Hackaday Prize Entry: Device For Seismic Noise Analysis

Whenever there is an earthquake somewhere in the world, our TV screens fill with images of seismic data. Those news report graphics with simplified bite-sized diagrams that inform the masses, but usually get something wrong. Among the images there will invariably be one of a chart recorder drawing a significant earthquake trace on paper, which makes good TV, but is probably miles away from the state of the art in seismology.

We are not seismologists here at Hackaday, so it was extremely interesting to find [Michael D]’s project, Device for Seismic Noise Analysis. In it, he gives a basic primer in seismic sensors, and outlines his take on the subject, a sensitive wideband seismic sensor designed to capture the seismic background noise. It seems that many seismic sensors are designed to capture big events, yet ignore the noise between them from which using suitable software one can glean advance warning of seismic events.

The sensor is a simple design, a ball of significant mass rests upon three piezoelectric microphone elements spaced at 120 degree intervals. An extremely high impedance op-amp circuit converts and integrates the charge from the piezo element to a voltage that can be read by an Arduino Yun which harvests the data. It is a bold claim, but the device is said to have already given advance warning of minor seismic events near its Tennessee test site.

Seismology has featured here a few times before. There was this seismometer using a subwoofer as its sensor, and this project using commercial geophones, just to name a couple of examples.

24 thoughts on “Hackaday Prize Entry: Device For Seismic Noise Analysis

  1. Project makes absolute sense; establishing baseline for measurement. Principle is used in nearly all fields of measurement from big 3d coordinate measurements to microwave network analyzer systems. Looks cool too!

  2. Very cool and I absolutely want to follow the results. In ancient times I studied seismic noise by recording on 16 channel FM tape machines and playing back though an analog forced harmonic oscillator as the frequency was adjusted – over and over.

    Seismic noise, or as we called it, the microseismic noise, varies by a couple orders of magnitude as storms hit shorelines or wave action on a continent’s edge changes, winds hit mountain ranges, etc. With enough sensors in a enough places it seems reasonable that earthquake precursor rubbing can be separated from the seismic noise. That is, provided the fault related motions are a different kind of signal from the noise (The way Russian underground nuclear tests were easy to identity), and that the vectors either point to a unique location, or a pattern of some sort, like East and up on this side of a fault when it is West and down on the other side. Sort of like the field around the poles of a magnet – sort of. What can I say? Deformation is a tensor.

    I assume the sensor output is computed in some way to get ground motion. Most sensor schemes produce a combination of position, velocity, and acceleration with the frequency determining which one is dominant and to get a vector you have to choose one. Unless you get lucky. Maybe Fourier Transform the data and make families of vectors by frequency. Can you take the data in I/Q?

    Hope to see more soon!

  3. GOOD subject for hackers! This field has been much overdue for some jumps in progress, and hackers have full freedom to think outside the academic box.

    In advance of an earthquake what is it animals and some small children detect that often spook them, and what are the sky lights often reported? Would monitoring the geolocal DC ground potential due to the conductance of auroral ground current be possible to calculate changing/trending stresses? What piezo spikes could be detected, geo-located, and how far away? Questions prompt ideas.

    1. Not to be a wet blanket, but there have been loads of 5.4 to 7.1 quakes in the last two weeks, and many in the same place, like Valparaiso or the Philippines. No reports of strange lights, etc. So “often” might be a stretch. Piezo signals from granite and similar rocks over the last 50 years have so far been useless.

      1. Useless so far, but create some phone apps and cloud-compute the data and we’re getting somewhere.
        This technology would be groundbreaking save lives.

    1. The basic requirements are that it be symmetrical and have mass! The way I see it, a dense metal like lead would be ideal because it is dense, easily molded into a spherical shape and it is cheap. Its density would mean that the overall size of the device could be decreased very slightly. I have actually tried a 4 inch steel ball instead of the 3 inch balls I am now using. Although it weighed quite a bit more and it led to more signal output, the weight bent the sensors I was using.

