A commentary on the DPWH guidelines for earthquake recording instrumentation. By Adam Pascale.
Adam Pascale is the head of the Seismology Research Centre (SRC) in Australia and has worked as a seismologist for over 24 years. He is currently on the Executive Committee of the Australian Earthquake Engineering Society and is on the committee for the 2016 Asian Seismological Commission general assembly, which was last held in Manila in 2014.
[note: this post was updated on 21 June 2019 with additional comments]
The Republic of the Philippines and the Department of Public Works and Highways (DPWH) should be applauded for their policy to improve building safety by monitoring urban environments for response to earthquake. This is a great initiative should be adopted in many more cities worldwide where the risk of damage from earthquake vibrations has a high short term probability.
The DPWH “Guidelines and Implementing Rules on Earthquake Recording Instrumentation for Buildings” outlines which buildings should be monitored due to their size, importance and classification, which is a reasonable approach that will cover a good selection of building types for response. The requirement to monitor three points in tall buildings (above 50 metres) is also solid policy, but the recommendation for buildings with only one accelerograph to have the instrument located in the Ground Floor/Lowest Basement should be reconsidered.
The main purpose of this policy is to gather information about building response to earthquake, to enable rapid evaluation of potential damage. By monitoring the ground movement only, nothing will be learnt about the building response during a seismic event. If a single accelerograph is to be installed in a building, I highly recommend that this instrument be located at the top of the building. Monitoring the ground will only provide us with input motion information, which will be very similar over large areas with similar soil profiles, so top-of-building motion relative to ground motion recorded at a nearby location would give a better measure of structural response.
Ground motion is useful as the input to engineering models that theoretically predict the response of the building, but recording building motion would provide actual structural response data, and it would be of most use when used in conjunction with synchronised ground motion recording.
To ensure that the quality of the data will not be compromised, an issue in the guidelines describing the recording equipment needs to be addressed. The full scale range of the sensor is to be ±2g, which is reasonable for the ground motion expected from a large nearby earthquake. This is described as “Sensitivity” in the guidelines, but that term describes how many volts from a sensor relates to acceleration (e.g. 10V per g). It goes on to discuss the RMS noise of the system being less than 40µg (the smallest signal level clearly visible) which is not possible using the specified 16-bit analogue-to-digitial converter (ADC). Using a 16-bit ADC gives only ±32768 counts of resolution (<96dB) in the recorder. Assuming you have a ±2g range sensor, that means that 1 count (the sensitivity/resolution of the 16-bit recorder) would be equal to 61µg, so a 16-bit system cannot mathematically achieve the noise level specified, even assuming there is no bit-level noise (which there always is).
With each step in the data being this large, there is insufficient resolution for analysis of structural response. Any modern scientific seismograph will be using 24-bit or 32-bit ADCs with at least 130dB of dynamic range, and with a high sensitivity accelerometer the system can achieve a real-world noise level of around ±1µg while still being able to achieve a ±2g full scale range. The basement of a typical urban building can have a noise level of a few dozen micro-g. This quality of instrumentation is required for structural monitoring so that the natural frequency and modal response of the building can be determined. Any system using MEMS technology accelerometers must be ruled out – even the best MEMS sensors have a noise level almost 10 times higher than the required 40µg, they have limited dynamic range (<100dB), and they are particularly poor at low frequency response – an important factor for high rise buildings that generally have long natural period response.
The main aim of all of this building monitoring is to rapidly assess the health of the building after a significant earthquake. With hundreds, possibly thousands of buildings requiring rapid assessment, collecting data manually will be a slow, possibly difficult process due to other emergency actions that may be operating at the same time. It would make sense to have building response data readily accessible to structural engineers so that they can rapidly assess buildings from central control centres. Adding data telemetry to modern accelerographs is a relatively low cost component (a cellular modem) with a small ongoing operating cost (a monthly data plan), so centralised data recording should be considered. Perhaps certain sectors of each city can share data centres to allow the work of building assessment to be spread among the available experts.
The guidelines are a huge step towards improving earthquake engineering knowledge, and with a few minor modifications I believe the program will produce a rich data set that will be of great value to seismologists and civil engineers for many years to come, and provide systems that will be of practical use in emergency situations.
For more information on the policy (and to download a copy) see this article.