At first I have to say that it is possible from powder data to obtain structure solution. Take a look at Rietveld Method and software like:

http://www.crystalimpact.com/endeavour/Default.htm

But it is possible in general from relatively small molecules(10-20 non hydrogene atoms) and accuracy of data will be low comparison with single crystal method. And there are some restrictions like low quality of powder experimental data, hard to indexing for low simmetry and even space groups determination, overlapping of reflections, and also some impurities in sample.

Because in sample many small unoriented parts crystalline powder, diffraction from different reflections on the same theta angles (not only symmetry equivalent reflections) spreaded as circles with radius~2Tetha but non as separete spots like in single x-tal method. So you will get only set of circles instead of many diffraction spots. Moreover intensities of such circles low because diffraction power spreaded in big area but not in spots, so it is impossible to observe small intensity diffraction spots and to separate diffraction ones with close diffraction angles.

For single crystalline material, the X-rays will be scattered only in certain directions. Which will cause high intensity peaks with a narrow width (not always true other variables can play a role in this). Whereas for polycrystalline, the material consists of so many single crystals oriented in so many directions causing the X-rays to be scattered among a wide range of 2-Theta instead of high intensity narrow width peaks. 

At first I have to say that it is possible from powder data to obtain structure solution. Take a look at Rietveld Method and software like:

http://www.crystalimpact.com/endeavour/Default.htm

But it is possible in general from relatively small molecules(10-20 non hydrogene atoms) and accuracy of data will be low comparison with single crystal method. And there are some restrictions like low quality of powder experimental data, hard to indexing for low simmetry and even space groups determination, overlapping of reflections, and also some impurities in sample.

Because in sample many small unoriented parts crystalline powder, diffraction from different reflections on the same theta angles (not only symmetry equivalent reflections) spreaded as circles with radius~2Tetha but non as separete spots like in single x-tal method. So you will get only set of circles instead of many diffraction spots. Moreover intensities of such circles low because diffraction power spreaded in big area but not in spots, so it is impossible to observe small intensity diffraction spots and to separate diffraction ones with close diffraction angles.

In case of single crystalline material, the XRD shows single peak. Whereas for polycrystalline, XRD shows several many peaks through a wide range of 2-Theta. instead of high intensity narrow width peaks.

In the case of new materials the only way to identifying the exact crystal system with their space group is the single crystal X-ray diffraction analysis.

For the known materials, we can go for the powder X-ray diffraction analysis and then we can compare with the existed xrd pattern to confirm the crystal system.

In fact, with modern computers it is often possible to determine the crystal structure having only the powder diffraction data. The solution will include a lot of the trial-and-error steps, but it can be done in reasonable time if the high quality powder diffraction pattern was obtained and the sample is free from the impurities. Please, refer to the modern textbooks such as V. K. Pecharsky, P. Y. Zavalij 'Fundamentals of Powder

Diffraction and Structural Characterization of Materials' for details and examples.

Mr. Thakur,

The simplest answer to your question would be that from polycrystalline sample you have collected set of reflections belonging to atoms from different unit cells.

Please pay attention to a discussion showing briefly the structural determination steps by single crystal XRD:

https://www.researchgate.net/post/Why_do_I_get_a_the_same_XRD_spectrum_after_crystallization

In spite of the fact that there are methods applicable to segregate each set of reflections to different unit cells allowing to obtain unit cell parameters even structural prediction, nevertheless that if you have polycrystalline objects of more than one polymorphs than this additionally reduced the applicability of powder XRD.

So that generally as an absolute method single crystal XRD has its strongly highlighted advantages for structiral determination over powder XRD.

It is harder to solve the structure of a material using data collected from a powder because powder data is 1 dimensional. Data from a single crystal is 3 dimensional. In both cases you see the same reflections but in the data from the powder there is more overlap.

There are many reasons why you may be unable to solve a structure with XRD data from a polycrystalline sample. If you provide more details about what you are trying to achieve it maybe more obvious why you are having trouble.

Dear Eduard Rusanov and Alexander B. Missyul, Could you please guide me a bit more about the solution of crystal structure from powder data? Please recommend me some studies to get step by step details and software used for it.

Which method is most appropriate for residual stress analysis in additive manufactured metal components consisting material of AlSi10Mg and SS316?

Good fundamental question! There appear to several components in the original question. 

It may be better to discuss these separately :-)

It is also important to discuss the method/optics used to obtain the "XRD Data" as well as the detector to record the intersection of the Ewald surface and the reciprocal space vectors.

Pranav! Good question! Post your question as a separate discussion and post the link right here for participants to contribute. You'd need to describe the mechanical processing history of the "manufactured metal components consisting material of AlSi10Mg and SS316" along with grain size and XRD equipment available.

https://www.flickr.com/photos/85210325@N04/11604959265/in/album-72157645018820696/

Can I get atomic position from XRD data? 

Yes we can get the XRD data from atomic position but too complicated may be but not exactly....

For single crystalline material, the X-rays will be scattered only in certain directions. Which will cause high intensity peaks with a narrow width (not always true other variables can play a role in this). Whereas for polycrystalline, the material consists of so many single crystals oriented in so many directions causing the X-rays to be scattered among a wide range of 2-Theta instead of high intensity narrow width peaks.

In summary, Single crystal shows intense and sharp narrower peaks in XRD pattern but polycrystalline materials shows less intense and broad XRD peaks.

it is possible in general from relatively small molecules(10-20 non hydrogene atoms) and accuracy of data will be low comparison with single crystal method. And there are some restrictions like low quality of powder experimental data, hard to indexing for low simmetry and even space groups determination, overlapping of reflections, and also some impurities in sample.

Because in sample many small unoriented parts crystalline powder, diffraction from different reflections on the same theta angles (not only symmetry equivalent reflections) spreaded as circles with radius~2Tetha but non as separete spots like in single x-tal method. So you will get only set of circles instead of many diffraction spots. Moreover intensities of such circles low because diffraction power spreaded in big area but not in spots, so it is impossible to observe small intensity diffraction spots and to separate diffraction ones with close diffraction angles.