Dark matter doesn't interact electrostatically but I don't understand why it does not collide with ordinary matter . Also can a chunk of dark matter interact with another chunk of dark matter if so how? .Also are there any established experimental protocols for detection of dark matter in a direct manner , that is not by unseen mass present in a galaxy.
In one sense you raise the question of whether a type of matter or elementary particle can exist without interacting with known matter or particles. Of course gravitation is a form of such interaction, but if that is the only way dark matter (DM) interacts with known matter then it is in a sense "forever dark."
The attempt to detect DM by it collision with Xenon atoms is outlined by EasyPeasy. The assumption has been that DM is most likely interacts by the weak nuclear force, and a condensate of the super-partners of the photon, Higgs particle and the neutral current Z particle called the neutralino is a candidate for DM. If so then a neutralino, or any weak interaction DM particle, a WIMP, would interact with an atom. This would then translate some of the tranverse momentum or energy into another form. The Cryrogenic Dark Matter Search (CDMS) attempts to find such an interaction with a cold crystal. A DM particle will then in principle induce a phonon or quanta of vibration on the lattice.
So far these searches have come up empty. If such searches as well as the LUX-ZEPLIN, a scaled up version of ZEPLIN that has come up null in the search, fail to find anything then we may have to abandon the WIMP hypothesis. The null data so far from the LHC for low mass or low energy supersymmetry means the neutralino (as well as s-quarks etc) may not exist at multi-TeV energy.
The alternative is the axion, which is found in the supergravity multiplet, and is also the carrier of the CP violation for QCD. This has some interesting connections between gravity and QCD in a supersymmetric setting at much higher energy. The axion is a scalar particle that obeys the wave equation $$ (\square~+~m^2)\phi~=~\kappa\vec E\cdot\vec B $$ where this inhomogeneous term term $\kappa\vec E\cdot\vec B$ means the axion can be exchanged with photons. There are searches for this. An axion in the presence of a strong magnetic field can be converted to a photon. The Axion Dark Matter Experiment (ADMX) is an attempt to find such conversions. This will be very difficult for the axion mass that is a measure of the CP violation is expected to be very small.
An interesting paper came out recently on an anomalous decay of $^8Be_4$ that suggests an $17MeV$ particle may be produced. This is argued to be a new type of gauge particle that mediates interactions primarily in the dark matter sector. This may or may not turn out to be how nature works. It does have some connections with axions, and it also makes the point that dark matter might be a whole "zoo" of various types of quantum fields or particles.
More than 99% of the atoms mass is in a very small part of it. Dark matter does collide with ordinary matter just not with the whole part of the atom. Dark matter is believed to be passing through everything, including us. As it passes through us and other materials sometimes, but very rarely a bit of the dark matter will hit the nucleus of an atom. This SHOULD heat up the atom an almost unnoticeable amount.
Also are there any established experimental protocols for detection of dark matter in a direct manner
Yes currently in a super cooled environment in deep underground labs scientists keep extremely dense packed materials at around .4 Kelvin. The have instruments that can detect almost any temperature change at this low temperature and they're waiting for some dark matter to collide with a nucleus. The reason they're so deep underground (in an abandoned mine) is that they need to filter out other subatomic particles that may interfere. So far they have met no success.
Currently there is a Prototype of LUX-ZEPLIN (large underground Xenon). It will begin testing for dark matter in 2020. It's a 5 foot cylinder filled with super cooled liquid gas. When dark matter hits the nucleus of a xenon atom it will release a flash of light and some electrons. It will hold 10 tuns of xenon and will have a magnetic field pulling electrons to the top of it. This is the most complicated dark matter test ever done. Currently LUX-ZEPLIN can't run stabley but scientists are working on it. More info -> Prototype of LUX-ZEPLIN
The title question
Why doesn't dark matter interact with ordinary matter?
Could be taken a couple of different ways.
How does it work?
One way would be equivalent to asking "What's the mechanism?", but the answer is a bit unsatisfying. It comes down to, "Well, it just doesn't participate in the electromagnetic or strong forces. It does participate in gravitation and we don't yet know about the weak interaction." And the reason for that is unknown. It just is.
How do we know?
Or we could take it as "How do we know that?", to which we have a much more complete answer, that can none-the-less be summarized in one sentence "Because if it did interact with ordinary matter we'd have found it by now."
We've spent decades thinking up ways to explain first the galactic rotation curves and then the velocity distributions of galactic clusters with only kinds of matter that we are able to produce or observe in the universe, and then we have looked very hard for signs of those kinds of matter that we have missed. Very often we have found a little bit of stuff with those searches, but it doesn't add up to enough.
For example it's been considered that a lot of small, dense, cold object floating among the star (MACHOs for MAssive Compact Halo Objects) might do the trick. So in the 1990's a large scale effort to detect them by means of gravitational microlensing of stars in globular clusters and nearby galaxies was launched. And they found MACHOs. Enough MACHOs to account for as much as 3% of the matter needed to explain the rotation curve of the Milky Way.
Just for completeness I'll mention that people have also asked if maybe the law of gravitation that we've been using (general relativity, but it reduces to Newton's Universal Gravitation with retardation at large distances) needed to be changed. For a while this effort when by the name MOdified Newtonian Dynamics (MOND), and reputable scientists pubnlished papers on the idea in good journals. But in the end it was mostly abandoned, killed by esoteric mathematical difficulties and the trouble it has with a few bit of observational data like the bullet cluster.