Situational Awareness for Construction and Facility/Infrastructure Operations and Management

Transcription

1 Situational Awareness for Construction and Facility/Infrastructure Operations and Management Burcu Akinci

2 Outline A little bit about me. Construction and facility/infrastructure operation problems Motivating technologies Research challenges Main research thrusts Conclusions and future directions

3 A little bit about me Born and raised in Turkey third generation civil engineer PhD at Stanford University Professor at Carnegie Mellon University Co-director of Smart Infrastructure Institute.

4 Smart Infrastructure Institute Performing research towards providing: sensor data-driven awareness of the usage and condition of infrastructure systems (both for components and the entire network) advanced information models and visualization approaches to provide more insight into trends and patterns of behavior proactive, intelligent decision support and control of these critical infrastructure systems over their lifetime The end goal is: a higher quality infrastructure system efficient, effective, secure and sustainable operations Data Acquisition Information Modeling Proactive Decision Support

5 Example Projects q vi (t) v i (t) q b (x,t) q vi (t) v i (t) Sensors Analytics Legacy Data Controllers Visualization Access

6 Complexities of construction projects/facilities challenge the situation awareness and assessment. Are there any deviations? Late detection of defects are costly (6-12%) (Burati and Farrington 1987; Josephson and Hammurland 1999) Where is the valve to shut off the water? Information Problems!!! $1.5 billion is wasted for staying idle. $4.8 billion for verification (Gallaher et al. 2004). How well is my HVAC system performing? Faults in the systems $ billion annually (DoE 2008; EIA 2008)

7 Many technologies exist/can help Virtual models/building Information Models Sensors

8 Virtual World Building Information Models (BIM)

9 BIM is great! html

10 BIM s limitation: Reality 30 Calendar Day Time Series for Production Rate: Activity Name: Excavation 2.35y y y y Estimated prod. Rate = y m3/(man-hr) 1.56y y y y y y y y y 0.76y y y y y y y y 1.46y y y y y y y 1.12y y y y y y y y y 0.52y 0.43y 0.5y 0.4y y y y /2/2003 6/3/2003 6/4/2003 6/5/2003 6/6/2003 6/7/2003 6/8/2003 6/9/2003 6/10/2003 6/11/2003 6/12/2003 6/13/2003 6/14/2003 6/15/2003 6/16/2003 6/17/2003 6/18/2003 6/19/2003 6/20/2003 6/21/2003 6/22/2003 6/23/2003 6/24/2003 6/25/2003 6/26/2003 6/27/2003 6/28/2003 6/29/2003 6/30/2003 Date

11 Reality - sensors ~700MB data per day ~100MB data per scan (in 2 mins) Work Work Work Order Order Order gigabytes of data per site visit 3.4 WOs per 100 sq. ft. per year.

12 Sensors are great!

13 Sensors limitation: Context

14 Integrate Sensors with BIM to Support Situation Awareness

15 Sample Research Issues/Challenges Technology experiments Developing data processing approaches Developing general data fusion approaches Developing representation approaches for depicting integrated Real life as-design is messy! and as-is information Developing approaches for living information models Modeling real life is even messier. Developing approaches for analyzing and visualizing federated information models...

16 Sample research projects Generation of As-Builts during construction using progressive laser scanning Integration of work-orders and BIM to support facility operations Integration of sensors and BIM to support fault detection and diagnosis of HVAC systems Indoor positioning through integration of BIM and various sensors

17 As-built documentation does not always accurately reflect the existing conditions

18 141 Geospatial Deviations in about 200,000 sqft Content Deviation Geometric Deviation Dim=0.9m Dim=1.2m PC Boundary Geospatial Deviation Missing Windows Dim=1.8m Dim=2.2m Δ=0.55m

