Extrapolation accuracy order

The two methods that stand out in terms of efficiency and convenience are BDF and extrapolation. Both require minimal programming effort, and can be extended to higher-order spatial derivatives. However, in the case of BDF, a limit is encountered. For the most convenient start-up methods such as the simple or the rational start, the accuracy from BDF is limited to 0(8T2). This means for one thing that one need not go beyond 3-point BDF (which is 0 8T2) in itself), but that no marked improvement can be gained from higher-order spatial derivative approximations, because there will then be a mismatch between the accuracy orders with respect to the time and spatial intervals. 

Note that (Fig. 14.11) both curves show similar accuracy at small Nt- If one measures the gradient there, it verifies the expectation of an accuracy order of 0(6T2). However, extrapolation then shows a sharp dip, followed by an approach to a constant error. The dip is due to the actual error crossing the zero line. BDF simply flattens out and approaches the same constant error. 

Note that the boundaries would have to be corrected via simple extrapolation in order to preserve the numerical accuracy wanted. The determinant of the matrix in Eq. (G.13) can be calculated by a Sturm chain and the eigenvalue is then solved for by bisection [495,1166]. 

Richardson extrapolation can be used to improve the accuracy of a method. Suppose we step forward one step At with a pth-order method. Then redo the problem, this time stepping foi ward from the same initial point, but in two steps of length Af/2, thus ending at the same point. Call the solution of the one-step calculation y and the solution of the two-step calculation yo. Then an improved solution at the new time is given by... 

Remember that the CBS models begin with a large enough SCF calculation to obtain the desired level of accuracy (see Chapter 7) therefore, no explicit extrapolation of the SCF energy is included. CBS extrapolation involves computing the second-order and infinite-order corrections to the energy. 

Owing to the necessity for extrapolating measurements on methyl acrylate to a conversion of less than 1 percent in order to avoid the pronounced autoacceleration occurring with this monomer, the data are of lower accuracy than for most of the other monomers investigated. 

The G2 extrapolation scheme - which is a prescription to extrapolate the quality of QCISD(T)/ 6-311++G(3df,2p) calculations - actually reaches chemical accuracy for the G2 test set of species, but only with an empirical correction, depending on the number of electron pairs in a molecule in order to better account for these effects. 

Before there can be any extrapolation there must be confidence in the model or rules being used. In practice this often has to involve an element of faith because of lack of validation data, particularly where the rule is empirical. The theory or model should be no more complex than is necessary to fit the data. The accuracy of fit to, for example, a creep curve can often be made more precise by applying ever higher order polynomial expressions, but outside the range of points these functions diverge rapidly to infinity (or minus infinity) leading to predictions that are ridiculous. 

The flow pattern is primarily determined by the particle Reynolds number, which is about 100 in the industrial converter (superficial velocity 0.35-0.55 Nm/s), but in order to improve the accuracy of the comparison of different catalysts, higher flow rates are also included. Although theoretical correlations can be used for extrapolating the measured pressure drops for a new shape to the industrial operation temperature, a more reliable method is to calculate the pressure drop from industrial experience for well-known shapes, e g. 10-mm ring, and assume the same relative pressure drop as in the cold measurements. 

Pioneering studies have also been published for weakly aligned protein complexes [61, 62]. Note that the ratio of bound ligand to free ligand must be optimized in order to achieve the desired accuracy when extrapolating the RDCs of the bound state. Since the protein does not interfere with the measurement of RDCs of the ligand, it is recommended to use a high concentration of the former. 

The prerequisites for high accuracy are coupled-cluster calculations with the inclusion of connected triples [e.g., CCSD(T)], either in conjunction with R12 theory or with correlation-consistent basis sets of at least quadruple-zeta quality followed by extrapolation. In addition, harmonic vibrational corrections must always be included. For small molecules, such as those contained in Table 1.11, such calculations have errors of the order of a few kJ/mol. To reduce the error below 1 kJ/mol, connected quadruples must be taken into account, together with anhar-monic vibrational and first-order relativistic corrections. In practice, the approximate treatment of connected triples in the CCSD(T) model introduces an error (relative to CCSDT) that often tends to cancel the... 

For experimental determination of diffusion coefficients, a large database is already available. Nonetheless, data for specific applications are often difficult to find because the data may not cover the right temperature range, mineral compositions, or fluid conditions. In geospeedometry applications, data often must be extrapolated to much lower temperatures and the accuracy of such extrapolation is difficult to assess. Because the timescale of geological processes is often in the order of Myr, and that of experiments is at most years, instrumental methods to measure very short profile are the key for the determination of diffusion coefficients that are applicable to geologic problems. 

---