As part of the Community Seismic Network (CSN), we are instrumenting selected buildings in the Los Angeles area. We are using low-cost class-C MEMS accelerometers
in order to allow each structure to be instumented densely.
The main scientific goal of these building arrays is to develop rapid assessment of damage following an earthquake. The belief is that detailed and density measurements are key to doing this. |
![]() |
The MEMS sensors are packaged in a box with a computer and (in some cases) a backup power supply. The computer is an ARM linux computer known as a Sheeva Plug. It interfaces with the sensor (a Phiget 1056), and with the cloud through an ethernet connection. Time is maintained with an ntp connection over the Internet. The data are packetized and sent every 10 minutes, providing continuous readings at every sensor location. |
We currently have sensors in a number of buildings. Two are highlighted here:
![]() |
![]() |
One of the buildings we have instrumented is the Dept of Water and Power (DWP) in downtown Los Angeles. This is a 15-story steel frame building built in 1964. We have placed 30 sensors in this building in communication closets near the north and south ends of each floor. The data are transmitted in near real time to the CSN in the Google Cloud. On the right is a sketch up model of the building developed by Monica Kohler and her students. |
We have recent received permission and funding to instrument the 52-story building show on the left with 104 sensors (2 per floor). These will also be sent in near real time to the CSN/Google cloud. On the right is a structural model of the building. The funded proposal can be found here. We currently have sensors on the 1, 12, 24, 26, and 48th floors, and plan to deploy the full completement by the end of summer, 2014. |
![]() |
![]() | |
The 52-story building is shown on the left and an ETABS model constructed from plans of the building is on the right. Data from the sensors in the builing allows us to calibrate the model. The model will allow us to simulate damage scenarios that can then be compared to the noise level on the sensors to the changes can be detected. |
![]() |
![]() |
Synthetic seismograms generated for each floor using the ETABS model with the measured seismogragram on the 1st floor to drive the model at that point. |
![]() |
![]() |
A comparison of the predicted seismograms (red) and actual measured seismgrams (blue) on the 12th and 24th floors. |
![]() | ![]() |
![]() |
The left panel below shows waves derived from ambient noise correlation
propagating in a 52-story building.
The right panel shows waves from the 4.1 Castaic
earthquake, approximately 70 km from the building.
The plot below shows the spectra for the sensor on each floor suring the
Castaic earthquake.
There insufficient energy in the 0.2 Hz range of the
source spectra to excite the fundamental mode, but the translational modes
up to order 9 are quite clear.
.
Accelerations, velocities and displacements for the ambient noise correlation.
The last column is the inter-floor displacements. |
The same quanities for the Castaic earthquake. |