Port passage taper vs. no taper Actually, this subject could morph into a complete story, but not here. We'll limit comments to the fact that both can have merits, under the proper circumstances. On the notion of "no taper," however, we'll offer one thought for you to at least consider because it's linked to the mean flow velocity issue.
Even though there is considerable "pressure excursion" activity along intake and exhaust paths, you can generalize these conditions as being akin to kinetic energy systems. As such, their kinetic energy content may be mathematically described as one half the product of mass times velocity squared (1/2mv^2). Since we suggested mean flow velocity at peak torque (v.e.) to be on the order of 240-260 feet/second, would you not want as much flow mass as possible moving at this critical velocity when it is reached? If there is taper in a passage (and velocity is a function of section area), less flow mass will be moving at this velocity when its associated rpm is reached. Multiple "degrees of freedom" intake and exhaust systems (those tuned to multiple rpm points) frequently contemplate the use of no-taper passages. It's something to think about.
What happens when flow direction changes? Here again, we're attempting to simplify a rather complex issue. Fundamentally, under any given conditions of flow in a passage, there will be a "pressure distribution" profile that represents the distribution of kinetic energy in the passage. While some amount of fluid friction at the interface between the "working fluid" and passage walls is unavoidable, the effects flow-direction changes have on pressure distribution across any given cross-section along the path can be more problematic.
Dissimilarities in mass (weight) between air and fuel exacerbate the problem, keeping fuel suspended in air (wet flow), and the effects of gas dynamics (dry flow) can combine to make passage design for directional changes a real challenge.
Consider some of the conditions of flow during a 90-degree turn (elbow shape), using a constant circular section area (see Figure 1). Note that the "short path" in this instance is along the passage's floor and, in fact, amounts to a single line of virtually no width. In addition, because of this feature, you can see how the pressure distribution "profile" is skewed to favor this shortest path. As such, the highest flow velocity is toward and along the passage floor. Neither the surface nor shape of the "outside" surface plays much of a role.
Were we to flatten the floor (Figure 2; still maintaining the same section area) to provide some additional "short paths" that are equal to the one in Figure 1, the pressure profile becomes less concentrated along a single line, essentially making the passage wider and more efficient, as compared with the one in Figure 1. The result is a time-honored D-shape sometimes found in exhaust ports. This D-shape concept can be applied to many flow passage requirements where the short-side can (1) be designed accordingly or (2) modified to produce the same results.
In both examples (Figures 1 and 2), you can see that there is a lack of flow activity on the "long side" wall, thereby causing some pressure distortion in the pressure distribution profile. One feature that can improve flow efficiency is to create a shape that provides a more uniform distribution of pressure across each section in a passage, including times when the direction of flow changes. Configuring section areas that work toward uniform pressure distribution is one possible solution. This leads us to the next topic for discussion: trapezoidal section shapes.