Journal of Mining Science, Vol. 46, No. 3, 2010
MINERAL MINING TECHNOLOGY INTEGRATED MODEL OF THE COAL OUTLET STREAM IN SURFACE MINING OF COAL SEAMS
A. A. Botvinnik
UDC 622.271.3/519.21
The paper describes an algorithm for quality stabilization in coal outlet stream formed in coal seams mining. The desired quality indices become achievable due to rating face excavation in coal seams and owing to accumulating storages for low-quality and high-quality coals. The algorithm-based modeling is exemplified with a case history of the coal quality control on the Elginskoe Coalfield. Quality index, coal seams, stabilization, accumulating storage, optimization
Efficiency of a surface or underground mine is governed by the quality of coal the mine serves out to a processing plant or directly to a power engineering facility, so it is important to know the natural quality of coal, to form an outlet stream of coal from mine and to beforehand impart it with the required characteristics. Coal is featured by nonuniformity of quality distribution over a seam, which causes quality changeability in the coal outlet stream a surface mine produces. In case of coal seams mining, the situation gets even more difficult as the outlet streams are delivered from different seams. So, prediction and monitoring of the coal outlet stream quality is a topical problem and is being included in recent programming projects aimed at the full-automatic mining planning and control [1]. The methods and algorithms developed within the frames of such projects are usually not off the press as these projects are generally commercial. The given research has been undertaken in order to develop an integrated model of surface mining of coal seams with realization of engineering solutions on the control of the surface mine geometry and monitoring of the coal outlet stream quality. It is possible to control coal quality by correcting and influencing the following factors: — direction, sequence and intensity of coal seams mining; — arrangement of backup faces; — inclusion of coal accumulation storages into the mining scheme. An effort to solving the task of the coal quality control by means of the outlet coal output blending and arrangement of accumulation storages is described in [2 – 4]. These researchers modeled coal seams by vertical or horizontal planes with a large pitch, and the quality control problem was solved almost by-hand at each stage of mine planning. An integrated quality control model for coal outlet stream in a surface mine must generate a system of the related mathematical models of coal morphology and qualitative composition, as well as of schemes allowing automated simulation of the development and regulation of mining. The mining regulation must include: Institute of Mining, Siberian Branch, Russian Academy of Sciences, E-mail:
[email protected], Novosibirsk, Russia. Translated from Fiziko-Tekhnicheskie Problemy Razrabotki Poleznykh Iskopaemykh, No. 3, pp. 63-72, May-June, 2010. Original article submitted January 19, 2010. 1062-7391/10/4603-0271 ©2010 Springer Science + Business Media, Inc. 271
1) the modeling of the stratigraphic geology of the coal seams; 2) the choice of the mining front advance; 3) the laying out of successive positions of mining operations in every coal seam; 4) the determination of coal amount and its quality variation dynamics within a production stage (cut); 5) the formation of the required quality coal stream from a number of faces; 6) the formation of the required quality coal stream from accumulation storages. The present author described the approaches to the objects 1 – 4 in the co-authored works [5, 6] and is presenting solutions to the other two problems below. The data source for verification of the algorithms and programs is the geological exploration and testing information obtained in seams N15 and N16, Elginskoe Coalfield, South Yakutia. An advantage of using this data source for the model verification is that these seams are nearly horizontal, with a weak-pronounced trough, and we can analyze mining advances in different directions. However, these coals are high ashy (18 – 40 %), and speaking about an improved quality of coal before beneficiation is only possible in terms of stabilization of the outlet coal quality and its conformance with consumer standards. The stratigraphic modeling of the chosen coal seams uses triangulation algorithms [5]. Parameters of geometry and quality have been defined by the data obtained in bores of one and the same prospecting boring pattern, therefore, we use the seam geometry modeling cells also for the coal quality calculation at any point of the seams, and the common calculation procedures for interpolation of each quality index and heights at each point of the (x, y) coordinate layout. So, first, we set a width of a processing cut, chose a mining advance direction, and then plot parallel cuts on the floor of the bottom seam. The cut-tocut advance of the mining front (MFA) reflects surface mine development dynamics (Fig. 1). Plotting of MFA on the top seam needs setting a highwall slope for the entire coal measure. Position of the mining front is defined by an intersection line of the top seam floor and the line of cutting on the bottom seam, and there is a given angle between this cutting line and the horizontal. Surfaces of floors of the seams are not planes, and the seams and their partings have various thickness in different directions, so the cutting lines on the top seam will not be parallel and only keep a general trend. A cut is a production stage and takes certain time. Removing a rock unit forms a spatial prism between the bottom seam floor and the top seam roof (or daylight surface) and planes generated by successive positions of the highwall in a respective cut. Coal volumes versus cuts are shown in Fig. 2.