19 Inaccurate as-built/as-is info issues exist in other infrastructure

20 Generation of As-Builts during construction using progressive laser scanning

21 Integration of work-orders and BIM to support facility operations and update of information Taxonomy of M&R changes & info modules Context-based information based on accessibility Filled Change and topology Change Capture Template Filled Change Capture Template FM BIM Taxonomy of changes Update need identification criteria BIM Generate 1 Link M&R 2 customized M&R change history BIM work change Generate info 1 change history with BIM and capture template identify patterns Request for M&R work customized M&R with BIM and User input by selecting from list: work change info identify patterns - WO location, cause - Get M&R request - Query history linked - WO type - Retrieve capture taxonomy template For the M&R work with BIM - Component type related info of the component - List the query results and ID - Retrieve BIM related info - Highlight the spaces - Generate a template - Query history linked and components Request for M&R work User input by selecting from list: - WO location, cause - WO type - Component type and ID Taxonomy of M&R changes & info modules Capture Template Context-based M&R information fieldwork based on accessibility activity and topology FM BIM Generate 1 Link M&R 2 customized M&R work change info capture template change history with BIM and identify patterns - Get M&R request - Retrieve taxonomy related info - Retrieve BIM related info - Generate a template Legend input M&R fieldwork activity Change Capture Template control For the M&R work output of the component process mechanisms Taxonomy of changes Link M&R Update need identification criteria 2 with BIM - List the query results - Highlight the spaces and components Space needs to be updated Space needs to be updated Query history Query history FM BIM + change history FM BIM + change history

22 BIM and maintenance & repairs

23 BIM and maintenance & repairs From to % of light replacements are in rooms without windows 90% of ceiling tiles are changed on floor 2 below the roof 60% of floor cleaning/replacement needed to be done in PH Floor A rooms with ceiling tile replacement

24 Integration of sensors and BIM to support fault detection and diagnosis of HVAC systems

25 Improving Energy Performance of HVAC Systems

26 Approach Self-Managing Framework for HVAC Systems Information base Condition measures HVAC configuration Building information Self-monitoring Self-recognition Self-configuration Information mediator layer Information repository Self-assessment Self-healing Self-improvement

27 Visual support for decision-making Integrating multiple information models for building and HVAC systems Monitoring and visualizing indoor air condition in 3D environment Diagnosing HVAC system at component level This prototype used the Open IFC 3D Viewer library (Tulke et al. 2010). 27

28 Indoor positioning through integration of BIM and various sensors No silver bullet for indoor positioning/localization yet Poor accuracy (5-10m) of low-cost positioning systems especially for a mobile user (Liu et al., 2007; Mautz 2009) Accuracies poor due to violation of spatio- Bluetooth temporal RFID, UHF integrity (eg. Fox et al., 2003; Taneja et al. 2010) Sensor Deployment Required Not required Infra Red, GPS, GSM, TV signals, IMU WLAN, UWB, Indoor GPS, Ultrasound 5-50m 2-5m < 2m Accuracy for 95% precision Actual path Path output by Inertial sensor Physically not possible to cross walls 28

29 Map-matching through curve-to-curve matching algorithm Average accuracy over 30 rounds ~3m 29

30 Conclusions Situational awareness is needed for effective construction and facility management. BIMs and sensors are complementary: Integrate virtual analyses capabilities with reality. BIMs need to be maintained and updated Living information models Many research issues exist, but opportunities are compelling.

31 Acknowledgements PhD Students: Jan Reinhardt Esin Ergen Kevin Tantisevi Frank Boukamp Hongjun Wang Chris Gordon Semiha Ergan Anu Pradhan Pingbo Tang Fernanda Leite Asli Akcamete Guzide Atasoy Saurabh Taneja Pine Liu Engin Anil Te Gao PhD Students (continued) Varun Kumaraswamy Varun Kasireddi Raghu Sunnam Recent Collaborators: Jim Garrett Mario Berges Daniel Huber Semiha Ergan

32 Acknowledgements

33 Thank you! Burcu Akinci