  4. Dear Michael D. I am a seismologist. I have spent more decades on this field of science and engineering in the Pannonian Basin, in the heart of Europe. Sorry for saying that, but your project might be interesting for outsiders, while at the same time full of scientific shortages. This is a typical example of playing with something without a willingness of preliminary learning about it. Nevertheless I appreciate your enthusiasm. Your sensor unfortunately represent a technical concept and level which was in the seismology about 100 years ago (not counting of course the connected electronics). Some of your statements in the description of your project is substantially false and misleading those ones reading that without a more potent knowledge background on the topic.

    The seismology and seismic exploration (mostly hydrocarbon) up today has became a very well studied and established science with excellent practical technical infrastructure, of course including sensors too. Thousands of scientists have done research, during many years, and have written hundreds of thousands of publications about that particular topic and problem of seismic noises you intend to solve by some your instant idea. I kindly recommend you just a little textbook or at least brochure reading about seismic sensors, about earthquakes, and of course about seismic noises. Let explain here the most important information.
    First of all: There is not any sound scientific evidence yet about that the earthquakes can be predicted using seismic noise analysis. If you have seismic noise related to an earthquake then already it IS the earthquake itself.

    Basically there are 3 types of sensors used widely in the seismic measurements. The piezoelectric (eg. hydrophone: https://www.google.hu/search?q=geophone&rlz=1C1AOHY_huHU708HU708&source=lnms&tbm=isch&sa=X&sqi=2&ved=0ahUKEwj0x6OJld7TAhXHjiwKHReID-wQ_AUIBigB&biw=1366&bih=677#tbm=isch&q=seismic+geophone&imgrc=_) the sensitivity of which is proportional the acceleration of the soil motion or the water pressure vibration, the dynamic (geophone: https://www.google.hu/search?q=geophone&rlz=1C1AOHY_huHU708HU708&source=lnms&tbm=isch&sa=X&sqi=2&ved=0ahUKEwj0x6OJld7TAhXHjiwKHReID-wQ_AUIBigB&biw=1366&bih=677#tbm=isch&q=seismic+geophone&imgrc=_), with sensitivity being proportional with velocity of the soil motion, and the static (pendulum) type (seismometers: https://www.google.hu/search?q=geophone&rlz=1C1AOHY_huHU708HU708&source=lnms&tbm=isch&sa=X&sqi=2&ved=0ahUKEwj0x6OJld7TAhXHjiwKHReID-wQ_AUIBigB&biw=1366&bih=677#tbm=isch&q=seismometer) which are sensing the displacement of the ground. The velocity of the vibration is the first derivative of displacement, and the acceleration is the second derivative of the displacement. All seismic events related to different origin has theirs own dominant frequency range. The frequencies are also very dependent on the distance of source and receiver. For example frequencies of several hundred Hz can propagate only no more then few dozen meters in the soil. So if you are recording a signal of such a frequency (because of using piezoelectric sensor) then the source is probably on the end of your backyard. Your dog might just scratch :-) and the source definitively can not be an active tectonic fault on the other end of the state. The seismic sensor types are having also particular frequency ranges, and those must be properly chosen for the seismic wave type being studied.

    The spontaneous acoustic events related to rock cracking (caused by many different reasons) typically have several hundreds Hz. For recording this type of waves are used traditionally the piezoelectric sensors. The hydrophones (pressure sensors) applied in marine seismic measurements are also piezoelectric. The man-made seismic waves generated during onshore (land) hydrocarbon exploration formerly by dynamite explosion and recently mostly by large vibrators have typical frequencies less then 100 Hz. For this type of waves the dynamic geophones are the best. By the way opposite your statements the geophones are very-very sensitive devices. Those can output electrical signal well above their own electric noise (that caused eg. by temperature vibration of metal lattice of the coil) corresponding such a small soil vibration, when amplitude of the displacement less then the size of a carbon atom! The biggest natural seismic events (strong earthquakes) and big man-made events (eg. nuclear explosions) can propagate through the whole planet. These waves and generally all from distant events (eg. ocean waves impacting the shore cliffs) have very low (<10 Hz) frequencies. For sensing these types of waves you should use seismometers (seismic pendulums) which are the typical devices of the seismological stations (observatories). These frequencies hardly can be sensed by a piezoelectric sensor.

