Dedicated to Professor Charles Fairhurst, who introduced me to the wild idea that crustal stresses can be measured by cracking rock.
ABSTRACT: Hydrofracturing has been the universally accepted method of deep-hole stress measurements. As tests are getting deeper and rock conditions more hostile, innovative ideas are surfacing to advance conventional hydrofracturing or carry out independent measurements using new concepts. In the 1990's it is likely that hybrid stress methods will become more common, combining the best that two or more techniques have to offer in order to corroborate test data and increase the reliability of the derived state of in situ stress.
I INTRODUCTION
Evidence of the importance attributed by scientists and engineers to the role of crustal stress in problems related to rock faulting, earthquakes, and plate tectonics, as well as in design of underground mines, tunnels, and hydroelectric powerhouses can be traced back to at least the 1940's. The introduction of quantitative physical models to explain phenomena in structural geology, tectonics, and seismology (e.g. Anderson, 1951), and the development of analytical and numerical methods for the rational design of rock excavations (e.g. Terzaghi and Richart, 1952), required knowledge of the pristine stress regime. It was soon realized, however, that rock stresses cannot be predicted but have to be measured, and thus a whole new field of study was initiated in the 1950's dedicated to finding a reliable method of determining stresses in the earth's crust. Most of the suggested techniques over the years have been in situ types. The first two decades of this study were dedicated to overcoring methods. Teams on three continents devised instruments for indirectly determining stresses in rock through the measurement of strain on the wall of overcored testholes (Hast, 1958; Leeman, 1964; Obert, 1962). These methods have suffered from several problems including theoretical (rock seldom fits the idealized characterizations used in strain-to-stress conversion, such as linear elasticity, homogeneity and isotropy), and technical (overcoring difficulties often limits depth of measurement to several tens of meters) rendering overcoring limited in scope.
The early 1970's saw the introduction of hydrofracturing. This new stress measuring method was revolutionary in that it was especially suited for rocks. It exploited rock fracturing characteristics, and yielded at least part of the stress tensor directly (typically the least horizontal principal stress S h and its direction) and independently of the material properties (Hubbert and Willis, 1957; Haimson and Fairhurst, 1970). This major advantage over previous techniques, plus the unlimited depth of testing afforded by its simple and rugged downhole equipment, turned hydrofracturing into the preferred in situ stress method both in earth science research and in preexcavation design of underground structures (Haimson, 1972, 1977a). Successful tests down to several thousand feet (Haimson, 1978a; Zoback et al 1980), consistent results over large areas belonging to same tectonic region (Haimson 1977b, 1978b) and good correlations with other geological or seismic indicators (Zoback and Zoback, 1980) have served to reinforce the confidence in hydrofracturing.