From VDC to the Digitalization of Construction Martin Fischer Professor, Civil + Environmental Engineering and (by courtesy) Computer Science http://www.stanford.edu/~fischer
[email protected] § Director, CIFE (Center for Integrated Facility Engineering) § Foreign Member, Royal Swedish Academy of Engineering Sciences
©2016
• 100% funded by industry – Building owners – Design and construction companies – Software and hardware vendors
• 1988-2000 – Building Information Modeling (BIM)
• 2000-2010 – Virtual Design and Construction (VDC)
• 2010+ – Optimize Facility Performance
VDC Certificate Program Graduates Award Ceremony on Sept. 6, 2016 in Sola, Norway
Construction labour productivity has not kept pace with overall economic productivity Last Modified 8/15/2016 5:47 PM Pacific Standard Time
Labor productivity (gross value added per hour worked, constant prices1)
$57 trillion of infrastructure investments required by 2030 (McKinsey Global Institute)
Printed 8/12/2016 9:10 AM Central Standard Time
1 Based on 2010 prices SOURCE: OECD Industry Stats (http://stats.oecd.org)
Data from Roberto Charron, McKinsey and Paul Teicholz, CIFE
McKinsey & Company
4
Challenges affecting the construction sector
Printed 8/12/2016 9:10 AM Central Standard Time
External challenges
Last Modified 8/15/2016 5:47 PM Pacific Standard Time
Internal challenges
▪ Shortfalls in accountability ▪ Talent management ▪ Reinventing the wheel ▪ Failure to adapt to new technology ▪ Coping with complexity ▪ Fragmented value chains ▪ Extensive subcontracting ▪ Complex portfolios ▪ Competitive pressure
Data from Roberto Charron, McKinsey
McKinsey & Company
5
We still produce too much rework …
©2016
… and work too inefficiently
©2016
Would your projects benefit if … Everyone is working on the right tasks at the right time all the time.
We are designing what the client wants.
We will be productive as possible next week.
We are We are certain We are sure that we are installing building everything safely that everything and with the best everything fits. fast and methods. right the first time. We are installing everything We gave the accurately based on the latest, client exactly correct information. Paper-free. what he ©2016
wanted.
What capabilities have really changed in the last decade? Digital modeling
Amount of data we can handle
Communication ©2016
Pictures courtesy SPS, Beck, and DPR
Illustrations of good VDC practice
Pictures courtesy DPR
©2016
Would your projects benefit if … Everyone is working on the right tasks at the right time all the time.
We are designing what the client wants.
We are certain We are sure that we are build everything safely that and with the best everything fits. methods.
We are installing everything accurately based on the latest, correct information. Paper-free. ©2016
We will be productive as possible next week.
We are installing everything fast and right the first time.
We gave the client exactly what he wanted.
VDC method to give the client everything he wants Open whole scope of hospital on budget and 30% earlier than typical Highly reliable construction
Confirm constructability of detailed design
Everyone works with the same plan
Combine everyone’s detailed design ©2016
VDC Overview Client Business Objectives Project Objectives
Integrated Concurrent Engineering (ICE)
Project Production Management
Product Modeling BIM++ ©2016
WHY is the client doing the Client/Business Objectives project?
WHAT does the project team need to achieve?
HOW is the project team accomplishing the project objectives?
BUILDING PERFORMANCE • Usable • Operable • Sustainable
Project Objectives
PROJECT PERFORMANCE
Integrated Concurrent Engineering (ICE)
PROCESS PERFORMANCE
• Buildable Safety, Budget, Schedule, Quality
• Production metrics
Process Modeling Production • Controllable factors of production Management
Product Modeling BIM++
©2016
Example of Application of VDC on a Multi-Family Housing Project in Lima, Peru from the CIFE-SPS VDC Certificate Program Implementation Phase
Vivienda Multifamiliar “Residencial Varela” Monthly Report N° 6 Date: 3 May. 2016 Javier Otiniano Project Manager PROYEC Contratistas Generales S.A.