Fig. 1. Layout: seam N15, Neryungri series 272
Fig. 2. Volume of coal versus cut
To find distribution of quality indexes in the production stages needs that a quality index value is characterized by one number. Integration of values in this case is a rough composition of the quality indexes varying length-wise a cut (on the assumption that the cut width-wise variation is insignificant). Such integration is made with a triangulation model of the seam surface as a set of triangles in space. Assume that the quality index inside a triangle crossed by a cut changes linearly and find its integral value for the cut domain. Let a cut domain L be covered with triangles {Tm } , m = 1,..., M . Then, denote by Lm = L ITm an intersection of the cut and the triangles carpet and by λ m an average quality index inside Lm , and obtain an integral weighted mean quality index for this cut:
∑ λm S ( Lm ) ΛL = m , ∑ S ( Lm ) m
where S ( Lm ) is area of Lm . As the quality index of coal outlet stream from a number of seams, we take Λ L , where weight factors are the areas of cuts on every seam. Figure 3 illustrates variation in the coal quality indexes per cuts.
Fig. 3. Variation of quality indexes per cuts: a — plastic layer thickness; b — ash content of coal; c — volatile content 273
Fig. 4. Variation in the coal ash content in cuts: a — seam N15; b — seam N16
Modeling production dynamics needs setting parameters of mining equipment for each seam, which allows determining the cutting rates and, on this basis, the cutting time. Besides, one should remember that MFA on the top seam is faster than on the bottom seam, as well as it is necessary to account for bench slopes. Cutting is implemented (and modeled) from one of the open pit ends, or from the both at the same time. An example of varying ash content in coal in the associated cuts in two seams is shown in Fig. 4; the coal volume in the cuts is about 3.7 Mt. In Fig. 5 there is a bar graph for the ash content distribution in the coal reserves discussed; and the average weighted mean value of the ash content is 34.7 %. To control an ash content index (or any other index) in the coal outlet stream overall for the coal seams, is it required to set an interval of allowable values for the quality index λ :
{a0 ≤ λ ≤ a1 } ,
(1)
which agrees with the object of having a coal outlet stream with a uniformly stable (in the set range and time) quality index. Determination of the range a 0 – a1 is governed by demands of a customer and also by the information about coal reserves and quality as exemplified in Fig. 5 in terms of ash content. The coal outlet stream is generated by daily coal batches, i.e. a daily face advance in meters, d i , along mining front is determined for each seam. Then, the ash content is determined in a batch, and when batches from different seams are mixed, the ash content is calculated as an average weighted mean of these batches. The daily face advance depends on both a shovel capacity and the coal measure stratigraphy. It is necessary to find a bench where mining advance has the lowest possible rate, and this minimum rate is then used for the lower lying benches in the measure.
Fig. 5. Ash content distribution in coal reserves 274
Let us analyze Fig. 4 as finding λi (s ) for ash content versus cutting length (i is the number of seams in a coal measure). Then, assuming the daily face advance limit [ s j , s j + d i ] , where s j is initial
position of a j-th cutting stage, the coal volume in this stage is found as: s j + di j
Vi = p
∫ ( Fit (s) − Ft
g
( s )) ds ,
sj
where Fit (s ) and Ft g (s ) are functions for the vertical level of the seam roof and floor, respectively; p is cut width. Consequently, an average weighted mean for the ash content in the given volume will be: s j + di
λij
= p / Vi
j
∫ ( Fit (s) − Ft
g
( s )) λi ( s )ds .