    Good luck for improvement your project and … knowledge. Do not hesitate to learn from experts.

    1. Oh.. and academic! I’ll add some relevant detail then, and clarify and expand a bit.

      The auroral current results in a difference of potential at every point of the globe from north to south poles. This pattern should vary with stresses in the various geologic layers by the same means it does in the common carbon pile rheostat, more pressure = less resistance. Has anybody bothered to even measure this? It’s there, and the earth itself is made up of uncountable such voltage dividers, just waiting to be measured at every point possible, but likely needing to be monitored at only a few hundred at most to cover the ENTIRE planet. HInt so you can catch up quicker… it will typically be a slowly varying DC with occasional deviations of interest to determine the location of by time reference and amplitude, but that’s just a matter of using math already well proven and used with precision in CT. Still dismissive? These DC currents are already proven. The melting of telegraph wires during the Carrington event and the auroral caused power outage of the late 80’s, which came from excessive currents the aurora lent to the planet, + at one pole, – at the other pole. A current flowing through resistances forms a voltage divider.

      The piezo effect, was not referring to the sensor in use. LOL. Listening skills… (and agenda prolly)… My reference is to the AC voltage output from the piezo effect of crushing rock/crystal within the planet. This should be measurable as a level of background noise with location of individual large events discernible from an array of sensors that can be up to global. Surely developing a 3D image from a spherical array of sensors could maybe be taught to geologists too?

      Hey… check with your weather guys! They’re already using an upside-down stainless steel salad bowl to measure the instantaneous magnitude and rate of lightning discharges worldwide to indirectly measure the average temperature of the planet’s atmosphere. They think planet-wide! Your turn!

      This is why I suggested “outside the academic box”. I grant though, that every idea should be challenged and tested, but that is a process of respectful back and forth reasoning and query rather than dismissal, and it is a process that IS CRITICAL to the success of the academic world as a whole, but works only when one reads carefully. Hackers have open minds and believe in ideas then search for ways to accomplish, and above all love collaborative constructive discussion, including when negative is needed… Academics have politics to establish reputation. Thank Goodness the academics didn’t get hold of the Wright Brothers… and look at what happened to Galileo! Lord help us, what probably would be done today with that wet nut running naked save for a towel through the streets yelling “Eureka”!

      Now. If you’re willing to review the above detail, there is more that I can discuss on the subject including how to develop a reference to measure the DC potentials against. And I would invite any others to join in as well. Hackers invited, nobody dissed

      1. Wishful thinking and lots of unsupported supposition.

        I spent years squeezing large pieces of granite and various minerals with high quartz content, plus set up femptoamp detectors with vibrating reed electrometers and kelvin connections to geological formations looking for any kind of correlation. I even had an arrangement with the US Dept. of Defense to get notified when underground nuclear tests were set off. I would get a phone call and a count-down and turn the qains way up. The tests shots gave a reference surface wave set and a reflection off the Earth’s core. The result? Zip. Nada. Nuthin. Easily detected on the pendulum (damped forced harmonic oscillator) seismometers. Nothing from the rocks. Electrical currents or changes in electric potential? No. Photon counters in mines? Nothing.

        The arrays of modern sensors in California is pretty darn impressive. What do they see?

        (BTW, watching the seismic “noise” build up, I refused to go set up some new instruments on the day Mt. St. Helens blew. SO I don’t just dismiss the utility.)

    2. Apologies. First read of your post appeared on the site here as a reply to my earlier post, now it does not. I am therefore, out of place speaking as thus. Apologies.

      Yet, my post may be of interest.

        1. Understanding the piezo effect is an easier concept and tells of movement. The auroral current will be more revealing as it measures stress…. which accumulates and leads to movement. They need both.

          1. “The auroral current will be more revealing as it measures stress..” This is the part I do not understand. As far as I know, granite and basalt do not behave like a carbon pile.

