To fight against earthquakes, we need to know ourselves and our enemy!

As the world becomes increasingly interconnected and interdependent, medium and large earthquakes can cause significant economic damage and losses. A large earthquake in any vulnerable urban center will have a knock-on effect on the national economy of the center and the ability of its businesses to provide services and engage globally. Recognizing that earthquake risk is a global issue, improving earthquake monitoring capabilities to mitigate this risk is a critical responsibility.

A key factor in improving earthquake monitoring is the realization of seismic sensor networks, which requires widespread deployment of seismic instruments and their interconnection. However, the cost and complexity of installing a large number of traditional seismic instruments are high. Integrating IoT technology can provide a low-cost solution while maintaining standard seismic data quality. This article discusses the physics of earthquake and ground motion sensors, the modern instrument standards they follow, and the features they extract. In addition, a system design using ADI solutions is developed for different seismic sensor network applications.

Earthquakes are events triggered by the movement and collision of tectonic plates. The energy generated by the collision propagates around the surface of the Earth's interior in the form of seismic waves. These waves have multiple directions and are divided into body waves and surface waves.

Figure 1. Types of seismic waves: (a) longitudinal wave; (b) transverse wave; (c) Love wave; (d) Rayleigh wave

There are two types of body waves: longitudinal waves (P waves) and transverse waves (S waves). P waves travel in the direction of propagation as a series of compression and rarefaction waves. Due to the nature of their propagation, P waves diverge spherically. Although their wave energy decay is the largest of all types of waves, their speed is the fastest, ranging between 5km/s and 8km/s. The rapid energy decay also makes it the least destructive type of wave. P waves can propagate not only through surfaces, but also through water or fluids.

S waves, also called shear waves, arrive just after P waves. They travel along the Earth's surface at about 60% to 70% the speed of P waves. These waves travel perpendicular to the direction of propagation and the Earth's surface. S waves have less energy attenuation and are more destructive than P waves. P and S waves are collectively called body waves.

Surface waves are 10% slower than body waves but are the most destructive. It is important to note that the speed at which seismic waves travel is strongly related to the type of soil they pass through. Surface waves consist of Rayleigh waves and Love waves. Rayleigh waves are a type of surface wave that travels near the surface of the earth in the form of ripples, which can cause prograde (in the direction of propagation) or retrograde (opposite to the direction of propagation) rotations. Due to the nature of its motion, it is also called ground roll. Love waves travel in a direction orthogonal to the direction of propagation but parallel to the surface of the earth. Figure 1 shows the different types of waves and their effects on the body of the earth.

Earthquake magnitude and earthquake intensity are often confused with each other. They are somewhat related, but they are measures of two different earthquake parameters.

The intensity of an earthquake (intensity for short) depends largely on the characteristics of the location where it is measured. It describes the impact of an earthquake on a specific area and is commonly used worldwide as a traditional method to quantify the vibration pattern and degree of damage. Therefore, there is no real value for earthquake intensity. Earthquake intensity values ​​follow the Modified Mercalli Intensity Scale (1 to 12) or the Rossi-Forel Scale (1 to 10). However, the Modified Mercalli Intensity (MMI) has now become the dominant standard worldwide. Table 1 lists the intensity values ​​in the Modified Mercalli Scale provided by the United States Geological Survey (USGS) and their corresponding impact descriptions.

Table 1. Simple version of the modified Mercalli intensity scale

There are many methods for determining earthquake intensity. These methods use data collected from past earthquakes to create their own ground motion prediction equation (GMPE) to predict intensity values. The derived equation uses at least one ground motion parameter or a combination of ground motion parameters, namely peak ground displacement (PGD), peak ground velocity (PGV), and peak ground acceleration (PGA). Early equations were mainly based on PGA, and several used PGV and PGD. Although GMPE uses data from multiple databases to establish correlations, the values ​​obtained by different models still vary greatly. For example, using Wald's GMPE, a PGA value of 10 cm/s2 gives an MMI value of 3.2. According to Hershberger's GMPE, a PGA value of 10 cm/s2 corresponds to an MMI value of 4.43. Note that most GMPEs follow a power law, where each increase in the MMI value requires an exponential increase in the PGA value. Equation 1 shows the correlation equation created by Wald and Hershberger.