sj
Coal beyond the limits (1) goes to accumulation storages. The simplest solution seems as following (we will discuss it in more detail below). Let V j be daily produced coal volume with ash content λ j . In this case, if a 0 ≤ λ j ≤ a1 , then V j enters the outlet stream; otherwise it goes to an accumulation storage. It is clear that such storages are required individually for low-rank and high-rank coals. Thus, we have that our model includes storages for high-rank and low-rank coal, respectively, S h , S l , in amount characterized by volumes Vh , Vl and average ash contents λ h , λl . When coal is fed to a storage or taken out of it, both the volume and ash content of coal in the storage are recalculated. After either S h or S l , or both are in use and not empty, coal that disagrees with (1) is upgraded in terms of the analyzed index and included into the outlet stream. The upgrading of the day batch V j toward λ j conditioned by (1) is implemented through addition of coal from the related coal rank storage. Specifically, when λ j > a1 , the coal is blended with the lower rank (high ashy) coal from storage
Sl in the volume determined as: ⎞ ⎛ λ j − a1 (2) V = min ⎜⎜V j , Vl ⎟⎟ . a λ − l 1 ⎠ ⎝ Corrected like this, the ash content of the coal outlet stream will take the upper limit value a1 , as this is the case of the maximum achievable volume of the output. When λ j < a 0 , the solution is analogous but the addition is from the storage S h and in (2) λl is
replaced by λ h , while a1 is replaced by a 0 . In a general case, neither storage will be empty, Vh , Vl > 0 , and the outlet stream will include extracted coal and storage coal. Whence it follows that the optimization problems is to be solved for every time period j. Let an i-th seam in a measure feed coal volume Vi j with ash content λij . Aimed at maximized output of coal with the assigned quality, we find a coefficient vector inside a unit multidimensional cube, {α i } , i = 1, ..., N + 2 , where N is number of seams in the measure, such that: N
∑α v
i i
i
j
+ α N +1Vu + α N + 2Vl → max
(3) 275
given that: N
a0 ≤
∑ α i λi vij + α N +1λuVu + α N +2 λlVl i
N
∑ α i vi i
j
≤ a1 ,
(4)
+ α N +1Vu + α N + 2Vl
α 1 ≤ α 2 ≤ .... ≤ α N ,
(5)
0 ≤ α i ≤ 1 , i = 1, ..., N + 2 .
(6)
Condition (5) exists owing to that extraction of seams is ordered, and the seams are numbered bottom-up. The problem (3) – (6) is a classical linear programming problem with linear limits and is solved by the simplex-method [7]. The solution outcome is what share of coal each source (seam, storage) pays to a current batch of the coal outlet stream with the required quality. Should there is no solution to meet the limitation set by (4), (5), the solution would be the degenerate vector {α i } = 0 , i = 1, ..., N + 2 . The resultant vector needs correction. The volumes of storages in our problem formulation, Vh and Vl , are unlimited and can accumulate large coal reserves that can be used in the solution of the problem (2) in case of the fulfillment of the conditions (4) – (6). At the same time, a real storage is serviced with loading machinery that has limited daily capacity Vld . Then, if: ∆ = max (α N +1V ,α N +2Vl ) > Vld ,
(7)
α is re-calculated: α i = α i Vld / ∆ , i = 1, ..., N + 2 . Accordingly, the factor Vld / ∆ , lower or equal to 1, is the multiplier for the daily face advance d along mining front, which means that the extraction rate decreases. Inclusion of the condition that max (α N +1V , α N + 2Vl ) < Vld , as opposed to (7), to the constraints of the optimization problem (2) seems inexpedient as this addition is nonlinear and will complicate the simple solution algorithm, whereas the described algorithm for the daily mining advance is natural and efficient. Consider the quality control results for cuts in seams N15 and N16 (see Fig. 4). Based on the distribution in Fig. 5, the chosen ash range for the stable quality coal is A0d between 34 – 36 %. Figure 6 illustrates solution of the problem (3) – (6). As seen in Fig. 6a, the required quality coal is obtained only during approximately a half of the considered time of the production; coal produced in the rest of time has very high ash content and can not be delivered to a customer. The periods when the coal quality meets the set demands are dotted (Fig. 6a and b); within these time periods the problem (3) – (6) has no nontrivial solutions, and coal is transported to a storage (Fig. 6c). Figure 7 shows the coal outlet stream distributed between its sources, i.e. the surface mine and storage. The time period in Fig. 7 is shorter than in Fig. 6 because when 330 days have passed and to the end of this cut development (as seen in Fig. 6c) the coal quality disagrees with the condition (1) and this coal is storaged and does not enter the outlet stream in Fig. 7. So, nearly 20 % of coal in the analyzed cuts is unsatisfying even with the storages, anyway. Inasmuch the coal outlet stream involves coal volumes that are storaged but initially fed from the surface mine, we have that the share of such coal is rather high (Fig. 7) and often exceeds the direct coal contribution from face. 276