          2. Consider the enormous magnitude of the auroral current. Note changes from baseline readings. Geolocate those changes. The Canadian plate that played a part in the late 80’s outage is acting as an insulator, any change would be of interest. Any changes at a caldera as well. Or go to a smaller scale like the movements of Mt. St. Helens before the blast, certainly the subsurface crawling would show changes, now take that small model and expand it to a continent. Anything even mildly ionic moving, cracking, separating, a water table changing, will have an effect.

            We already have a massive dc power supply hooked up to the planet, and nobody is looking at the changes in the paths current is flowing that can be calculated by taking voltage readings across the surface. Remind again, Carrington Event the telegraph wires melted… There is an excess of current available to monitor for changes. It’s there, it’s free, but nobody measuring it, but we’ve seriously experienced it.

    3. Sort of difficult to sort through your pretentious snark to find something to actually respond to. You start off establishing your spotless credentials as a seismologist. No, you are a guy named “geoman” posting anonymously on the internet. You then state that my sensor represents a “concept and level” that was in the literature 100 years ago – ummm, sorry, Mr. Snarkmaster, that’s one of many falsehoods you made in your post. Start off your reply with a link to the century old article you are alluding to, wontcha? I would love to see who was calculating the 3D vector magnitude and origin of seismic waves in real time a century ago.

      You say that I am “misleading those ones reading that without a more potent knowledge background on the topic.” Sorry, Mr. Potentate of all Seismological Pretension, your summary of seismology that follows your accusation is too incoherent to follow – nothing but google searches of pretty seismometer pictures. You then make a lot of statements about frequencies of rocks breaking, etc that has nothing to do to this project. The sensor I am using looks at seismic noise through a lowpass filter passing 15 Hz and below.

      Verbosity does not make one single word of your disjointed, rambling reply meaningful or relevant.

      Oh- when/if you reply, please tell us what you have done for the world.

      1. Pssst….

        They never respond. God does not answer questions! Needs no starship. Students learn this and give up and just haul his VW up on the roof! A few subservient enough and deemed useful may get a pat on the head and doggy biscuit from time to time, but little more. One posing as such a God has to follow the same script lest they be busted, so indeed yes, could be a poser. Let’s assume not though, in which case I then say pity them for they suffer working in a world full of it. I know, I had to work with a building full of them for several years. Better than a third of them were absolute delights to work for, and oft you worked WITH them in wonderous moments of their needing your collaboration so they teach… But also there were those that were nothing more than a pushy god suffering from a disease made of ambition, pride, tenure, and pomp.

        As for piezo event detection, I concede it may well be near non-existent. But seeing as you’re also monitoring changes in dc auroral current (indirectly as local ground voltage) there will be noise as water fills new cracks, water tables rise and fall, so does magma, you WILL see detectable changes. You have a sensor station out there already, may as well watch for the AC noise too, even if all it does is alert to a rare rapid change. Can’t form a theory without a properly documented observation to back it up! Someone also said granite and basalt do not pass current so would not be of any use… but these would be places to still monitor for unusual changes… much of the rest is compressible and do conduct current so do act as a carbon pile. We know this because the Earth IS passing a DC current we easily observe as aurora. It is also already proven that we CAN measure this DC, that’s what melted the telegraph wires… they were grounded at each end. Enough amps to melt wire and start wooden buildings on fire, and enough to cause wide area power outages! Do you actually believe that DC did not fluctuate? Explain how it starts and ends then… at what point/freq do you declare a changing DC to now be AC?

        I don’t care if they listen, my job is to get the idea spoken…. How am I to know in advance I’m speaking to one that’s already done a few iterations, and is destined to do a few more iterations, as an earthworm? I don’t. I just stuck the words in his ear and he can’t get them out! Just doing my part… Can’t make a horse drink! BS! He’s gotten a full glass and it’s in his head now. One cannot un-hear a thing. You have massive power in your voice… use it. Mother Nature trusts casting seeds to the wind… but it works better if you plant it in a place it could grow.

        1. I did not expect him to reply. His sign on name links to a blank facebook page with no information on it whatsoever and the actual name on the facebook page does not lead to any google hits, related to seismology or otherwise. Classic internet troll.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.