Investigation of error estimation method of observational data and comparison method between numerical and observational results toward V&V of seismic simulation

Suzuki, Yoshio; Kawakami, Yoshiaki*; Nakajima, Norihiro

Mechanical Engineering Reviews (Internet), 4(1), p.15-00525_1 - 15-00525_18, 2017/01

https://doi.org/10.1299/mer.15-00525

The method to estimate errors included in observational data and the method to compare numerical results with observational results are investigated toward the verification and validation (V&V) of a seismic simulation. For the method to estimate errors, it is found that errors are caused by the resolution, the linearity, the temperature coefficient for sensitivity, the temperature coefficient for zero shift, the transverse sensitivity, the seismometer property, the aliasing, and so on. Thus, it is needed to estimate errors individually for those factors and integrate them. For the method to compare numerical results with observational results, it is found that six methods have been mainly proposed in existing researches. Evaluating those methods using nine items, advantages and disadvantages for those methods are arranged. The method is not well established so that it is necessary to employ those methods by compensating disadvantages and/or to search for a solution to a novel method.