SNO - STR - 1998 - 021

                          Revision 0

       DESCRIPTION OF THE SNO MICROSEISMIC DETECTION SYSTEM

                     by    J.D. HEPBURN

                         Dec, 1998

*********************************

Table of Contents:

1.	Preamble

2.	Basic Detection Scheme

3.	Mine-Wide System

4.	Sensor Array within SNO

5.	SNO Data Analysis System

6.	Operational Philosophy for the SNO System

7.	Applicability of the SNO System to SNO Neutrino Detector Data

*********************************

1.	Preamble

Any large body of rock has a variety of inhomogeneities in it, such as
faults, high stress zones, dykes, and interfaces between different
rock masses within the rock body.  Normally the masses are stably held
against each other by friction, assisted by clamping forces arising
from pressure from the surrounding rock.  However, when lateral
stresses overcome friction, one rock mass moves relative to another
and a "seismic event" occurs. If these are naturally occurring, they
are called tremors or earthquakes.  Changes in stress pattern might be
caused by plate tectonics, gravitational effects in the earth's crust
caused by the sun and the moon, or man-induced activities such as
mining or building power dams.  Within the mining industry, a great
deal of effort and expense is devoted to analyzing stresses,
stabilizing mining areas, and back-filling mined-out areas.  In spite
of this, some changes in local stress patterns are inevitable.

Additional events occur as rock bursts, where stress is relieved by
rock fracturing and moving in the direction of a stress gradient.  The
more obvious of these are where "loose" (pieces of rock no longer
joined to the rock mass) is created at an opening in the rock.

Blasting in the mine also produces a seismic-like signature.

Acoustic noises are created by all these processes, and occur
frequently throughout the mine.  Activity in a given area is an
indicator of rock instability.  For this reason INCO maintains a
network of hundreds of sensors throughout the mine that is monitored
on a continual basis; SNO also maintains a local system.

2.	Basic Detection Scheme

The detection of seismic activity consists of locating the source of
the energy emission plus estimating the magnitude of the event.  In
this regard, the microseismic detection process is very like that used
for detection of neutrinos in SNO:
 
-- energy radiates from a point source

 -- the energy is detected by sensors for which the time of arrival of
the energy is recorded

-- if enough sensors see energy in a given time window, then the event
is captured as data

-- the times of arrival at the individual sensors triangulated back to
give the location of the point source

All the events listed above give frequencies in the acoustic, audible
range of typically 500 to 5000 Hz so the sensors used are just
microphones that are embedded in holes in the rock.  A crucial feature
of the heads is that they are directional, and have a cosine response
function.  For practical purposes they record sounds in a forward and
backward cone of 45 degree half-angle about the central axis.  The
velocity of sound in rock varies from 18,000 to 20,000 feet per
second.

3.  Mine-Wide Systems

While there are many systems in the mine, the ones that concern us most are:
-- INCO MP250 (source location)
-- INCO / Canmet (magnitudes) (being decommissioned)
-- "Mines Design" (developmental system used by P. Vasak, working
 as part of the CAMIRO consortium, on development of full-waveform analyses)
-- SNO system

The current relevant data flow from 6800 to surface is: analog signals
go by twisted pairs up #8 shaft to 6400 where there are two 64 channel
multiplexers that send both the SNO and the Mines Design data to
surface via fiber optics.  There signals are de-multiplexed, sent to
computers, and digitized by cards in the computers.  SNO channels are
sent to the SNO computer, and in addition certain Mines Design analog
signals are split and sent to both the SNO and the Mines Design
computers.  There are about 12 channels available for transmission of
SNO signals to surface (limited both by a 12 pair cable from 6800 to
6400 and by available channels in the main fibre optics multiplexer).

4.	Sensor Array within SNO


Figure 1 shows the sensors that have been installed at various times,
in the SNO area, with their original numbering.  These are indicated
as numbers within circles.  The three geotechnical stations (numbered
1 and 2 in the utility room, and # 3 in the ramp) are also indicated.
Figure 2 shows the sensor array as it is currently connected at this
time.  Note that while a variety of numbers has been assigned to
various heads (some of which are either not functioning or lost or
buried in sandfill) in the past, as in Figure 1, the present scheme
uses the channel number in the data logger to identify the heads.
Unless otherwise noted, the holes containing the heads are drilled
about 8 feet deep.

