There are two methods of determining the value of T0: the single-temperature method, which is used when the tests have been conducted at a single tem­perature; and the multitemperature method, used if testing is performed at more than one temperature. In the first case, the analysis is simpler and can be performed in analytical form, but the results do not provide further insight on temperature dependence or the lower shelf behavior. The multitemperature analysis can be performed only iteratively, but it gives more comparative information on temperature dependence, scatter, and the lower shelf behavior. As for their effects on T0, both methods give statisti­cally equal confidence levels ifthe numbers ofthe valid test data are equal and the measurements have been made close enough to the final T0 (see Figure 7). Ifsufficient specimens are available, the multitempera­ture method is generally preferred since there is less risk in exceeding the limits of T0 ± 50 °C (discussed next) and the maximum KJc from eqn [14].

4.14.2.1.2 Data qualification

Data qualification is described in detail in ASTM E 1921 and is not repeated in detail here. In addition to actions associated with the test procedure and equipment, there are measures which have to be undertaken after each test or test series to ensure that only valid data will be included in the final T0 analysis. The optimum test temperature range for the T0 determination is generally selected iteratively by taking into account the already measured data. For example, three or four tests may be conducted at a selected temperature and then a preliminary T0 is determined from the data; subsequent test tempera­tures are then based on that preliminary T0.

As discussed previously, the Master Curve cleav­age fracture model is accurate only in the transition area, where stress state and cleavage crack initiation are the main controlling factors for the fracture event. The data used for T0 determination should therefore be populated in the mid-transition area rather than near the lower or upper shelf. The shape of the transition curve causes the uncertainty of the T0 determination to increase when data measured near the lower shelf are used. There can be an optimum range within the ASTM E1921 validity range of T0 ± 50 °C (Figure 9) depending upon test speci­men size. The resultant data outside of the validity range are excluded from the analysis but still can be compared to the Master Curve determined for the

valid temperature range data. The closer the test temperature is to T0, the fewer the number of test specimens that are needed. However, when testing small specimens, the maximum KJc limit is closer to the 100 MPa Vm level at T0, and the test temperature generally has to be moved to temperatures below T0 with more specimens than the minimum number needed for a valid analysis.