The document is ISO-8859-1 encoded. It works when uploading it and changing the document encoding to ISO-8859-1, but what comes out, doesn't look Chinese.
I don't know how to get back to Chinese characters starting from your document - it might be not possible.
Anyway, somewhere in the process of creating your document, you mixed up encodings, with this as a result. I think it's better to use UTF-8 for Chinese.
I made an example document. I don't know any Chinese - I just got some characters from a Chinese page somewhere, so the content of my file is complete nonsense and hopefully not offending
I don't know if it is possible to figure out what went wrong and revert that. I think in this case, the only thing you can do is start over and make sure you know during the whole process of creating the file which encoding you use and make sure that that encoding supports Chinese.
ISO_8859-1 doesn't support Chinese - it is a single byte encoding, only supporting 191 characters. There have been several encoding standards for Chinese, but not this one. So something went wrong in the creation of this file: the original encoding got lost and replaced wrongly by ISO_8859-1, thereby damaging the file.
It is weird that this works on a 2.8. A lot more weird then that it doesn't work on a 3.6.