General features of the hydrogeology of the Kherzet Youcef deposit:
Hydrogeological studies carried out successively by SONAREM (1973-77) and EREM (1981-83), have revealed the existence of two (02) aquifer systems separated by the Kherzet Youcef fault:
(1) The Barremian carbonate system (CBS and CBi) west of the fault where the mineral deposit is located,
(2) The Neogene-Paleogene Quaternary system east of the fault, in the Ain Azel Depression. (NLT).
The first system is characterized by the presence of centimetric karsts (3-5 cm) and by strong cracking (cracks, diaclases and faults with low East-West rejection) due to local brittle tectonics. These two characteristics define the filtration and storage capacity of groundwater, which is very abundant judging by the pumping rates recorded at different levels of exploitation of the deposit (approximately 300 l/s).
The pumping tests (SONAREM study 1973-77) and the observation of hydrogeological borehole on either side of the fault, showed that the two aquifer systems could be considered as having no connection. (Sonarem, 1979).
This finding was confirmed by Bellouche (1997). This author performed a statistical analysis (PCA) of the drawdowns caused by the pitting pumping of the mine to study its hydrodynamic functioning. He concluded that "the use of this method allowed us to support the hypothesis on the sealing of the Kherzet Youcef fault".
Finally, all of these aquifers are fed by rainfall and by underground flows from the Jebel Boutaleb and Hadjer Labiod reservoirs, located south and southwest of the deposit.
After the accident on June 02, 1990, mine engineers discovered that the second system actually consists of two aquifers.
(1) A shallow quaternary aquifer of low productivity (a few l/s) where the free water table is located in sand, pebble and gravel formations, interspersed with clays.
(2) A Neo-Paleogene aquifer with a charged water table (recommended operating flow: 10l/s) located in limestones, sands and conglomerates, attributed to the Miocene in the North and the Eocene in the South. (NCK)
These two aquifers are isolated from each other by an impermeable layer of argillite and marl (Fig.5).
A longitudinal geo-electrical section was drawn (Fig. 6). It extends from the Barremian calcareous dolomitic outcrops in the West (Not far from Kherzet Youcef), to the Miocene mound of Draâ Ben Aibouch in the East.
The cup finds the same collapsed structure, with a very resistant base (>500 ohm.m) probably calcareous dolomitic in nature, corrugated and notched in several places by sub-vertical accidents.
The sedimentary cover, very heterogeneous, has a thickness that varies from 75 m at the location of the shoals (IIIK), to 600 m at the right of the tectonic pits (IIId).
Apart from the upper level, 170m thick at most in the third, everywhere else the filling deposits contain a fairly large portion of clay.
Note also the position of the Kherzet Youcef normal fault as well as that of the gallery (overlap) made by the Sonarem at the 775m which clearly confirms the hydraulic interconnection linking the Barremian compartment East and West, on either side of the fault.
Estimation of water inflows by operating level:
From the hydrogeological point of view the Kherzet Youcef deposit is complicated among the typical deposits. The highly developed network of tectonic accidents and the presence of karstified rocks give rise to rather complicated conditions for the circulation of groundwater. The filtration capacities of the rocks are extremely irregular both horizontally and vertically.
The main quantity of water flows into the mine shaft and into the mine works flow through certain open cracks in the tectonic zones, most often enlarged by the process of kartstification to considerable dimensions.
Therefore, we had to apply a conventional averaging of the calculations for the entire surface and depth of the deposit.
For the calculation of possible water flows in the mining works, the formula ‘the large well’ proposed by Troyanski was adapted using the Dupuit formula. According to this method, mining works will have the same area as a well of the same size.
- H: thickness of the aquifer, (m)
- s: lowering of the water level in the ‘large well’, taken equal from the static level to the operating depth, (m)
- R: radius of influence, (m)
- r0: radius of the ‘large well’, (m)
- K: permeability coefficient, (m/24h)
The mean permeability coefficient is calculated according to the Dupuit formula, using the results of the pumping tests (tracking the drawdown) at piezometers (H1, H2, H4, H5, H6, H7 and H8) around the mine (Fig.7) and the pumping test from the mining well Pit III.
The radius of the large well, it computes according to the following formula: where:
F, represents the area of the deposit, taken equal to 60 000 m2, (length of the mineralized field is of the order of 1000 m and the average width is about 60 m).