Equation 1 shows the ground motion prediction equations: (a) Wald; (b) Hershberger.

The Japan Meteorological Agency (JMA) has designed an earthquake intensity scale that can be calculated based on strong motion triaxial acceleration data. The acceleration time signal of each axis is Fourier transformed. The bandpass filter shown in Figure 2 (composed of a period effect filter, a high cutoff, and a low cutoff filter) is applied to the frequency signal of each axis. The mathematical representation of each sub-filter is also given in the figure.

Figure 2. Bandpass filtering of the accelerometer output signal used to calculate the JMA intensity: (a) period effect filter equation; (b) high cutoff filter equation; (c) low cutoff filter equation.

After inverse Fourier transforming the filtered frequency signal for each axis, the magnitude of the vector sum of the corresponding time domain signals for all three axes is calculated. The highest acceleration value accumulated for 0.3 seconds or longer is designated as a0. The instrumental seismic intensity is then calculated from a0 using Equation 2, i.e., the JMA seismic intensity equation is solved using the highest acceleration with a duration of at least 0.3 seconds.

While seismic intensity measures the effect of an earthquake felt at a specific location, spectral intensity (SI) measures the amount of destructive energy an earthquake imparts to a specific structure. The SI value is calculated from the velocity response spectrum using the equation shown in Equation 3. The velocity normal period for highly rigid structures is 1.5s to 2.5s. The SI value is for the vibration velocity spectrum, so it is easy to distinguish between seismically active earthquakes or other sources. Therefore, the SI value can be used as a criterion for the effect of an earthquake on the structural health of a building. In addition, the SI value involves simpler calculations than the JMA seismic intensity, which makes it more suitable for low-power applications.

Equation 3 gives the spectrum intensity equation, which is the integral of the vibration velocity response spectrum over the normal velocity period of the building:

The magnitude of an earthquake (or simply magnitude) indicates the energy released by an earthquake at its source. Its value does not depend on the location of measurement. In fact, it has only one true value, which is a number assigned on the Richter scale. The strongest earthquake ever recorded was the one that struck Valdivia, Chile in the 1960s, with a magnitude of 9.4 to 9.6.

The correlation between earthquake magnitude and intensity is not completely well defined. Defining a clear relationship between the two depends on many factors, including the depth of the epicenter, the geological composition of the area surrounding the epicenter, the type of terrain between the epicenter and the measuring device, and the location of the device or its distance from the epicenter. For example, the May 2017 earthquake off the coast of Oregon was determined to be a magnitude 4. According to the July 2017 USGS shake map, the earthquake was felt in Montana at a magnitude of 5 to 6, and the same earthquake was felt in Idaho, but only at a magnitude of 2 to 3. This suggests that even though Idaho is closer to the epicenter than Montana, it does not necessarily mean that the former will feel the effects of the earthquake more strongly.

Earthquake detection is the process of measuring and analyzing seismic waves. Seismic waves refer not only to the motion produced by an earthquake, but any force applied to the ground, even as small as a person walking on the ground, can cause disturbances sufficient to produce seismic waves. The range of ground motions of interest for earthquake monitoring applications is very large. The ground motions produced by an earthquake can be as thin as a piece of paper or as tall as a house. .

Ground motion can be characterized by displacement, velocity, and acceleration. Ground displacement is measured by the distance traveled by the Earth's surface. The change in position can be horizontal or vertical. Ground velocity is the speed at which the Earth's surface moves, while ground acceleration is the rate at which the ground velocity changes with respect to time. Ground acceleration is the most important factor in determining the stress induced in a structure during an earthquake. A presentation from GeoSIG shows the relationship between magnitude, ground motion, and intensity.

Equipment used for earthquake detection is specialized equipment. Applications involving earthquake detection can be categorized according to their frequency range. Therefore, the frequency response curve of the instrument must be appropriate for its use scenario. A chart from GeoSIG shows different earthquake detection applications and the frequencies they cover.

