THE PERMEABILITY OF OIL-BEARING STRATA 371 



of the coefficient of increase in resistance for petroleum strata of different 

 deposits (see Fig. 8). 



2. The curves proposed for the Devonian sandstones for the dependence 

 of the coefficient of increase in resistance on the permeabihty <^' *> ^^' i^> 

 differ considerably from the actual curves (Fig. 6). As a result, in the 

 determination of permeability from the specific resistance using the Dolina 

 curve or similar curves (see Figs. 2 and 3) there is a large systematic error, 

 considerably distorting the results (see Fig. 4). 



The main error which is introduced in estabhshing a connection between 

 the coefficient of increase in resistance and the permeability, consists of 

 the fact that the effect of the permeability on K^. ^, and consequently, on 

 Q; is over -read. In other words, it is assumed that the coefficient of increase 

 in resistance is connected with the permeability by the relationship: 



Q = ^Kler 

 where ^ is a very large value. 



Thus, according to Dakhnov ^ ^^ 1, according to Dolina q = 1.56, 

 according to Morozov ^ = 2. In actual fact, as follows from the curves given 

 in Fig. 6, q = 0.3-0.66. 



Of the curves Q =f{Kp^j) suggested for the Devonian sandstones the 

 closest to actuality of the empirical curves is the Dakhnov curve (2 on Fig. 2), 

 although it also differs considerably from them. 



3. It can be seen from Fig. 5 that the connection between the coefficient 

 of increase in the resistance and the permeability are insufficiently close. 

 Therefore, even with the help of the curve Q =^f(K) there is a considerable 

 error in determining the permeability from the specific resistance. Table 5 

 gives the probable and mean square errors in the separate determinations 

 of permeability from the specific resistance using empirical curves Q =f{K) 

 for a number of strata. 



The probable error is calculated directly from Fig. 5 and from similar 

 curves for other regions (the distance along the abscissa between the dotted 

 and continuous curves), the mean square error was taken as 1.48 of proba- 



bility(^7,pagel80)_ 



The error in determining the permeability from the specific resistance 

 even in the most favourable case is high (35-50%;. 



In some cases, for example, when the coefficient of increase in resistance 

 changes little with variations in permeability (a small slope of the curve 

 Q =f{Kper) relative to the abscissa axis), the error is so great that the determi- 

 nation of permeability from the specific resistance loses all meaning; as 

 for example, for the D^ stratum of the Sokolovogorsk deposit. 



24* 



