verticale

IoT & Artificial intelligence lanciano una nuova generazione di CMMS per l’Industry 4.0?

(in lingua inglese)
Digital revolution
Artificial Intelligence & Cloud Computing
Digital Equipment Platform
Project

Scarica il PDF Scarica il PDF
Aggiungi ai preferiti Aggiungi ai preferiti


Atti di convegni o presentazioni contenenti case history
MCM ottobre 2018 La manutenzione dell’era digitale: strumenti, automazione, metodologie e persone

Pubblicato
da Benedetta Rampini
MCM 2018Segui aziendaSegui




Settori: 

Parole chiave: 


Estratto del testo
Veronafiere 17-18 ottobre 2018 Vi aspettiamo a mcT Petrolchimico Milano, 29 novembre 2018 Cogenerazione Termotecnica Industriale Pompe di Calore 27 ottobre Cogenerazione Termotecnica Industriale Pompe di Calore Alimentare Alimentare Petrolchimico Alimentare 28 ottobre Alimentare Petrolchimico Alimentare Alimentare Petrolchimico Visione e Tracciabilit 28 ottobre Luce Energia Domotica LED Luce Energia Domotica LED Petrolchimico Alimentare Petrolchimico Petr 17 Ottobre 2018 IOT & ARTIFICIAL INTELLIGENCE LANCIANO UNA NUOVA GENERAZIONE DI CMMS PER L'INDUSTRY 4.0' Laurent TRUSCELLO ' Product Manager 2 CARL SOFTWARE 1. 3 CARL SOFTWARE - INNOVATION STRATEGY 2,2 M' invested in R&D in 2017 Axes of innovation ' BIM, ' GIS : Cartographic information system, ' Energy management, ' Mobility tools, ' IOT A strategy focused
on innovation 4 23,6 M' INVESTED IN RESEARCH AND DEVELOPMENT 45% INNOVATION CONCEPTION DEVELOPMENT 55% EVOLUTION AND MAINTENANCE (technology, regulation.. RESEARCH & INNOVATION 5 OUR PARTNERS TECHNOLOGY RESEARCH & INNOVATION COMMERCIAL 33 69 25 Examples 127 RESEARCH & INNOVATION 6 Digital revolution 2. 7 Things'on Internet 50 billions - 2020 200 billions - 2030 8 Industrial IoT market 109 billion $ in 2016 933 billion $ in 2025 Data GrandView Research 9 IT'S ABOUT WORLD SENSING We need data 10 Artificial Intelligence & Cloud Computing Intelligent machines and
Artificial intelligence (AI) in full
takeoff 6 billion Intellingent objects in
2020 95% of the smartphone's users
have used a virtual assistant
(Siri,...) 11 11 TECHNOLOGIES 'Broadband networks
'Virtual Reality/3d
'Cloud 2.0 & Artificial Intelligence (machine
learning)
'Security and Blockchain
'Api ' 12 Digital Equipment Platform 3. 14 ANALYZE DATA TO PREDICT AND RECOMMEND ' Understand: when, why and how a failure occurs ' Predict the consequence (s) of the failure ' Recommend the action (s) to be carried out: maintenance and / or equipment configuration and / or
action on the environment (temperature, geographical
position, . .) Predicting "when the failure wil occur" is not the real issue! Issues Holistic IIoT platform for predictive maintenance 15 PREDICT AND RECOMMEND: A NEW APPROACH Maintenance will no longer be an autonomous function
and a cost center, with the aim of minimizing downtime. Instead, maintenance will be a production function whose
objective is to optimize production and quality! 16 50% failures
30% maintenance costs 20% power
consumption Knowledges & Security 30% Time To Repair Flexibility PREDICT AND RECOMMEND: BENEFITS 18 TYPICAL ARCHITECTURE OF OUR PREDICTIVE SYSTEM Any reproduction prohibited nan os ec on de se co nd m in ut e ' ' ' 20 PROTOCOL & NETWORK INTEROPERABILITY 22 INTEROPERABILITY WITH MANAGEMENT SYSTEMS 25 PROJECT
ENERGETIC (AIR-ENERGY) OPTIMIZATION OF
DATACENTER
26 BUILDING DESIGN There will be a datacenter and office spaces. 2500 m office
240 Kw (heating / air
conditioning)
' Datacenter 30KVA
30Kw air conditioning
The Atrium has a large volume with a
temperature gradient that is
probably high . 27 DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING ' ' ' 28 PARAMETERS TO TAKE INTO ACCOUNT The control of the global heating / cooling system includes about thirty configuration
parameters depending on severals environmental and operational conditions: Outside temperature
Front of servers temperature, etc.
Effective electrical power of the DataCenter (instant energy consumed by servers)