Earthquake detection equipment, commonly called seismometers, has evolved from using traditional pens and pendulums to using electronic and electromechanical sensors. Advances in the design of these sensors have resulted in instruments with different operating frequency ranges, detection mechanisms, and measuring different ground motion parameters.

Historically, seismic instruments could only record ground displacement. Advances in technology have made it possible to measure ground displacement through different mechanisms. A strain seismometer, or strain gauge in general, refers to an instrument that records and measures displacement between two ground points. Traditional models use a solid rod buried or installed in a borehole. The rod is often infused with quartz and other materials that are highly sensitive to changes in length and strain. Changes in length are attributed to small displacements caused by ground motion.

Another implementation is called a volumetric strain gauge, which uses a cylinder with a fluid-filled tube installed in a borehole. The deformation of the container volume causes a change in the liquid level, which is converted into ground displacement by a voltage displacement sensor. Volumetric strain gauges are more widely used in this field because they do not require the special materials required by traditional models.

Recent developments in laser technology have made it possible to create laser interferometers, which have greatly improved the accuracy of strain gauges. These strain gauges use the same principle as the unequal-arm-length Michelson interferometer, with a sensor, laser source and short arms at one point and a reflector at another, which is located at a certain distance. The device converts the changes in the interference fringes caused by the movement of the reflector into ground displacements. The sensitivity and accuracy of this displacement measurement method are proportional to the length of the measurement distance. Therefore, laser strain gauges require very deep underground facilities.

Strain gauges can be accurate to within one part per billion. These devices are often used to measure deformation of the earth or movement of the crust caused by fault movement and volcanic activity. They can measure very low frequency seismic wave signals. However, the differential ground motion of the strain gauge is very small compared to the movement of the suspended mass relative to the ground. Therefore, it is not recommended to use strain gauges to detect ground motion caused by earthquakes.

Inertial seismometers determine ground motion parameters relative to an inertial reference, which is usually a suspended mass. Specifically, ground motion parameters refer to the linear velocity and displacement of the suspended mass. Although the resultant ground motion includes linear and angular components, the rotational effects of the seismic waves are negligible. These velocity and displacement values ​​are obtained from sensors that convert the motion of the suspended mass into electrical signals. The mechanical suspension that controls the motion is related to the inertial forces acting on the suspended mass. The velocity and displacement sensors and the mechanical suspension are the two main components of an inertial seismometer. Developing precision instruments for these two parts is the main design effort of modern inertial seismometers.

Mechanical suspensions require a small restoring force to increase sensitivity so that small accelerations can produce large displacements in the suspended mass. However, when large accelerations from strong earthquake motions act on the suspended mass, the small restoring force will not be able to balance the resulting motion. Therefore, the accuracy and sensitivity of passive mechanical suspensions are limited to a limited range of ground accelerations. Force balance accelerometers (FBAs) remove this limitation by adding a negative feedback loop to the mechanical suspension.

Electromagnetic sensors generate compensation forces based on the position of the suspended mass. This position is converted into an electrical signal by a displacement sensor, which then passes through an integrator module to produce an output voltage proportional to the ground acceleration. The dynamic range of the FBA is significantly greater than that of seismometers with passive mechanical suspension. Therefore, this device is often used in strong earthquake applications. However, the delay caused by the feedback loop limits the bandwidth of the device.

Seismic waves caused by vehicle motion and human disturbances (e.g. mining) have high-frequency ground accelerations. At very low frequencies, ground accelerations are dominated by unbalanced suspensions, ground tilt and thermal effects. Therefore, the bandwidth of a seismometer using ground acceleration is limited to a specific bandpass response. The bandpass response of ground acceleration is equivalent to the highpass response of ground velocity. Therefore, to obtain a wider seismometer bandwidth, seismic signals are recorded in ground velocity. The VBB seismometer is based on the FBA, but instead of using the acceleration of the suspended mass as feedback, its velocity and position are used. The response of the device is very similar to the theoretical response of a conventional inertial seismometer, but its sensitivity and accuracy are not reduced for a wider range of forces.

