Front Fan Design – P&W’s Problem

The RPM numbers already given are enough to show the difficulty of the design problem. To compete with the CJ805-23, P&W wanted the same frontal area, and hence the same fan diameter, namely 53 inches. But their low-pressure design speed was 6500 rpm, 13 percent greater than GE’s fan RPM. This meant that P&W needed to live with a tip-speed over 1430 ft/sec, not only far above GE’s 1260 ft/sec, but also well above the 1400 ft/sec of NACA’s most successful supersonic design. The implied tip Mach number was totally out of the range of P&W’s compressor design technology. P&W had hired none of the central figures from NACA’s supersonic compressor research program. They had access to the NACA reports, but they had done nothing significant toward pursuing very high Mach number stage designs in house. Their compressor design technology was built around a huge data-base of two-dimensional airfoil performance, which they had developed through their own cascade wind-tunnel testing. They had extended this data-base to progressively higher increments in Mach number through testing double-circular-arc blade profiles. This had given them an empirical base for designing stages with pressure – ratios in the 1.25 range, but they had yet to utilize this capability in an engine, and it offered them no basis for designing highly loaded blades with tip Mach numbers above 1.2.

Exacerbating the problem was the fact that P&W had not developed any streamline-curvature computer programs. Their axial compressor computer program had been based on the non-iterative streamtube method that they had employed, in hand calculations, in designing the two spools of the J-57 compressor in the late 1940s. The great virtue of this computer program was that it incorporated their airfoil performance data-base, allowing them not only to identify preferred airfoils in the design phase, but also to predict compressor performance at off-design conditions. Radial equilibrium effects were taken into account in a rough way in the program by transferring flow radially from one geometrically pre­specified streamtube to the next in between blade rows. This approach had proved adequate for their designs in large part because neither of the spools of their compressors contained a great many stages, and hence the cumulative effects imposed by radial equilibrium were not that severe in the back stages of their spools. This streamtube method, however, offered no way for tailoring arbitrary blade contours in the way GE had. P&W was going to have to employ pre-defined airfoils in their fan.