Table 1 shows the head-location and interconnections in the system.
The top part of the table shows defunct/not used heads.  Note that
there is a functioning head (original # 3) part way down the ramp
(half way between present numbers # 11 and # 10) that is not
connected/used due to lack of a functioning channel to surface.  The
bottom half of the table lists the actual channels.
  Of the 16 possible channels, 12 are now functioning.  The
geometrical data are

 	"X" = lateral coordinate, northing, in feet
 	"Y" = lateral coordinate, easting, in feet
  	"Z" = depth/ elevation in feet.
for example:   cavity centre =         6837n  3149e  6823d (approx) 
               chiller =               6140n  3940e  6780d (approx)

The bars shown for each head in Figure 1 indicate the approximate
depth and orientation of each bore hole, hence the orientation of the
respective heads.  This is also described in the "location" column of
Table 1.  (Orientation coordinates will be worked out in due course.)

5.	SNO Data Analysis System

The data system runs on a computer with Windows 95 and Access 97 as
the basic operating and database systems.  The microseismic data
logging software is proprietary and purchased from ESG Canada.  The
basic blocks are:
--	WCRON: program scheduler
-- 	GPS: universal time, taken from a signal generated at and used by INCO
--	WINAQ: captures waveforms, performs trigger and capture functions
--	ALOC: locates events from captured waveforms
--	ARCHIVE: saves data from ALOC onto hard drive and optical disk
--	WAVEVIS: allows viewing of seismic waveforms off the sensors
--	SEISVIS: shows locations of reconstructed events, within a
           defined  viewing volume
--	DATABASE: high-level access to stored data

Because the heads are all uniaxial ones, it is impossible to get a
Richter-type magnitude for the event (although, we do have a measure
of signal strength for each event, subject to the directional-
sensitivity issue).

6.	Operational Philosophy for the SNO System

The purpose of the SNO system is to see if there is any seismic
activity in the environs of the SNO cavity, which the INCO systems do
not cover well.  This purpose has allowed/forced the following
conditions:

--	has an "nhit" of 8 with only 12 sensors
--	many events that can be heard at SNO will not be reconstructed because:
--	the limited number of sensors, plus the directionality of
          them, may mean too-few sensors  receive useful signals
--	may be too weak to record
--	source is not within purview of the SNO array therefore does
             not  reconstruct
--	may reconstruct but not be located within the fairly tight 
          boundaries defined within the software as the zone of interest


7.	Applicability of the SNO system to SNO neutrino detector data

a. Use of detected events by the SNO microseismic system in analyzing
SNO detector data

-- events captured by the SNO microseismic system must be fairly
   "robust" ones, because at least 8 heads have detected significant
   sounds.

-- one might expect, therefore, that if there is indeed an effect of
   microseismic noise on the SNO detector, the effect should be there and
   be detectable for most if not all reported seismic events.

-- there is the problem of correlating the two GPS systems (SNO and
    INCO).

-- any such comparisons are off-line

b. Finding microseismic events and times of interest, starting with
the SNO detector data

-- as stated above, many "loud" events will not be captured by the SNO
   microseismic system yet could perturb the SNO detector.

-- using the INCO system as a data source is much better, as it will
   include all significant events

-- there is the problem of correlating the two GPS time signatures

-- the comparison is off-line and, because of the large number of
   events in the INCO system, cumbersome.


c.	Incorporation of microseismic data in the SNO detector data stream
-- the constraints of nhit and reconstruction in the SNO system mean
   it is unlikely to be a good source of complete data, and getting that
   data from the microseismic system to the SNO data stream is difficult.

-- on the other hand, the signals from the heads could be used
   directly within the SNO lab, as:

-- amplify, rectify, add and integrate signals from several heads that
    have a variety of orientations

-- a discriminator set on the resulting signal would trigger an entry
   in the SNO data stream.

-- note that the CMA system is not useful for this purpose, as:
-- it is not based on GPS time
-- it is a polled, not interrupt driven, system so could miss events
-- the advantages of such a system include:
-- depends only on loudness, not location
-- the integration time and discriminator level can be set, with some
     experience, to match the sensitivity of the SNO detector to the sounds

-- timing is based on only one GPS: that of the SNO data stream.
-- a measure of loudness could be recorded in the data stream.









Figure 1: SNO Seismic Array

Showing all heads by the original numbering scheme



Figure 2: SNO Seismic Array

Showing head identification by data logger channel number