The trend for an increasing number of seismic applications is to develop networks and arrays of seismometers or seismic sensors, such as for earthquake monitoring, oil exploration, and structural health monitoring. The implementation, shielding, and installation of seismometers are three common constraints for these applications. Mass production and rapid deployment of equipment can directly overcome these three common limitations, which requires a corresponding reduction in the size and cost of seismometers. There are currently two types of sensor technologies that can detect ground motions; compared to FBAs and VBBs, they are very small in size and low in cost.

A geophone is a ground velocity sensor that is lightweight, rugged and does not require any power source to operate. A modern geophone has a magnet fixed to its housing and surrounded by a coil of wire. The coil is suspended by a spring and can move over the magnet. This movement relative to the speed of the magnet induces an output voltage signal.

Figure 3 shows the simulated frequency response of a 4.5 Hz geophone. The frequency response of the geophone is flat in velocity for frequencies above its resonant frequency and rolls off for frequencies below this frequency. Small, low-cost geophones typically have resonant frequencies above 4.5 Hz.

Figure 3. Simulated 4.5 Hz geophone frequency response with a damping factor of 0.56

An equivalent electrical model can be created based on the mechanical specifications of the geophone. Figure 4 shows the electrical model using the mechanical parameters of an SM-6 4.5Hz geophone.

Figure 4. Equivalent electrical model of the SM-6 4.5 Hz geophone using the mechanical parameters from the product datasheet.

To extend the bandwidth to cover lower frequencies suitable for earthquake detection, a period extender can be used. The three most common methods of low frequency response extension are inverse filters, positive feedback, and negative feedback.

At frequencies below the resonant frequency, the inverse filter compensates for the roll-off of the geophone. An inverse filter is constructed by cascading an inverting high-pass filter at the resonant frequency and a low-pass filter with the desired reduction in cutoff frequency. Figure 5 shows the response of the inverse filter and the resulting transfer function when applied. This approach has a number of disadvantages that result in a low signal-to-noise ratio (SNR) for the overall result. Pink noise is amplified by the inverse filter, and it has poor thermal stability at low frequencies.

Figure 5. Frequency response of the inverse filter transfer function and its effect on the simulated 4.5 Hz geophone frequency response.

Positive feedback is achieved by feeding an external current into the geophone coil, which generates a force on the suspended mass. This external current signal is derived from the geophone output signal through a positive feedback filter (e.g., an integrating filter), which amplifies the low-frequency suspended mass motion. In practical situations, the design of the positive feedback filter is difficult to keep stable.

In contrast to positive feedback, negative feedback damps the motion of the internal suspended mass. One approach is to overdamp the current through the geophone coil by reducing the damping resistance. However, this is physically limited by the coil resistance. To reduce the damping resistance to a value significantly lower than the coil resistance, a negative resistance should be added. Negative resistance can be achieved with active devices such as negative impedance converters (NICs). This can be achieved using operational amplifiers (op amps), as shown in Figure 6. Bandpass filters and high gain filters can be added to shape and stabilize the frequency response.

Figure 6. Basic architecture of a negative impedance converter using an op amp.

MEMS accelerometers are motion sensors packaged as a single IC device. The typical structure uses a pair of capacitors and a tiny silicon mass with a metal plate in between. A very thin area of ​​silicon suspends the mass in the middle. Changes in the position of the mass cause changes in the device capacitance, which in turn is converted into a voltage signal proportional to the acceleration of the suspended mass. MEMS devices require power to operate, and some MEMS accelerometers have built-in digitizers to eliminate unwanted noise and eliminate the need for matching sensors and recorders. As shown in Figure 6, the frequency response of a MEMS accelerometer behaves like a low-pass filter with a cutoff frequency at the resonant frequency.

Figure 7. Frequency response of a MEMS accelerometer (ADXL354) on the X-axis.