Fig. 6. Extraction at one end of the surface mine: a — variation of the coal ash content; b — coal volume in the output stream; c — coal volume in the storages; 1 — uncontrolled production; 2 — controlled production
The above example describes the situation when a surface mine is developed from one end of the pit, while the other end is not involved into production. Mining from the other end somewhat differs though is of the same character. The results are qualitatively different when a surface mine is developed from the both ends simultaneously (Fig. 8). It follows from the comparison of Figs. 6 and 8 that the ash content in the outlet stream of coal generated on the both ends of the pit is less dispersed (1.4 in Fig. 6a and 0.9 in Fig. 8a); this is achieved by reducing the quality index span through blending the low-ash and high-ash coal extracted on different ends of the surface mine. As a consequence, the quality control (dotted line in Fig. 8) is more stable as against the one-end development of the mine field (see Fig. 6). The coal outlet stream generation by a forming source is also more uniform (coefficient of variation is decreased more than by a factor of three) as compared to the one-end extraction (Fig. 9).
Fig. 7. Forming sources of the output coal stream 277
Fig. 8. Characteristics of coal in two-end extraction: a — ash content; b — volume in the outlet stream from the open pit; c — volume in storages; 1 — controlled quality; 2 — uncontrolled quality
Fig. 9. Forming sources of the coal outlet stream in the two-end extraction
So, the pre-set characteristics of the coal outlet stream quality can be formed at any stage of surface mining, which increases coal processing and offtake efficiency. The calculation of the extraction alternatives (see Figs. 6 – 9) takes almost no time. Once calculated and built, the triangulation models of seams and surface mine field (see Fig. 1) are then used on-line and posses many archived realizations for selecting a quality planning alternative by choosing the limits (1). To conclude with, this paper illustrates feasibility of the coal quality control in the outlet stream from a surface mine based on the integration of geodata algorithms for planning open pit geometry change and coal quality dynamics by stages of surface mining. On this basis, an integrated control instrumentation can be developed for the coal product quality, considering future utilizations of this product. The author is thankful to Prof E. V. Freidina and PhD A. N. Dvornikova for the invaluable advice and consideration. 278
REFERENCES
1. Yu. E. Kaputin, Information Technologies in Mining Planning (for Mine Engineers) [in Russian], Nedra, Saint Petersburg (2004). 2. E. I. Ridel, Coal Quality Planning and Optimization for Surface Mines [in Russian], TSNIEIugol, Moscow (1980). 3. V. P. Kapustin, V. N. Sin’chkovsky, M. E. Red’kin, et al., “Computer-aided short-time planning of mining based on blending,” in: Open-Cut Mining Mechanization [in Russian], Issue 2, Nedra, Moscow (1983). 4. P. P. Bastan, E. I. Azbel’, and E. I. Klyuchkin, Ore Blending: Theory and Practice [in Russian], Nedra, Moscow (1979). 5. E. V. Freidina, A. A. Botvinnik, and A. N. Dvornikova, “Geoinformation technologies for coal quality control,” in: Industrial, Scientific and Educational Resumes and Development Problems in the Area of Surface Mineral Mining. Proceedings of the International Science-and-Technology Conference Dedicated to the 75th Anniversary of Prof V. S. Khokhryakov [in Russian], UGGA, Ekaterinburg (2002). 6. A. A. Botvinnik and A. N. Dvornikova, “Geoinformation technology based control of the quality of coal stream from a number of faces,” Gorny Inform.-Analit. Byull., No. 4 (2007). 7. V. R. Karmanov, Mathematical Programming [in Russian], Nauka, Moscow (1980).
279