field data on the Cloud platform.
We then correlated it with the BIM model and the project schedule to identify the corresponding task and its relative construction elements. This linkage was performed on Node-RED in collaboration with the platform developer.
Finally, the implementation of the contextual recognition was proposed and realized by writing a JavaScript code directly integrated in the Cloud platform’s workflows.
Automatic mapping of an industrial environment network
During a security audit, mapping objects connected to the Internet (TCP/IP) and to loT networks (zigbee, LoRa, Bluetooth …) is essential to identify potential vulnerabilities.
Working in partnership with Red Alert Labs, a company specializing in IoT bcybersecurity auditing, we have developed a tool that allows us to map a network, recovering as much information as possible, such as IP address, MAC address, open ports or OS.
To do this, we use sniffing to detect the devices communicating between them on the network. In a second phase, we perform ARP queries to find the devices that are not communicating. However, the firewall can cause problems during this step, so we have separated it from sniffing.
Finally, we are interested in the IoT communication protocol Zigbee to identify potential gateways on the network, and the objects connected to it.
All this allows us to develop the most exhaustive mapping of the network to facilitate the auditor’s work.
Thanks to this, he can identify any anomaly or weakness in the network, using the results given in a json format and the map generated in png.
This post was last modified on %s = human-readable time difference 1:06 pm