MEMS accelerometers perform better at higher frequencies below resonance due to offset drift. Conversely, geophones perform better at lower frequencies (but still above resonance) due to their mechanical structure. A small, low-cost seismometer can be implemented to utilize both geophones and MEMS accelerometers to achieve higher device bandwidth. The sensor outputs of geophones and MEMS accelerometers can be converted to different ground motion parameters when convolved with the appropriate sensor transfer function. The paper “Earthquake Sensing: Comparison of Geophones and Accelerometers Using Laboratory and Field Data” discusses the geophone and MEMS accelerometer sensor outputs for the same ground displacement Ricker wavelet based on a common transfer function for each sensor21.

In order to provide repeatability and consistency and support the analysis of seismic signals using seismometer arrays or seismic sensor networks, a set of standards and specifications for the instruments used is required. The USGS has set standards for its instruments to be deployed in the Advanced National Seismic System (ANSS). This section discusses the different specifications required to achieve the desired device performance for a wide range of applications based on the experience and technology trends mentioned by the USGS.

The USGS classifies modern seismometers as data acquisition systems. Compared to traditional seismometers, a standard DAS includes seismic sensors, a data acquisition unit, and peripheral and communication hardware. Equipment can be divided into four categories: A, B, C, and D. Class A instruments are close to state-of-the-art seismometers, while Class D instruments are comparable to traditional seismometers. For a detailed discussion of specifications, see the Instrument Guide.

Seismic sensors that measure velocity and acceleration have different rated bandwidths and frequency responses. The higher the instrument class, the wider its bandwidth and the better its frequency response. Broadband sensors are all Class A instruments with a bandwidth of at least 0.01Hz to 50Hz. Their frequency response to velocity is flat over the frequency range of 0.033Hz to 50Hz.

Short-period Class A sensors have a low bandwidth of 0.2Hz to 50Hz. Their frequency response to velocity is flat only in the frequency range of 1Hz to 35Hz.

Class A accelerometers have a flat frequency response from 0.02Hz to 50Hz, while Class B accelerometers only have a flat frequency response from 0.1Hz to 35Hz.

The sensors used in DAS are classified by the amplitude and frequency range of the seismic signals they capture. Strong motion sensors measure large amplitude seismic signals and are typically accelerometers. Strong motion accelerometers can measure accelerations up to 3.5g with system noise levels below 1μg/√Hz.

Weak motion sensors can measure very low amplitude seismic signals with noise levels below 1ng/√Hz. However, broadband sensors are already capable of measuring low amplitude seismic signals, so weak motion sensors are rarely used.

The sensitivity of the broadband speed sensor is 1500Vs/m. When the maximum output voltage is ±20V, the output clipping level or the maximum measurable speed is ±0.013m/s.

Short period velocity sensors are more sensitive than broadband sensors over a smaller frequency range. For a 1 Hz signal frequency, the clipping level is typically ±0.01 m/s.

The clipping level for Class A accelerometers is greater than ±3.5g while the clipping level for Class B accelerometers is ±2.5g. The dynamic range of a sensor is the ratio of the maximum measurable rms value of the seismic signal to its rms self-noise. However, the rms self-noise of a sensor varies with its bandwidth. Table 2 lists the dynamic range of different seismic sensors over different frequency ranges.

Table 2. Dynamic range of different sensor types: Wideband sensors

Wideband sensor dynamic range (dB)

Table 4. Dynamic range of different sensor types: accelerometer

The linear ground motion components generated by seismic waves exist in all three Cartesian axes. The traditional standard orientation of triaxial seismic sensors is east, north, and up. However, the construction of traditional (and even some modern) seismometers is different for horizontal and vertical sensors because the vertical sensor must account for gravity. The homogenous triaxial arrangement allows the use of similarly constructed sensors to determine the linear ground motion components in the Cartesian axes. The sensors are located at three equally spaced points on a circle centered on the instrument and tilted 54.7 degrees to it (with respect to the vertical). The modified axes can be converted back to Cartesian axes using the equation shown in Equation 4.

Equation 4 shows the transformation matrix for transforming the homogeneous three-axis arrangement into a Cartesian coordinate system.

However, most modern sensors are packaged and designed to support triaxial measurements. These sensors have very small inherent cross-axis coupling effects. Instrument guidelines require that the cross-axis coupling must be less than –70dB of the output signal.

