The Study On Multi-attributes Joint Inversion Method For Reservoir Prediction

With the development of the oil and gas exploration,the geological conditions of the oil research are more and more complex,and the reservoir research is also facing many challenges.Seismic data contains a large number of reservoir information,including the stratum structural information,as well as the information of reservoir lithological characters,physical properties and fluid,etc.It is the key of reservoir research to excavate reservoir information from seismic attributes,to recognize the seismic attributes,to optimize the seismic attributes and to carry out the seismic multi-attribute inversion.Firstly,the seismic attributes can be divided into the following categories,geometry,dynamics,inversion and mixed attributes,and can be further divided into ten categories,including structural properties,body attributes,AVO analysis,amplitude,phase,frequency,attenuation,pre-stack inversion,post-stack inversion and mixed attributes through previous research results.In this paper,the geological significance and physical meaning of seismic attributes are analyzed in a list to provide theoretical support for the optimization of seismic attributes,in addition,the extraction of seismic attributes and reservoir prediction methods are discussed in detail.We construct the seismic forward modeling according to the four factors of reservoir thickness,physical property,fluid change and thin interbed,and summarize the following rules through the analysis of the impact of reservoir parameters on the seismic response,the variation of reservoir thickness,lithology and petrophysical properties mainly affects the amplitude of seismic reflection,therefore,the properties of wave amplitude and wave impedance inversion can better identify the thickness,lithology and physical property of sandstone and mudstone reservoir;the seismicresponse is relatively weak in fluid change,comprehensive analysis of multiple techniques and a variety of materials can reduce the multiple solutions of fluid prediction;the analysis of thin interbed model shows that the existence of interbed not only affects the amplitude of seismic reflection,but also affects the frequency and phase characteristics of seismic waveform,therefore,it is necessary to combine the amplitude,frequency and phase properties to predict the interlayer.Seismic attributes can reflect different characteristics of reservoir.However,due to the complexity of the actual geological conditions,the application of a certain attribute to solve the actual geological problems is often faced with the problem of multiple representations of seismic attributes.Therefore,for the production needs of the study on reservoir thickness and interlayer,we analyze the sensitivity of the wedge model which describes the variation of reservoir thickness and the forward seismic data of the thin interbed model which reflects the development of Interlayer.Then we select a set of sensitive seismic attributes that reflect the two models,which lay a theoretical foundation for the optimization of seismic attributes and joint inversion of multi attributesIn order to solve the multi solution problem of single attribute reservoir prediction,we study the seismic multi-attribute inversion and reservoir parameter quantitative prediction,then the method and implementation of seismic multi-attribute inversion technique are described in detail,including: the construction of the objective function of seismic multi-attribute inversion,the analysis of multiple stepwise regression sensitive seismic attributes and the seismic multi-attribute inversion based on probabilistic neural network.Finally,through the practical application of M oilfield,the seismic attribute analysis of sand body,the selection of sensitive parameters of reservoir parameters and the quantitative prediction of reservoir thickness and porosity are achieved,and good application results have been obtained.