©2015
USING THE BIM FOR INSTALLATION OF PRECAST SLABS IN BASEMENT CEILING (FLOOR PARKING)
PRECAST SLAB MODULES ACCORDING TO THE SECTORIZATION “7 SECTORS”
-
©2015
INSTALLATION TIME OF PRECAST SLABS: 8-10 min c / u INSTALLATION TEAM: 3 People
ECONOMIC BENEFIT / ITEMS FOR USE OF PRECAST SLABS
31% savings in structure cost
$. 8,430.00
9000 8000 7000 6000 5000 4000 3000 2000 1000 0
$. 5,411.00
$. 15,956.00
$. 1,742.00 $. 373.00 Concrete Savings
Steel 1/4" Savings
Joist and Flooring Savings in Plastering, Blocks Savings painting, ceiling (basement ceiling)
(-)
DIRECT COST FOR USE OF PRECAST SLABS 10000
$. 8,873.00
8000 6000
$. 11,032
4000
$. 2,159.00
2000 0 Lightened Precast Slabs
Solid Precast Slabs
ECONOMIC BENEFIT FOR USE OF PRECAST SLABS
$. 4,924.00 ©2015
WORKING MEETINGS WITH TECHNICAL TEAM
18
EVALUATING THE 4D PLANNING MEETING, WITH TECHNICAL EQUIPMENT ON THE CONSTRUCTION SITE
PROPOSAL FOR IMPROVEMENT MEETING, WITH THE DESIGNER OF THE GAS SYSTEM AND TECHNICAL TEAM
©2015
IMPLEMENTATION VDC
MANAGEMENT, PLANNING, AND PRODUCTION PROCESSES TO ACHIVE THE TARGET
IMPROVEMENT PROJECT
4 D MODEL
SECTORIZATION/WORK TRAINS MAPPING PROCESSES I C E S E S S I O N S
Confirm attendance
DESIGNERS
Identify deficiencies in design
Contributions to improvements
Check queries
Convene Meeting
Present Schedule
Contributions to improvements
PROJECT MANAGER
Decision Making
YES
Review by levels and specialities
BIM SUPPORT
BIM MANAGER
COLLABORATIVE MEETINGS
Develop Modeling Plan
Model Presentation
Identify deficiencies in design
Make queries of the Model
Contributions to improvements
Contributions to improvements
NO
Specify contributions
R E S I D E N C I A L
VARELA REDUCE CONSTRUCTION TIME BY 5% - TARGET GOAL
OPTIMIZED SCHEDULED
CREATE VALUE CUSTOMER ©2015
Making of improvement metrics
Key technology and management developments Mobile • from just-in-case to just-the-right information Cloud • anytime (push and pull, bi-directional, ”unlimited”) Parallelization • fast Location / dimensional measurement • accuracy, dimensional control, off-site / on-site Machine learning • experience and data Robotics, additive manufacturing • virtual ßà real, safety, environmental impact Internet of Things (IoT) • virtual ßà real Virtual Environments • test! Collaboration • concurrent knowledge Lean • lower uncertainty, lower risk, customer, pull, purpose à value ©2016
Wake-up Calls The following slides show several innovations that show dramatic improvement opportunities.
©2016
Daily Process & Feedback Loop START OF DAY
u Reduced data-entry (only once at start and end of day) u Minimal Training required
Anirudh Reddy
u All data and activities verified by Device based Locational and Transactional Data, in the context of the project plan and operations
• Create Material Requests for planned • Collect Material at work on SiteMan App Plant • Write Requests on • Data captured by Smart Card Cabin Controller and • Verify at Weigh-Station Vehicle Unit • Data captured by Cabin Controller and Vehicle Unit
• Whole day’s activity visualized and analyzed on Software portal • Work plan for next day created based on recommendations by Einsite
END OF DAY
• Drop material at worksite, and perform work using relevant machinery • Acknowledgement made by supervisor ID card on • Enter Daily Progress device Reports • Entries cross checked with device Data
einsite
Fully automated design of a house from a sketch (elilbre)
©2016
REDUCING THE COST OF STEEL STRUCTURES USING C O M P U TAT I O N A L D E S I G N O P T I M I Z AT I O N
Wo rk by Fo rest Flager in Co llabo ratio n with Arup and John Haymaker CASE STUDY RESULTS DESIGN PROBLEM
Objective: Minimize steel weight Constraints: Safety and serviceability
conventional design method
FCD (128 cpu) design method
4 hrs
3 sec
39
12,800
216 hrs
151 hrs
2,728 met t
2,292 met t
-
$4 M (-19%)
Variables: 1955 size and shape variables Possible design alternatives: ~ 102435
BiOPT METHOD http://cife.stanford.edu/sites/default/files/TR202.pdf
PROCESS Design cycle time Alternatives evaluated
GEOMETRIC MODEL GEOMETRIC MODEL 1
Total design time PRODUCT
ANALYTIC MODEL 2
Total steel weight OPTIMIZE SIZING
Est. cost saving (USD)
3
OPTIMIZE SHAPE
4 SEQOPT Algorithm (Booker, et al. 1999) FCD Sizing Algorithm http://cife.stanford.edu/sites/default/files/TR201_0.pdf ©2015
• Orders of magnitude reduction in design cycle time • Evaluation of a greater number of design alternatives • Improved product quality
A Tale of Two Moment Frames KEY
3 Continuity Plates
W14X233 W14X176 W14X132 W24X117 W18X86
Doubler Plate
Steel Weight Total Cost Procurement Time