At very low frequencies, the amplitude of ground motion caused by an earthquake can be very small. Data recorders used for seismic instruments are capable of recording signals at a wide range of sampling rates with high resolution. Broadband seismometers require at least 20 bits of data resolution with sampling rates ranging from a minimum of 0.1 SPS (samples per second) to a maximum of 200 SPS. Short-period velocity sensors and Class A accelerometers require at least 22 bits of data resolution with sampling rates ranging from 1 SPS to 200 SPS. Class B accelerometers have lower data resolution requirements, at least 16 bits.

The sampling rate specification takes into account the instrument and its internal data storage. However, advanced seismometers are equipped with more storage space and have access to large network data spaces (such as cloud data services), so they can support sampling rates exceeding the rated specification, which can enable more accurate data analysis and seismic studies.

Seismic signals are only associated with a specific measurement location and time. It is standard for every seismic instrument's data to have a timestamp and a known global location. Every recording from every seismic instrument must be able to have its location attached to it, either through manual user input or through a Global Positioning System (GPS) device or service. Modern seismometers also have built-in real-time clocks or can be synchronized to a precise reference time, such as through an online Network Time Protocol (NTP) server.

There are two main data formats used by seismic instruments around the world: SEG-Y and SEED. The SEG-Y format is an open standard developed by the Society of Exploration Geophysicists (SEG) for processing geophysical data such as three-dimensional seismic signals. Each record includes a timestamp, sampling interval, and the coordinate position of the actual measurement. Details of the format specifications and revisions can be viewed on the organization's website. It should also be noted that there are a variety of open source software for seismic analysis that use the SEG-Y format, but most of them do not strictly follow the specification.

The Standard for the Exchange of Seismic Data (SEED) format was designed to simplify the exchange of unprocessed seismic data between institutions and instruments while ensuring accuracy. Although it is primarily used for archiving seismic records, there are different versions of SEED (such as miniSEED and dataless SEED) for data analysis and processing. miniSEED contains only waveform data, while dataless SEED contains information about seismic instruments and stations.

In order to quickly deploy and implement seismic networks, especially for urban and structural monitoring stations, the design of traditional seismometers must be changed. Remote instruments must comply with current instrument guidelines to make modern seismic signal measurements consistent with and correlated with established data standards. However, the cost and size of the solution should be greatly reduced. Using small seismic geophones and MEMS accelerometers as ground motion sensors, coupled with high-performance ADCs and digital signal processors (DSPs), is a reasonable solution.

The main design consideration for the data acquisition unit (DAU) of a DAS is the analog-to-digital converter (ADC). Traditionally, this is performed by a digital field system (DFS) that acts as a linear successive approximation register (SAR) type ADC and an instantaneous floating point (IFP) amplifier. Figure 8 shows the block diagram of a traditional DFS.

Figure 8. Block diagram of a conventional DFS using an IFP amplifier system.

Discrete implementation of the preamplifier (PA), low cutoff (LC), high pass filter, notch filter (NF), anti-aliasing (AA) high pass filter, and IFP amplifier increases system noise and power consumption. The use of multiplexers increases switching, crosstalk, and harmonic distortion. Most importantly, the quantization error caused by the SAR ADC limits the dynamic range and resolution of the system. Therefore, it is better to design the DAU using other architectures and other converters.

∑-Δ converters take advantage of changes in a signal and add them to the original signal. This reduces the quantization error inherent in SAR ADCs and enables higher resolution and dynamic range. With modern ∑-Δ ADCs, there is no need to implement signal conditioning filters discretely. These ADCs have rich and configurable digital filters that can perform the same functions as traditional signal chains. This effectively reduces system noise and design complexity. In addition, high-end precision ∑-Δ ADCs are capable of detecting multiple channels simultaneously with at least 24-bit resolution.

FIG9 shows a general block diagram of a low-cost seismic sensor node implementation that is flexible enough to be adapted to different applications.

Figure 9. General block diagram of a low-cost seismometer using three homogeneous triaxial geophones and a triaxial MEMS accelerometer.

