Business Intelligence Industry 4.0
Ricerca e sviluppo di modelli di descriptive, predictive e prescriptive analytics per la business intelligence finalizzata all'efficienza energetica e alla manutenzione predittiva di macchinari industriali in ottica Industria 4.0
The project aimed to develop descriptive, predictive and prescriptive analytics models for Business Intelligence for energy efficiency and predictive maintenance of industrial machinery from an Industry 4.0 perspective. Specifically, Business Intelligence Industry 4.0 aims to:
minimize the use of ad hoc systems, ensuring compatibility with existing infrastructures;
create interfaces within the reach of different types of users for the realization of objectives at different levels;
build a database of the plant history, so as to predict future trends; improve the efficiency of the workforce and avoid wasting time; minimize the waste of hardware and computational resources, guaranteeing maximum security;
customize the service, providing decision support.
Idea75, with Business Intelligence Industry 4.0 (BI_I4.0) has developed descriptive and predictive analytics algorithms for energy efficiency and predictive maintenance.
The developed solution uses machine learning processes, learning automatically and over time thanks to adaptive algorithms that have the ability to use a large amount of data, coming from the field sensors (for example energy consumption data: electricity, water, gas), they learn from them and are able to elaborate, in an intuitive and inductive way, forecasts for the future.
The function implemented for the energy efficiency improvement, combining computational analysis, statistics, mathematics manages to predict what the energy consumption in a given plant will be for a certain time during a given production activity. The forecast data thus elaborated are fundamental information with which it is possible, through the software, to understand where to intervene to optimize consumption and make energy efficiency, analyzing the Key Performance Indicators (KPIs).
In the context of prescriptive analytics, Idea75 has developed a DSS (Decision Support System): it is a decision support system that adapts dynamically to the available data, selecting the most appropriate forecasting model based on criteria specified by the user.
Step of the energy efficiency and predictive maintenance algorithm (descriptive and predictive anlytics)
Simulink model for the efficiency computation
Predictive analytics: analysis of performance indicators
DSS: application of prescriptive analytics models
For the validation of BI_I4.0, two high energy consumption machines were used, used during the milling production process; each of these has production constraints (batch size, processing times, storage times, etc.).
Production planning based on consumption and breakdown forecasts was carried out by analyzing and monitoring different KPI categories:
programming / production;
Furthermore, the out of samples indicators were evaluated for six different forecasting models and for the different optimization methods used.
The implemented decision support system dynamically adapts to the available data, selecting the most appropriate forecasting model based on criteria specified by the user (accuracy criterion or variability criterion).
With its technological and innovative solutions, BI_I4.0 has made it possible to obtain the reduction:
production waste of 20%;
energy costs, estimated at 15%;
of the final customer's complaint rate of 5%.
Optimization methods used: comparison
DSS: criteria selection for the most appropriate forecasting model
BI_I4.0 - Predictive Maintenance: analysis of a system not subject to breakdowns
BI_I4.0 - Identification of a system non sujbect to breakdowns
BI_I4.0 - Predictive Maintenance: deteriorating system analysis
BI_I4.0 - Deteriorating system identification