Expertise
Support
The Great Hansueli
Knowledge
Field Observations
Prototype and Experiments
Technology
Demonstration Site
1st Pilot Installation
One Step Ahead
Information
Utility
Arfang is designed to permanently monitor the natural and artificial avalanche activity, within a delimited area of interest, as a complement to the intrinsically lacunar local human observations (weather, night). In the case of natural avalanche activity, Arfang is a mean of gathering short-term information about the possible transition between a calm
period and a period with increasing activity, and then the end of the activity interval. Also this is a way of detecting isolated avalanches. In the case of artificially released avalanches by means of a remote system, operated at distance with no direct view, Arfang operates as an independent tool for the control of both the fire success and its result in terms of
both avalanche flow, and topographical extension. In other words, Arfang’s main benefit is to automatically disseminate intuitively usable information about part of the non-observable natural and artificial avalanches in a given delimited spatial coverage area.
By late 80’s, a Swiss governmental research unit depending from the military department developed an infrasonic goniometer prototype called “Peiler”, which from its German origin can be translated as “direction finder”. The objectives of the applied researches were to evaluate the potential of passive low-frequency acoustic detection by means of a local array of sensors: acoustic and distance ranges, azimuthal and elevation precision, typology of noise sources emissions and factors affecting detection.
Soon a four sensors configuration prototype was developed for short and middle range detection, on the basis of standard laboratory and computing materials, using as sensors 1Hz lower bandpass limit electret microphones, a MicroVAX computing unit and arranged on a mobile unit for tests in various environments. In early 90’s, the system was able to display real-time detection information in terms of azimuth/elevation/intensity incidence vectors, and the following infrasonic
sources classes in the indicative 1-20 Hz band were well identified or confirmed from previous works published on the same topics : storms, civil and military air transport, heavy industry plants, avalanches. Typical one year monitoring from a remote site in the central alps was conducting to the detection of some 10’000 to 100’000 “infrasonic events”, adopting as a rough definition that an event is something different than noise in the sense that there is “a coherent acoustic wave propagating” through de detection array, by opposition to ambient noise periods where only non-coherent noise is picked-up by the microphones.
Within this huge number of events, the researches were oriented to the related sources identification and classification under various characteristics: trajectories, estimated range given by the vectorial spread of the detection, estimated power related to the intensity, periodicity for civil air transportation, acoustic signature, duration, and others.
As a country mainly concerned by avalanche danger, with strong engineering culture, and remaining a very small geographical entity, people developing the Peiler were early in contact with people from acoustics, and others from snow. At this point, in 1993, crossing all these leading opinions and according to the confirmation of pre-shared physical intuitions and futuristic operational expectations, it becomes evident that the infrasonic goniometry was a well born and high potential application in the field of snow avalanche monitoring: Arfang was born.
Within this enthusiastic context the only deception came from inside Switzerland in 1995 with an astonishing refusal from the world known Institut Fédéral de la Neige et des Avalanches to support more strongly the project - probably for internal friction reasons between two heading federal competiting institutions, of the same small country but each across the language barrier river.
History
Nowadays, after more than a 10 years interval of constant efforts with support and involvement of very various people, ground workers, engineers, motivated public agents, gathering French and Swiss
state institutions and small private companies, among of all and with constant unconditional support Charly Wuilloud head of the Dangers Naturels State Department in Valais, Arfang enters
practice as a powerful intuitive decision-making tool in the field of avalanche danger management.
To localize infrasonic events and recognize potential avalanches, Arfang first determines the direction of incidence of the infrasonic wave with a typical 2 sec. time resolution, i.e. an {azimuth, elevation, intensity} vector. The {azimuth, elevation, intensity} components of the incidence vector are calculated on the basis of the arrival time delays of the infrasonic wave on each pair of the four detectors of which the system is comprised, and that are distributed on the ground in 30 m radius. In parallel the system determines the projection of this
moving vector on the topographical map of the site. The angle of incidence of the wave is provided with a precision of about ± 1° in the horizontal plane. Cumulated with the effects of variation of incidence induced by the local relief, the air temperature and wind speed profiles, the resulting azimuthal precision is approximately ± 2°, i.e. an arc of 350 m at a distance of 5 km. To be able to have the same degree of precision for elevation, the vertical extension of the sensor’s array must be maximised by choosing a medium steep topographical
area for installing the array. Should the vector point in the direction of a preset sector of the spatial coverage area, with a suitable path trajectory and duration, and fit into a given gauge of frequency distribution of energy the event is considered as potentially an avalanche, allowing information dissemination. If the event direction is outside of the area of interest, or too far away (narrowed angle of detection), it is automatically rejected from the event dissemination process.
Events localization and recognition
The detection range depends mainly on the following factors: the infrasonic power of the avalanche, which depends on the type and on the volume of snow mobilized by the avalanche, the flow characteristics, the characteristics of topography, the propagation conditions between the avalanche and the system, the local ambient noise, the relief between the avalanche and the system, the weather conditions, and finally the possible interfering infrasonic contributions.
