**The length of the specimen was reckoned in running meters.
 
Fig. 1. Schematic diagram showing straightening of rolled product:
a) initial moment; b) completion of straightening.
Displacing the movable grip to the left, a straightening force is applied to the rolled product, as a result of which the latter is elongated. The measured (specific) section le1 of the rolled product is elongated by the amount Δl = Prle1 ，and the measurement base, as determined by a special apparatus, attains the end value le=le1+Prle1 =le1（1+Pr）.In tensioning the rolled product, the moving support, owing to the scale mechanism (scale K), is displaced by the amount εre = ΔlK as a result of elongation le1 of the base of the measuring unit, presses against the limit switch, and terminates the straightening process. Thus, constant comparison (scale K, which increases the accuracy of the measurement) of the actual elongation Δl of segment le1 of the rolled product, which is selected as the measurement base, with the reference standard εre takes place during straightening. Since the parameters of the process are selected proceeding from Eq. (I), the throw δ of the limit switch is a part of the value of εre, which is no greater than Δl is of le1, i.e., Δl/le1 = δ/εre≤Pr. Thus, the error Δδ of the throw δ of the limit switch does not exert a significant influence on the accuracy of the reading, on the error of the strain measurement Δl, nor on the quality of the straightening, i.e., the error Δδ of the throw is considerably smaller than the magnitude of the switches throw itself.
The method described eliminates the error of deformation, since (irrespective of the length and waviness of the rolled product) the relationship between the deformation alone and the properties of the metal is maintained automatically. This method is used in straightening batches of rolled products formed from the same grade of material; this leads to an increase in labor productivity, improvement in component quality, simplification of design, and a reduction in the net cost of stretching machines.
Fig. 2. Nomograms of allowable strip sections (b and c denote the width and thickness of the strips): a) St0, St2, St3, 0, 8, 10, 20, 09G2S; b) 12KhI8NIOT, 08KhlSNIOT, 12KhlSNg; 1) σa.2 = 200 MPa; 2) σa.2 = 240 MPa; 3) σa.2 = 300 MPa; 4) σa.2 = 320 MPa
During the experiment, strips were straightened on a series-produced stretching machine under a force P = 1000 kN, stretching strips of different sections to 1000 mm. The stretching ranged from 1 to 3%. The deformation was measured over a measured section with a length le1 = I00 mm, which was adopted as a measurement base.  The deformation in this section was monitored by comparing the elongation of the adopted base with a reference standard calculated from Eq. (2) for Pr = 1-3. An indicator (pisions of 0.01; K = I00) was used; εremin = 100mm; and εrema x = 300 mm.
One complete revolution of the indicator's pointer corresponded to Δl = 1 mm. An MP2 101U4 limit switch, which is secured over the scale of the indicator and which interacts with its pointer, was used to shut off the machine (after completion of the straightening process).
For a switch throwδ= 0.5 mm, the built-in measure was determined from Eq. (3); εa=99.5 mm; this corresponds to 99.5 indicator pisions.
Since the ratio δ/εremin = 0.5/100 = 0.005 < Pr = 0.01, and δ/εremax = 0.05/300 = 0.0017 < Pr = 0.03, we can assume that the accuracy of the measurement is guaranteed within the range of 0.01 mm and exceeds the required accuracy. When εre is maintained with an accuracy to one scale pision of the indicator, Δl= 3% (0.03le1）is maintained with an accuracy to 0.01 mm, the final value of waviness does not exceed 1.1 mm/m, and the extent of the crescent does not exceed 1.2 mm/m (see Table I); this corresponds (in a number of cases, exceeds) to GOST requirements for rolled steel. 矫直机英文文献和中文翻译(3):http://www.youerw.com/fanyi/lunwen_4869.html