Environmental Study through NARL Synergy & CEPERC
Post on 15-Jan-2016
32 Views
Preview:
DESCRIPTION
Transcript
Page 1
Environmental Study through Environmental Study through NARL Synergy & CEPERCNARL Synergy & CEPERC
Associate Researcher
Jyh-Horng WuJyh-Horng Wu
Page 2
Objective
• Introduction • Real-time flood watching system• Mobile Real-time flood watching system
Page 3
Introduction
Page 4
Application of GRID Tec. for Natural Environment
Infrastructure
Field monitoring
Flood simulation
Disaster Prediction
Protocol/Working flow
Rescues Dispatch
Resource AllocationResource Allocation
Super computingSuper computing Storage CenterStorage Center Network BackboneNetwork Backbone
Real-Time Data SharingReal-Time Data Sharing
Middleware
SensorsSensors
Grid Application ServiceRemoteRemoteControlControlRemoteRemoteControlControl
P2P P2P SteamSteamP2P P2P
SteamSteamImage Image
RecognizeRecognizeImage Image
RecognizeRecognizeCommunicationCommunicationCommunicationCommunicationGISGISGISGISUser User AccountAccount
User User AccountAccount
Computing GridComputing GridComputing GridComputing Grid Data GridData GridData GridData Grid Sensor NetSensor NetSensor NetSensor Net StreamingStreamingStreamingStreaming
Data Data ProtectionProtection
Data Data ProtectionProtection
Breakdown Notification
Job Job SubmitSubmit
Job Job SubmitSubmit
Ecological Research
Agriculture Research
Customize
Application of GRID Tec. for Natural Environment
Page 5
The most damage typhoons in TaiwanThe most damage typhoons in TaiwanDate Typhoon Name Agriculture Damages in~106
US$Persons Missing
and Deaths
06/30/2004 MINDULLE 2.99 41
09/15/2001 NARI 1.63 104
08/21/2001 BILIS 2.32 21
07/28/2001 TORAJI 2.68 214
10/29/2000 XANGSANE 1.63 89
10/13/1998 ZEB 2.06 38
08/28/1997 AMBER 0.84 1
07/29/1996 HERB 4.92 73
08/23/1994 AERE 0.58 29
08/09/1994 DOUG 1.47 15
07/09/1994 TIM 1.26 23
08/31/1993 DUJUAN 0.89 3
Page 6
Apply Grid Tec. to Flood mitigation
• Flood forecast– Require computing and data storage resources.
• Decision maker– Group-to-Group Communication.
• Sensor Net (Real-Time Data)– Un-limited number of users.– Integrate GIS
• Indicate what and where you are watching– Integrate Mobile devices
• Integrate cell phone systems for popular usage
Page 7
Flood Mitigation Grid: The NetworkFlood Mitigation Grid: The Network
C. River Bureau
Hazard Mitigation Research Center
N. River Bureau
E. River Bureau
S. River Bureau
Taipei Office
Taichung Office
Research Center
Research Center
TWAREN
Linking TWAREN to WRA offices, Research centers, River bureaus
Page 8
Flood Forecast System Flood Forecast System (Develop by NTU)
• After typhoon warning announced by CWB (Central Weather Bureau), WRA (water resources agency) will active the Flood Forecast System
Typhoon & RainfallModel
Forecast
Typhoon related data1.Collect sensor data
2.Search similar history typhoon(Require computing resource)
3. Flood forecast (Require computing resource)
Page 9
Problems of Flood forecast system
• Huge amount of data I/O
• No parallelism
• Only can simulate a single watershed each time (20 min)
• Taiwan has more than 20 rivers.
Page 10
Improve forecast efficiency
• Large scale flood forecast– Integrate computing resource and storage of
NCHC– Simulate all rivers concurrently
• Record and monitor the progress of forecast– For debug/interrupt/pause during forecasting, if
necessary– Computing resource usage report – Record the huge data of forecast result for further
research
Page 11
Grid-based Forecast system architecture
Page 12
Flood Simulations
Page 13
Computing resource monitor
Page 14
Grid-Tec. Decision maker
Page 15
Decision meeting
鳳凰颱風期間員山仔分洪情況
Chair of decision meeting, Director-General, Water Resources Agency; plus 12 River and Water Resource Bureaus during Typhoon Phoenix
Page 16
Real-time flood watching system
• Install remote sensors (including videos) around Taiwan.
• Monitor the conditions of rivers, water reservoirs and pump stations through internet.
• High resolution and frame rate of video stream data.
• Sharing the real-time sensor’s data with other people.
Page 17
Problems• Limitation of bandwidth.
• Limitation of Computing resource.
• Un-limited number of user clients.
User UserUser User
Bottle NeckEx. ADSL 512k
Source
Page 18
Solutions
• Prevent user clients from getting data from data source directly.
• Design a scalable middleware– Use “Multicasting” to reduce the loading of
network bandwidth from data source.
– Use “P2P” to distribute the loading of middleware.
Page 19
Implementation of Multicast
User
Middleware
UserUser User
100M or more
Source 1 Source 4Source 2 Source 3
One Connection
Duplicate Data
512K or less
Page 20
P2P - share middleware’s loading
Duplicate Data
User UserUser User
Source SourceSourceSource
Middleware Middleware
User UserUser User
Page 21
Real-Time Flood Monitoring System
Abnormal detection Device
Remote control Real-Time
Video
GIS map layer control
Search Camera Control
GIS map
Page 22
Auto-Recognize water level
Page 23
History Images
The Recoded Images.
The Date of this Image
Searching Recorded Images
Select a Date
Play Recorded Images Save Recorded Images
Bright/Contrast
Page 24 24
Sensor net river monitoring system found flow over crest of levee in Den-Pau Creek
Flood Sensor Monitoring System was officially used in Disaster Prevention & Response Center During Typhoon KALMAEGI in a decision meeting chaired by premier Liu
More than hundred times of log-in Sensor Net System during Typhoon KALMAEGI
Provide real-time flood scene for premier
Premier Liu
Page 25
Mobile Real-time flood watching system
PTZ control
Position adjust
Change resolution
Page 26
Work flow of mobile flood watch
Page 27
Flood images report
Page 28
Watch mobile data via flood watching system
Page 29
Flood report
Middleware
Database
BT Devicesearch
Upload
Upload Module
Cell Phone Internet ServiceMonitor System
Active CCD
Connect BT GPS
Location Capture
GPS+
Internet
Page 30
Conclusions
• The powerful cell phone has affect human life
• Integrate cell phone, more people will join for disaster mitigation
Page 31
Thank youThank you
top related