After the numerical application of Dupuit's formula, the results obtained are recorded in the following Table 1:
Table 01: Estimated water inflows by operating level
Depth (m)
|
Operating level
(m)
|
Flow rate
(m3/h)
|
160
|
835
|
420
|
280
|
715
|
596
|
340
|
655
|
700
|
400
|
595
|
778
|
460
|
553
|
828
|
520
|
475
|
851
|
Fig. 8 shows the correlation between the flow rate of the water flows and the depth. We note that there is a strong positive correlation between the depth and the flow of the water flows.
The exhaust flows actually realized are higher than these calculated flows (ENOF, 2002). The actual discharge rates for operating levels 775 and 835 are shown in Table 2 below:
Table 2: Exhaust flow rate by operating level
Operating level
(m)
|
Discharge rate
(m3/h)
|
Expected flow rate (m3/h)
|
Margin of error
(%)
|
835
|
531
|
420
|
26
|
775
|
740
|
596
|
24
|
692
|
823(1000 before accident)
|
700
|
18
|
We notice that the margin of error is not large enough (an average of 23%). The expected flow is less than the flow actually extracted from the mine. Due to the fact that we did not take into consideration, during the calculations, the discharge rate that is carried out during the tests (according to the documents of the mine).
The disaster of 02/June/1990:
The disaster of 02/June/1990 caused the death of 19 workers. The rupture, of this day, is located at level 775 on the side of the Kherzet Youcef fault, and caused the emptying of the captive sheet formed by the karst limestones. The piezometry of this sheet is not known before the accident. The emptying of the captive web is translated by a rise of 200 m and a filling of 225000 m3 of vacuum created by the mining works.
Cracks appeared in the ground a few months before the accident on June 02, 1990. These cracks developed several days before the accident and evolved into gaping cracks.
In the opinion of the engineers working on site, the cracks were minimal and limited around shaft I: they did not begin to evolve until August 1989.
These cracks were attributed to surface effects due to collapses of old galleries. Work was then undertaken to map and record these surface cracks (Fig. 9).
Interpretation of the causes of the accident
The activation of the pumping system in the well N°3, intended to ensure the progress of mining work, changes the hydrodynamic situation in a sudden way. The large drawdowns and partial drying of the CBs causing essential hydraulic disintegration of the hydrogeological units. This hydraulic decay occurs at the "macro" level (significant difference between the piezometric levels in the different units), as well as at the "micro" level (because of the well-expressed stratification of the CB and the embarrassed bonds among the different permeable ones).
Over time, we arrive at a highly disturbed piezometry, with large differences in hydrostatic pressures and hydraulic gradients in the different parts of the massif, that is to say, we arrive at a new distribution of tensions in the massif. By adding the presence of numerous mining works (galleries, shafts, descendents, etc..), it is not difficult to assume by the appearance of several cracks and deformation of the terrain.
The catastrophic accident is only a result of this general destabilization of the massif. The movements, even if quite small, threaten to produce new underground cracks, especially at more vulnerable points. Such a point turned out the overlap of the gallery at level 775, located in the immediate vicinity of the Kherzet Youcef fault, respectively of the karst captive sheet (Fig.10).
The large load (more than 120 m above level 775) in the NCK and the karst character of the water table, resulted in a huge influx of water (several thousand m3/h at the beginning). In a few hours the mine is completely drowned. It is also to sign the extraordinary rains fallen (and the flood passed in the wadi crossing the site) during the 3 - 4 days preceding the accident. The penetration of the waters into numerous cracks (open and quite deep) contributed to the further weakening of the massif (mainly clay layers) and to the internal movements that occurred.
Pumping over the following months allowed the lowering of the level in well III and currently it is supported at coast 770 – 775. The flow drawn mainly comes from the source at level 775 (the flow drawn from the lower levels 692 cannot be specified, but it does not exceed 200 to 250 m3/h).
At present, there is a progressive drying of the NCK carried out by the source at the level 775. At the start of pumping, the piezometric level of the NCK is lowered by a few hundred m (based on Pz03 data) and is now below the top of the water table (approximately at hill 800).
The drying of the water table will continue, which means a subsequent decrease in the flow of the source.
The data of the flows pumped during the months of September and October 1990 interpreted, in order to find a valid value of the drying coefficient ∝ of the source.
We obtained: ∝ =0.0043 l/day. Based on this value, we predicted the evolution of the flow rate (Q) over the following 15 months. At the end of 1991, a decrease in the drawn flow can be expected to about 800 m3/h (Fig. 11).
However, it is possible that the decrease in the flow rate will be slowed down if the piezometric level of the NCK later reaches a strongly karstified zone.
The significant lowering of the piezometric level of the NCK could lead to some deformation of the terrain east of the fault. (About very recent cracks have already been noticed in this area).