A Data Driven Model for the Optimization of Energy Consumption of an Industrial Production Boiler in a Fiber Plant

(GENFI – Gerenciamento de variáveis de processo e ativos para eficiência energética)

Initiated in 2019 in partnership with an industrial fiber chemical plant, a data-based approach is proposed to optimize the process of an industrial boiler in a fiber plant. Initially, the production variables with the most impact on the plant’s energy consumption were identified. A machine learning model was developed next using the historical data collected for these parameters at the factory. It formally relates these parameters to the energy consumption of the analyzed system. An approach based on genetic algorithms was then used to search for the ideal values of the model’s input parameters that optimize energy consumption for given production demands. As a result, we obtain a tool capable of guiding the daily operation of the modeled system, helping to reduce energy consumption per unit of production.

Intelligent System Development Platform For Intelligent And Sustainable Society (IMPRESS)

The year 2013 saw the approval of the EU-Brazil IMPRESS project. The consortium consisted of academic and industrial partners. This builds a platform (middleware) for the rapid and robust development of applications in the Internet of Things (IoT). During Impress an energy management system was developed as a proof of concept. The project also developed energy control hardware, power meters and IoT devices as part of an energy efficiency proof of concept deployed within the University campus.

A Monitoring System for Energy Management

(SIMONE – Sistema de Monitoramento e Gerenciamento de Energia)

Initiated in 2019 in partnership with an industrial polymer manufacturing plant, a real-time monitoring framework is developed. It includes energy consumption prediction using Machine Learning techniques, the setup of high level consumption indicators and a production planning model based on an analytical model that seeks to schedule production in a way that minimizes energy consumption.

Virtual Base Station Robotic Front-plane Operation (vRBOT)

Initiated in 2020 in cooperation with a partner from industry, its main goal is to enhance our previous prototype solution aimed at remote maintenance of a Radio Base Station (RBS) and integrate it into the context of Virtual Radio Access Networks (V-RAN).  During the project, a modular gripper design with embedded sensors for the UR5 robotic arm was achieved. Camera calibration was enhanced in this phase.

Human Robot Safe Collaboration HOSA

Initiated in 2019 together with a partner, HOSA develops and evaluates a platform that guarantees safety to operators in the presence of robots. It captures the knowledge of the environment sufficiently to infer risk situations. The proposed architecture supports an advanced augmented reality interface, the visualization of site-specific real time scene constraints, safety Verification: managers can simulate different scenarios in terms of human risk. Other modules include risk knowledge capture and representation and intent recognition.

Base Station Robotic Front-plane Operation (RBOT)

Developed in 2018, this industrial collaboration built a testbed for radio base station remote maintenance through the manipulation of a robotic AR5 arm. Robot operation could be (1) tele-operated mode, with the aid of augmented reality, and (2) semi-automatic mode, where some inspection and maintenance activities will be carried out by the robot itself.

SFC orchestration for Distributed Clouds Availability (SFC-DCA)

To allocate SFCs in a distributed scenario (e.g., a geographically distributed data center) and to achieve a high level of performance and availability is a complex task. Projects DCAV and DCAV II, focused on proposing a set of stochastic availability models to represent the complex cloud data center infrastructure. They considered three basic subsystems: power, cooling, and IT; and we have modeled them separately as independent models. SFC-DCA, started in 2020, develops an SFC orchestration solution focused on the contexts of distributed clouds and micro data centers. It considers the issues of placement and the dynamic composition of VNFs.

Data Center Availability II — DCAV II

This Project started in 2019 to expand models and algorithms for representing subsystems of a cloud data center (based on the results of the previous project, called DCAV). The focus of this project was on the automatic acquisition of information about the configuration of the IT (Information Technology) subsystem of a data center, as well as on the proposition of models for analyzing the availability of vPODs and algorithms for allocation of vPODs. It evaluated DC availability considering availability requirements and restrictions on hardware resources (processing, network and storage constraints).

AVTS-Marcação Avançada Precisa de Tempo de Pacotes em Ambientes Virtualizados

In 2017, this project was supported by the FAPESP research foundation. It focuses on carrying out experiments using virtualization with the Kernel Virtual Machine (KVM) for sending and receiving packets. The AVTS project aims to provide a flexible testbed environment for evaluating the effect of virtualization on the time-stamping process under different operating conditions, with different CPU loads, different packet sizes, packet frequencies, packet arrival times, etc.