Concerning the weather conditions, the main factor to be considered is wind. Local wind turbulences induce non-coherent low-frequency noise on the detectors which
deteriorates the detection performance by reducing the avalanche infrasonic signal-to-noise ratio. In case of strong wind, the increase of the aerodynamic ambient noise on the detectors gradually “deafens” the system for the weakest signals of interest.
Factors affecting detection range
Even if the detection of avalanches is still possible with an unfavourable signal-to-noise ratio when using these solutions, wind exposure of the system must still be considered as the major deteriorating factor of the detection performance.
The optimal conditions for detection are thus met for avalanches with a high infrasonic emission level, a propagation in direct sight (no masking relief between the
avalanche and the detectors), and low wind (the other weather conditions do not at all influence the range of the system). Under such conditions, Arfang can easily detect small avalanches at a distance up to 5 km. On the other hand, under unfavourable conditions of detection, the range of the system can remain about 5 km for biggest avalanches but decreases significantly for small avalanches.
Based on the criterion of the duration of the event (width of the infrasonic signature), and on the frequency distribution of the associated infrasound (there is more low frequency energy in the case of the large avalanche), Arfang displays under precautions an intuitive information about the size of the avalanche.
As an automatic system integrating measuring, analysis and transmission functions covering natural and artificial avalanches activity, and for operation within inhabited and developed area, Arfang might be used to provide:
• Warning by using its capacity to disseminate information about the exact time of occurrence of non-observable natural avalanches, which could constitute information that can potentially be interpreted as precursory
of other avalanches, in the view of a coming major episode (e.g.: notification of several consecutive non-observable avalanches after a period without avalanche activity).
• Triggering of alarms by means of its capacity to disseminate information about the downstream topographical extension for both natural and artificial avalanche (e.g.: notification of an avalanche flow reaching a road portion after a remote artificial release).
Both features are commonly confused with the more general terms of surveillance system, or monitoring system. In all cases one must always remember that Arfang’s only capability is to detect “avalanches after they occurred”, i.e. that in any case Arfang can not be expected to be a means for detecting any kind of predictory infrasonic signals that could possibly indicate potential avalanche instability of the snow cover.
Surveillance, Warning and Alarm
Technological Realities
24th, Jan. 2005, seven audibly transposed and visible events within a 10 minutes observation window:
|
1.
|
Helicopter flight, Lama or Alouette turbine type (17.5 Hz), with visible Doppler frequency shift effect indicating non stationary flight sections.
|
|
2.
|
Second helicopter, stationary.
|
|
3.
|
Series of explosions at 15h14m35s, azimuth 239°, 15h16m28s, azimuth 264°, other unmapped explosions at 15h12m10s,15h14m12s, 15h15m09s.
|
|
4.
|
Avalanche at 15h16m36s, duration about 1min46s, potentially in sector 5 with extension down to the road.
|
Example
To balance this disadvantage, the system uses two features. First passive acoustic filters, integrated to the detectors.
Second a technique of numerical cross-coherence signal processing, based on the fact that the individual aerodynamic friction of low frequency wind noise components are uncorrelated from one detector to another, contrarily to the propagating avalanche signals of interest (see example hereafter : 10 minutes raw signal upper graph, showing increasing wind noise contribution patterns, and, lower graph, discriminated avalanche after coherent filtering of the raw signal).
Practical Recommendation
To listen to the event audio file the majority of the computer loudspeakers are far not suitable, in particular concerning the avalanche infrasound’s, of main interest. To be able to listen adequately to the audio transposed file, you will really need to use a good quality headphone, equivalent to the following
model, or an amplified
loudspeaker able to
reproduce low
frequencies
down to 10 Hz.
model, or an amplified
loudspeaker able to
reproduce low
frequencies
down to 10 Hz.
This article is available here (1.5MB) in pdf format. To view a pdf file, you will need Adobe Acrobat.
Introduction
Arfang is a remote system for local range detection of natural and artificial snow avalanches. The usefulness of Arfang is to bring real-time intuitive information about avalanche activity to local people in charge of avalanche hazard management. The operation principle is infrasonic goniometry. Detection is based on a 30 m radius ground array of four infrasonic sensors. The array monitors and digitalized continuously the ambient infrasonic noise, and transfers the recorded signals to a central computing
and web server unit. Should a distant avalanche have induced a coherent infrasound wave propagation through the local detection array, the central computing unit determines with 2 sec. temporal resolution its azimuthal and elevation incident direction, and the associated infrasonic signature characteristics. With no delay, after a first level sorting process to eliminate evident parasitic infrasonic events, the Arfang central then proceeds to the online dissemination of the full event detection
information: azimuthal map, analysis graph, detection parameters, audibly transposed audio file. Those unique features allow concerned people having online all possible information about every detected infrasonic event potentially originating from an avalanche. Added to all other available information, this information is used in the final self-decision process confirming or invalidating the avalanche occurrence