You can't just increment h and s!
H is the hash of the message and is not predictable let alone be incrementable and s is computed using k, h, r, d so again it is also no incrementable.
The article also makes little sense to me, since you can't have the same k but different r values since r is computed using k (multiplied by G) meaning rf should be equal to r already otherwise it is computed from a different k value which means the equations used there are all wrong.
The article is about fault injection, where a faulty signature (in this case, made with a r value that's different from what it should be) and a valid signature is used to extract the private key. As OP have finally noticed, you can't create faulty signatures without the private key, so this attack means nothing if you already only have a single, valid, signature.
OP: Any kind of manipulation of an existing signature will create two linearly dependent signatures, which in practice is just the same as having a single signature. You can't use the info in the article to get a private key out of a single good signature in any way.