Original Frame
Value-Engineered Frame
-
+8% -13% -20%
Work by Forest Flager, Pratyush Havelia, Henry Hamamji, Filippo Ranalli, Bo Peng, Thomas Trinelle In collaboration with SOM, Herrick, Autodesk CIFE 2016
Copyright Ó 2016
Steel Frame Cost Components Design Cycle: 8-24 weeks SUPPLY CHAIN EXPERTISE
DESIGN DECISION BASIS
Other Material 27% Erection Fabrication
CIFE 2016
Copyright Ó 2016
Automatic as-built BIM from laser scans
In collaboration with Iro Armeni, Silvio Savarese, Amir Zamir, and Ozan Sener
©2016
“The automated execution of processes changes everything.” (Alan Perlis, 1961)
©2016
Flow-based Construction Site Management
IN
N E L LY G A R C I A - L O P E Z C O L L A B O R AT I O N W I T H G R A Ñ A Y M O N T E R O
September 2016
Case study: Applying the flow-based site management method Project info: • Graña y Montero jobsite in Peru • 11 basements + 21 floors • 18-week period (8 weeks on site) • Structural phase Objectives: 1. Can the flow-based model represent the look-ahead plan? 2. Does the method help field managers make decisions during look-ahead and daily planning?
Case study project adopted best practices for production planning
Master plan • Processes • Gross constraints
Takt plan • Sector definition • Quantities • Trade sequence • Crew balancing
Look-ahead plan • Constraints analysis • Productivity • PPC + reasons
Daily plan • • • •
Quantities Productivity Daily PPC Visual planning
19th floor
Example of daily plan Activities
Crew (type & number)
Time
Risk
18th floor
32
So, why was it so difficult for the project team to stick to the plan?
Video link: https://youtu.be/sT6kkvE2WXU
Activities require a set of flows to start execution Koskela (1999)
Flow Key: Labor
Workspace Precedence Materials/components Information Equipment External (Permits, inspections)
34
Activity
Existing construction models do not formally represent, track, or quantify the activity flows Existing transformation view What flow? Build deck form A1
Install slab rebar A1
Pour slab A1
What caused delay? ∆
Build deck form A2
Install slab rebar A2
Pour slab A2 How are these flows affected?
Did one of these flows fail? By how much? Was it something else? Are we representing all the flows?
Existing construction models do not formally represent, track, or quantify the activity flows Proposed flow view What flow? Build deck form A1
Install slab rebar A1
What caused the delay?
Pour slab A1 ∆
Flow Key: Labor
Build deck form A2
Install slab rebar A2
Precedence
Information
Did one of these flows fail? By how much?
Equipment External (Permits, ispections) 36
Pour slab A2 Steel crew availability
Workspace
Materials/components
Concrete crew idle
Are we representing all the flows?
How are these flows affected?
Weekly data collected by the app Aggregate metrics
Activity metrics
Flow metrics
WIP
Activity info
Flow info
Aggregate activity statuses
Activity execution variability
Flow execution variability
10 variables
17 variables
24 variables
High level of detail data about the plan, its execution, and reasons for variability
GyM case study data (18 weeks) Total
Per week
# Activities
1,153
64
# Flows
4,192
232
415,008
23,056
# Data points
Large dataset for supporting:
Performance analytics
Predictive models (data mining + machine learning
Total additional data collection effort: 45 hours
Feedback from field management team “The analytics allow me to dig deeper into the design of the production system … [and] which teams have the best and worst performance.” Project Superintendent “Keeping track of the historical flow performance is key. We might have a hunch about what flows are consistently late, but we don’t have the data to identify performance issues.” Project Engineer “It’s very useful that we now have a tool that formally maps the flows that are needed to execute an activity … we think about these things, but there is no formal tool that allows us to check that all the flows are ready so the activity is not in danger.” Project Engineer
It’s not experience OR data it’s Experience AND data ©2016
I have made all my generals out of mud. Napoleon
©2016
Interact in a virtual environment
Screenshots courtesy Brandon Fischer
• Rapidly understand target, team, roles, progress, obstacles, challenges • Frequent communication, feedback
©2016
Scaffolding
70kg/100kg Tor
Gunnar
200kg/150kg Steinar
82%/100% 1,640kg / 2,000kg 54 #8, 5m
©2016
The future looks different, but the future is now Strategic Implications Ø Projects vs. Corporate Ø Staff Development Ø Partners
44
©2016
Do you want to continue to be a player or become a spectator?
©2016