Temperature of the atrium and temperature gradient between the DRC and the 2nd
Floor,Ensoleillement des faades, btiment occup ou inoccup,
Transition between FreeCooling and Air conditioning mode,
FreeCooling of the Data Center used in FreeEating mode ("heat the building").
Season (summer / winter / inter seasons)'. We did not understand some behaviors and the correlation between the different
parameters. We did unitary system control! 30 DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING The digital platform of equipment cooperates with CARL Source (CMMS / EAM) to bring the representation of
reality (sensors, physical measurements in real time) and thus consolidates the digital twin 31 31 DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING For the creation of the digital twin of the freeCooling / freeHeating system the platform collects 65 sensors (among 800) every 15s ! On average:
~ 160 000 measurements / hours.
The collection step is calculated automatically by algorithms to optimize the rate / need for analysis / variability of measurements (Edge
Analytics with unsupervised learning) 34 In addition to the ability to track the status and operation of equipment in real time, the platform offers 'forecasting"
functionalities to predict
the evolution of certain measurements (eg. temperature, air circulation, etc.).
These forecasts will contribute to the development of predictive models The green curve
corresponds to the
prediction of the
temperature in front of
the servers, the red curve
to the temperatures
recorded. The analysis is
carried out over the
period [now -10 hours;
now 3 hours]. The average error
between the predicted
and measured value is
<5%
DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING 35 The mechanisms of
machine learning and
modeling highlight 7
modes of operation
whereas we thought to
have only 3 of them
(freeCooling-freeHeating,
air conditioning, free air).
We have thus better
understood the unstable
states of the system
(energy consumer) and the
abnormal states (leading
to premature degradation
of certain equipment); the
detection of these states
triggers a work order in
the CMMS.
This work order is based
on prepared word order,
some operations contains
automaton configuration
values wich are computed
by the self learning
recommendation system . DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING 36 The mechanisms of machine learning and modeling highlight 7 modes of operation whereas we thought to have only 3 of them (freeCooling-
freeHeating, air conditioning, free air).
We have thus better understood the unstable states of the system (energy consumer) and the abnormal states (leading to premature
degradation of certain equipment); the detection of these states triggers a work order in the CMMS. This work order is based on prepared word
order, some operations contains automaton configuration values wich are computed by the self learning recommendation system . DATA CENTER ENERGY OPTIMIZATION AND HEATING BUILDING 39 In 3 months of operation: Observation of the gap between the planned (from specifications and simulations to the
design of the building) and the real!
Energy optimization: 1% of the time in cooling mode; 99% of the time in freeCooling mode (~
12 KW of electrical energy saved) during winter and 30% of the time in cooling mod; 70% of
the time in freeCooling mode during spring
Understanding the dynamics of the system,
Anticipation of operating modes leading to premature wear and equipment and ultimately a
better prediction of breakdowns
Automated recommendations on GTB control parameters (~ 30 parameters) ' Optimal
operating assistance
Identification of modifications (reverse engineering) to meet the objectives of energy
optimization and optimization of equipment (rate of operation and service life) BENEFITS 40 INDUSTRY PROTOTYPE (IN PROGRESS) ' CUSTOMER PROOF OF CONCEPT Prototype Pilot Site ' ALSTEF AUTOMATION Predictive maintenance of handling system and conveyor 48 www.carl-software.com


© Eiom - All rights Reserved     P.IVA 00850640186