Challenging Geotechnical Doctrine with Datalogged Instrumentation
What is “Square of the wall”?
Smoky River Coal – 9G Mine Rocky Mountains, Alberta, Canada 200m wide wall 128m depth of cover 3m seam
Rear Pillar Stress Cells 200m wide wall 128m depth of cover 3m seam
0m 95m 128m 200m Rear Pillar Stress Cells 200m wide wall 128m depth of cover 3m seam Geology? Depth Square 5m cell 10m cell Wall Kickoff 15m cell 1.5m cell yielded immediately
Horizontal square (no change in stress cells) Geology (stress peaked and dropped) Depth of cover =wall width (vertical square, stress cells stabilised)
Theory Start of wall loading peak is due to geology, wall width or is it depth dependent? Rear abutment effect is depth or width dependent, whichever is smaller?
Longwall Abutment – How far does it redistribute? 15m? 50m? 9.3 x √h?
Phalen Colliery Nova Scotia, Canada Seam thickness 2.4m Depth 550m
Principal Horizontal Stress Direction Pillar Stress Cells Phalen Colliery Nova Scotia, Canada Seam thickness 2.4m Longwall 256m wide Depth 550m
Not great for magnitude of stress Difficult to install in coal Excellent for change in stress Extremely sensitive to small changes
This graph only shows immediate response to production startup 170m away however stress cells showed 500m away.
Longwall Abutment – How far does it redistribute. 9 Longwall Abutment – How far does it redistribute? 9.3 x √h = 120m Measured 500m!!!!! Kirsch equation = 2-3xD? (wall width 256m)
Kirsch Equations (2 x D)? Phalen Colliery Nova Scotia, Canada Seam thickness 2.4m Longwall 256m wide Depth 550m Kirsch Equations (2 x D)?
Longwall Vertical Stress Abutment Peaks at Gate End?
Pillar Instrumentation Site
USBM Chart recorders (windup spring powered)
Demonstration of peak just behind wall and then pillar yield (3m in, post failure loading and then second yield.
8.8m into pillar still increasing load 120m behind the wall
No yield at the face except at MG corner
Load still increasing when the shearer cut the pressure cells out of the face and stainless steel tubing was ripped off these chart recorders, drawn through the boreholes and wrapped around the shearer drum
Resulting face and pillar stress contour allows real observation of stress redistribution subtleties MG Corner yield Pillar offload (upper roof cave?)
Longwall Stress Abutment Exponential Longwall Stress Abutment Exponential? Rate of roof acceleration is dependent on face position?
Vertical stress increase in pillar as face approaches Readings on 12hr intervals show linear stress buildup in pillar
Datalogged results show something very different Massive rate (shearer cutting through) Low or negative rate (stress redistributing across pillar)
Longwall Stress Abutment Rate of change in vertical stress (pillar load) is dependent on shearer cutting through at gate end Rate of change in horizontal stress (roof movement) is dependent on pulling gate end shields (goaf cave and redistribution of stress around new void) TARPs that are based on rate are extremely dependent on / sensitive to; reading time relative to shear (shield advance) cycle
Are there other things going on that we are not aware of?
Pillar stress cells (lower seam) being read by a datalogger during a time when no mining was in the vicinity
If read daily this would be the resulting data
If read 3 hourly this would be the resulting data
Read on 30 minute intervals
40m 550m Pillar was 510m below the ocean with a 1m tide
Interburden Ocean Seam Scale of tide relative to seam and stress cells
The government predicted high and low tide peaks plotted against data and discovered not only correlation in time and stress magnitude change (on all stress cells in the pillar) but also demonstrated when tide varied from prediction due to onshore and offshore winds. Even shiftly readings would never have discovered this
Result Accurately measured the effects of 1 Result Accurately measured the effects of 1.2m of tide on a pillar 550m depth of cover (However did not alter mining process at high tide!!)
Same pillar stress cells underwent a strange offload / reload while longwall was 100-200m away, over an 8 day period (*Note the tidal effect still occuring)
Upper Seam Dry Workings Flooded Workings Longwall had recently undermined from a flooded workings to a dry workings Upper Seam Lower Seam Upper Seam Dry Workings Stress cells (lower workings)
Water draining out of overlying goaf Water filling overlying goaf adjacent to cells Water draining out of overlying goaf
Upper Seam Dry Workings Flooded Workings Lower Seam Upper Seam Dry Workings
Water filling a goaf can accept a redistribution of vertical stress - off loading adjacent areas. (Flooding workings reduces load on pillars?)
Can you control subsidence (Smax, Strain) by rate of retreat?
Instrumentation site in sealed/reentered gateroad in overlying workings Longwall in lower seam (2.4m thick, 138m below) undermines instrumentation site
370m 1.8m seam Gateroad 140m 2.4m seam
Automated Piezometer Instrumentation Setup Constant Head Water Tank Continuous Water Hose Survey Monument Piezometers Stations Piezometers in hose but pinned to the floor, as subsidence occurred, hose and survey monuments subsided changing head relative to constant head tank thereby measuring subsidence on 30 minute intervals Programmed Datalogger Lower Seam
Rate and magnitude of subsidence (and strain were different when the rate of wall retreat was different (faster – more)
Undermining at 6m/day Undermining at 4m/day
Subsidence response 140m above is to every 1m shear (geometrically) and almost immediate (< 1 hour) Smax greater when undermined at 6m per day than 4m per day (what would 30m per day do?) Could we actually plan to mine under rivers or infrastructure slower to control subsidence/water inflow risk?
Blast Monitoring for effect on Roof Movement Are daily or shiftly readings sufficient?
Shiftly (8hr) readings on roof exto in grunching section
8hr readings over top of 30 minute intervals
Example 1 Readings on 30 minute intervals
Example 2 What geometry led to these 2 fundamentally different ground responses? Who blasted and didn’t put it on the shift report?
What could have been learned from 10 second intervals? Blast effect is immediate on nearby roof movement but was stepped movement in some cases and led to time dependent stabilisation in others. What could have been learned from 10 second intervals? What could we learn from 10 second intervals on continuous miner cutouts Cut out distances? Time to bolt effect? Importance of pretension? Monitoring drilling of roofbolt holes?
Conclusion What are we missing in the understanding of ground movement and stress redistribution by not continuously monitoring? What could we learn from 10 second intervals on continuous miner cutouts Cut out distances? Time to bolt effect? Importance of pretension? Monitoring drilling of roofbolt holes? Convergence on longwalls? Pull tests? On set of floor heave?
The Challenge Don’t accept the excuse that we don’t have approved instrumentation – push for equipment approval, support suppliers in development, make use of recent developments in comms to CMs and Longwalls to bring it directly to surface Think of what other geotechnical understanding may benefit from continuous monitoring