Imagine a blunt object like a space-capsule entering the atmosphere. It experiences a decelerating force, right? If you divide this force by the surface area of the blunt front facing surface, we get an effective pressure. The atmosphere has to create this pressure in front of the object, otherwise there would be no force (a molecule streaming by the sides of the object without hitting it can't create such a force). This compression also heats the gas in front of the object.
The hot, dense gas now streams along the sides of the capsule. If we want to keep the capsule cool, then we certainly don't want this hot gas to touch the body again, which is why capsules are entering with the broad side and are not flying like planes with a sharp nose cone. The body angle has to be small enough that the gas can pass the entire body before it expands enough to reach the walls.
You can see these effects nicely in old NASA images showing the supersonic bow shock around models of their capsules, the dark areas are dense gas under high pressure, the light ones are less dense, low pressure regions:
The majority of the kinetic energy in the capsule will be converted to heating of gas in this bow shock, only a fraction of it will be absorbed by heat shield and an ever smaller amount will heat the backside walls. Without this phenomenon re-reentry would be an even harder thermal problem than it already is.
Comets do orbit in ellipses, but there is no requirement for their orbital planes to match the Earth's exactly. Short-period comets have planes that are relatively parallel to the Earth's (with a good degree of lee-way), and long-period comets can come in from any direction and with any orbital inclination.
Here is a good representation of the orbit of comet 109P/Swift-Tuttle:
When the comet's orbit crosses the orbital plane of the Earth again, it is way out past Jupiter, and this means that there is just the single meteor shower from this debris track.
Also, it's worthwhile noting that while ellipses in general can intersect four times, keplerian orbits share a focus, which means that they can intersect at most twice (which they do when they're coplanar and the sizes are right).
Best Answer
Meteoroids come in a very large range of sizes, from specks of dust to many-kilometer-wide boulders. Explosions like that of the Chelyabinsk meteor are only found meteors that are larger than a few meters in size but smaller than a kilometer.
Though the details are argued endlessly by those who study such phenomena (it is very hard to get good data when you don't know when/where the next meteor will occur), the following qualitative description gets much of the important ideas across.
The basic idea is that the enormous entry velocity into the atmosphere (on the order of $15\ \mathrm{km/s}$) places the object under quite a lot of stress. The headwind places a very large pressure in front of it, with comparatively little pressure behind or to the sides. If the pressure builds up too much, the meteor will fragment, with pieces distributing themselves laterally. This is known as the "pancake effect."
As a result, the collection of smaller pieces has a larger front-facing surface area, causing even more stresses to build up. In very short order, a runaway fragmentation cascade disintegrates the meteor, depositing much of its kinetic energy into the air all at once.
This is discussed in [1] in relation to the Tunguska event. That paper also gives some important equations governing this process. In particular, the drag force has magnitude $$ F_\mathrm{drag} = \frac{1}{2} C_\mathrm{D} \rho_\mathrm{air} A v^2, $$ where $C_\mathrm{D} \sim 1$ is the geometric drag coefficient, $\rho_\mathrm{air}$ is the density of air, $A$ is the meteor's cross-sectional area, and $v$ is its velocity. Also, the change in mass due to ablation is $$ \dot{m}_\text{ablation} = -\frac{1}{2Q} C_\mathrm{H} \rho_\mathrm{air} A v^3, $$ where $Q$ is the heat of ablation (similar to the heat of vaporization) of the material and $C_\mathrm{H}$ is the heat transfer coefficient. Since the mass-loss rate scales as $A \sim m^{2/3}$, sublinearly with mass, smaller objects will entirely ablate faster, setting a lower limit on the size of a meteor that can undergo catastrophic fragmentation before being calmly ablated.
Meteors that are too big, on the other hand, will cross the depth of the atmosphere and crash into the ground before a pressure wave (traveling at the speed of sound in the solid) can even get from the front to the back of the object. There simply isn't time for pressure-induced fragmentation of the entire object to occur, meaning the kinetic energy isn't dissipated until the entire body slams into Earth.
[1] Chyba et al. 1993. "The 1908 Tunguska explosion: atmospheric disruption of a stony asteroid." (link, PDF)