You can only access things via HTTP that are served by an HTTP (e.g., "web") server which recognizes and handles the URL you give it. It would be possible to configure an HTTP server to allow you to access any file by full path, but it would almost certainly be an extraordinarily bad idea to make **all* files on a machine accessible via HTTP.
Your question contradicts itself by saying both that the file isn't in a web site and that you can access the file via http://5.135.140.184:8080/tpi/documents/D773.JPG. If you can access the file via that URL, then an HTTP server at 5.135.140.184:8080 is running and handles that URL by giving you the file. That's what you should stick with.
In a comment you've added
I want load an other file that is not in the web sites.
You can't, not via HTTP. HTTP requires a web server that understands the URL. If your web server at 5.135.140.184:8080 isn't configured to deliver the file you want, you cannot access it via HTTP (without changing the server configuration).
http://5.135.140.184:8080/tpi/documents/D773.JPG, which means it is in a web site. Which is it?var/lib/tomcat/webapps/tpi/documentswould not be accessible through this exact path by URL, but bytpi/documents). There can be redirecting rules that can affect an URL destination too