Maybe there is a bug in the programs

gwbearthqukae commented 4 years ago

Thanks for your code. It's very helpful, but I found a bug recently. When I computed the phase velocity of rayleigh wave with these parameters

thickness=np.array([0.3,0]
vp=np.array([2.6, 5.29])
vs=np.array([1.12, 3.14])
rho=np.array([2.12,2.58])
periods=1/np.arange(6,1,-1)

surf96 threw an error. But If I make the thickness=np.array([0.301,0]), or even [0.3001, 0], it runs sucessfully! Why? This problem confused me . Could you help me find the answer?

miili commented 4 years ago

Hi, I cannot reproduce the error. Can you please paste the output?

gwbearthqukae commented 4 years ago

I'm doing the Monte Carlo inversion, and try to calculate thousands of models. I found that the thickness=[0.25,0] also produced the error, "surf96 threw an error! This may be due to low velocity zone causing reverse phase velocity dispersion and mode jumping. Due to looking for Love waves in a halfspace which is OK if there are Rayleigh data" . But thickness=[0.251,0] is OK.

I run the programms on a laptop with ubuntu system, and python3.7, gfortran.

I have put the error information in the attachment.

发件人: Mi! notifications@github.com 发送时间: 2020年1月20日 20:44 收件人: miili/pysurf96 pysurf96@noreply.github.com 抄送: gwbearthqukae gwb4133@hotmail.com; Author author@noreply.github.com 主题: Re: [miili/pysurf96] Maybe there is a bug in the programs (#2)

Hi, I cannot reproduce the error. Can you please paste the output?

― You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/miili/pysurf96/issues/2?email_source=notifications&email_token=AF5B6O2RLOQMTCAOWBVCJG3Q6YELZA5CNFSM4KI676Q2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEJNYYSI#issuecomment-576425033, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AF5B6O2SJAYQMU2QKRKC5ATQ6YELZANCNFSM4KI676QQ.

miili commented 4 years ago

This is due to wrong input data.

miili / pysurf96

Maybe there is a bug in the programs #2