ADI's three-axis accelerometer solutions that support seismic imaging are the ADXL354 and ADXL356. The digital output versions are the ADXL355 and ADXL357, which integrate a 20-bit ADC and can be directly connected to the processor.

Low-cost, compact geophones detect only a single channel, typically with a resonant frequency greater than 4.5Hz and a sensitivity greater than 25V/m/s. A homogenous triaxial arrangement allows three similar single-channel geophones to be combined into a triaxial ground motion sensor. A period extender is required to extend the geophone bandwidth downward to meet standard instrument specifications for broadband sensors. When the design is powered from a single supply, the period extender can also be used as a gain amplifier and sets the offset of the input signal to the center of the ADC range.

The inherent frequency response of MEMS accelerometers makes them susceptible to offset drift and high frequency noise. Bandpass filters improve the acceleration signal in the frequency range of interest for local seismology. Both the geophone period extender and the accelerometer bandpass filter require precision op amps with low noise, low offset voltage, and low input bias current, such as the ADA4610-1.

The reference voltage sets the measurement range of the ADC and the output signal swing of the period extender. If an analog output sensor is used, the reference voltage value should also take into account the voltage swing of the three acceleration signals. The offset voltage temperature drift of the reference voltage must be very low, especially for outdoor installations (typically 0˚C to 50˚C). The ADR45xx series of ultralow noise and high precision voltage references from Analog Devices is the industry benchmark and can easily meet these requirements. For installations with power lines, such as buildings and stations, the power supply for seismic sensors can be obtained from wired DC power converters; for remote and field installations, the power supply can be obtained from batteries. When obtained from wired DC power converters, low noise switching regulators and low noise, low dropout (LDO) regulators are suitable for the application. Analog Devices LDO regulators, such as the ADM717x series, have high power supply rejection ratio (PSRR), low temperature drift, and low noise. Battery-powered designs require high load efficiency and low power consumption charge controllers and battery chargers to keep the instrument running for a long time without maintenance. In addition, it is better if the instrument can harvest energy from easily accessible energy sources such as solar and thermal energy. The ADP5091 ultra-low power energy harvester features maximum power point tracking and hysteresis mode to ensure the highest efficiency in energy transfer. It has a power path management function that can switch between the harvester, rechargeable battery, or primary battery, allowing self-powered instruments to operate reliably.

If an analog output accelerometer is used, the sigma-delta converter receives three channels of velocity signals and another three channels of acceleration signals from the period extender. This design requires a converter with at least six input channels. If possible, the velocity and acceleration signals must be sampled simultaneously. For multichannel ADCs that switch between channels during sampling, the sampling rate needs to be higher. The maximum frequency of the target signal for seismic is 100Hz. For these signals, the sampling frequency without aliasing should be at least 200Hz or 5ms per cycle. Each acceleration and velocity channel should be sampled at a sampling rate of at least 1.2kSPS. The analysis of seismic signals drives the oversampling of each channel. Therefore, an ADC with a sampling rate much higher than 1.2kSPS should be selected. The AD7768 is an 8-channel, 24-bit sigma-delta ADC that supports simultaneous sampling, eliminating the need for a higher sampling rate. Its maximum sampling rate is 256kSPS, but in low power mode, the sampling rate can be reduced to 32kSPS. It is very flexible, allowing different implementations and applications of seismic instrument design and can easily meet the standard requirements of Class A data acquisition units.

The functionality of the low-cost processor varies depending on the application. For remote nodes that use an external computing device for data analysis, the processor is a data logger that stores and packages the seismic data from all channels into a standard format (SEED or SEG-Y) and then sends it to the computing device through a data interface. The processing requirements for this application are low, so a low-power microcontroller can be used. The ADuCM4050 is an ultra-low-power ARM ^® Cortex ^® -M4 microcontroller recommended for IoT applications. It has low-power modes, with a sleep mode power consumption of 650nA and a fast wake-up shutdown mode power consumption of 200nA. In addition, it has two real-time clock (RTC) peripherals for timing and time-synchronized data sampling.

For stand-alone instruments with built-in data analysis capabilities, the DSP calculates earthquake signatures and other parameters depending on the application, such as building health indicators for structural health monitoring. Earthquake data analysis requires the calculation of various mathematical and statistical functions. For example, the calculation of earthquake intensity requires logarithmic functions and peak detection windows for acceleration and velocity. In addition, the processing time should be short enough to allow continuous data sampling and processing. The ADSP-BF706 is a low-cost, low-power DSP with a processing speed of up to 400MHz, making it the industry's preferred DSP32 for field instrumentation applications. It provides multiple seamless peripheral interfaces, making it easier to connect external devices such as data interfaces and ADCs.

The instrument’s location data can be extracted from a GPS module or set manually during installation. For time data, a low-cost DSP can use its internal RTC peripheral or use NTP through a data interface. There are several options for the data interface, depending on the type of installation. The instrument can use an industrial RS-485 interface for wired communication (especially inside a building), or an Ethernet interface to easily connect the device to an existing data network. For wireless communication, the instrument can use Wi-Fi equipment or Analog Devices SmartMesh® IP33, which enables full data reliability in dynamic environments.

As the number of seismic sensors deployed at various locations increases, the reliability of seismic data also increases. A wealth of information can be extracted from seismic data, which can be used in a wide range of applications such as structural health monitoring, geophysical research, oil exploration, and even industrial and home safety. This section provides an overview of three common applications for seismic sensor networks.

Volcanology and seismology research deploys seismic sensors in rugged (and sometimes dangerous) terrain. Monitoring the internal processes of volcanoes requires ground motion monitoring at multiple points. After certain phases of volcanic activity, these locations may become dangerous and make it impossible to retrieve the seismic sensors. Low-cost, low-power seismic sensors would reduce the cost of the research while maintaining a long service life. Another similar situation is the characterization of plate motion, which also requires the deployment of a large number of seismic sensors along fault lines.

S waves and surface waves are more destructive seismic waves, but they travel slower than the least destructive P waves. This characteristic can be used to implement an earthquake early warning system that detects early signs of an earthquake. In this way, all types of systems have a short time to respond and prevent the earthquake from causing major damage. Residential and commercial buildings will be able to shut down power systems and gas pipelines just before severe ground shaking occurs. Using a network of seismic sensors deployed in multiple locations around the protected area will help increase the allowable reaction time. In addition, false alarms caused by non-seismic sources will be minimized. Figure 10 shows a possible setup for an earthquake early warning system to protect a specific area or structure.

Figure 10. Earthquake early warning system using a network of seismic sensors deployed at multiple locations 6 to 12 miles apart. Image by Erin Burkett (USGS) and (Orange County Register). Courtesy of the USGS ShakeAlert Program35

The response time allowed by the early warning system is proportional to the radial distance of the seismic sensor from the protected structure, as shown in Equation 5. Assuming that P waves travel at a speed of 3.5mi/s or 5.6km/s, and S waves travel at a speed of 2.0mi/s or 3.2km/s, it can be calculated that for every 7.51km increase in the distance of the seismic sensor from the protected area, the response time will increase by one second. In addition, placing multiple seismic sensors at a shorter spacing will provide a higher time resolution for the response time.

Equation 5 shows the relationship between the response time of the early warning system and the radial distance of the earthquake sensor from the protected area.

The seismic safety of buildings can be improved by monitoring and modeling their response to forced vibration testing. Installing seismic sensors in buildings will aid in post-earthquake assessment, response, and recovery. In cases of widespread damage, a widely distributed network of seismic sensors can locate areas of structural damage, reducing the risk and cost of visual inspections. A study on strong motion instrumentation applied this to the Atwood Building, a 20-story steel MRF building, using 32 accelerometer-based seismic sensors deployed on 10 floors to accurately monitor the building's structural health.

Seismic sensor networks are widely used in industrial technology, earthquake research, and structural health monitoring. Application requirements have changed the sensor and system requirements of seismometers, favoring remote systems and lower operating costs. Modern low-cost ground motion detection technology has achieved measurement capabilities comparable to traditional instruments. Using a variety of products from Analog Devices, a detection device that meets different earthquake detection applications can be implemented.