Multi-lever Fuzzy Comprehensive Evaluation Of Reservoir Induced Seismic Risk Based On Association Rules And Cloud Model

As one of the important hydraulic projects, huge artificial water reservoirs are builtfor generation of hydroelectric power, flood control, water supply and irrigation purposes.Water conservancy projects make life more convenient and bring economic benefits, whilethey may also cause terrible damage to the ecology and environment. For one thing, sincereservoirs always cover a large geographical area, the construction may influence thetopographic features, the hydrogeology and the environment geology of reservoir and itssurrounding region; for another, since water flows in fractured rock may change the stressof rocks, seismogenic environment can be affected by reservoir impounding.Reservoir-induced seismicity is indeed one of the most serious hazards in waterconservancy projects. Analysis of the possible seismic risks of a reservoir and itssurrounding region is one of the major considerations in geological analysis andenvironmental impact assessments of hydraulic projects.Previous researches on reservoir-induced seismicity focus on three major aspects:The initial work focus on analyzing and comparising reservoir-induced seismicityexamples; on the basis of these studies, the cause mechanism and physical mechanism arediscussed; Since1990s, many workshops and symposia have been organized onreservoir-induced seismicity, and the development of this topic has transferred toquantitative prediction and evaluation. However there is no agreement on the causemechanism of reservoir-induced seismicity. Uncertainty exists both in the model structureand in the parameters, and it is thus not an easy work to evaluate the risks inreservoir-induced seismicity. Most of previous researches consider the environmentalfactors as unchangeable factors, but ignore the uncertainty of the seismogenicenvironment in reservoir-induced seismicity. The seismogenic environment ofreservoir-induced seismicity is a comprehensive and fuzzy system with multi-lever,multi-criteria and multi-objects. As an extension of previous researches, an uncertaintyassessment model based on the multi-hierarchy risk index system of reservoir-inducedseismicity is proposed, and take Yangtze Three Gorges Reservoir as an sample to illustratethis model. The main contents in this dissertation are as follows:(1) Select Xinfengjiang reservoir, Geheyan reservoir and Koyna reservoir asexamples, and analyse reservoir-induced seismicity causing mechanism and seismicactivities in details. According to the cause mechanism proposed by previous research, thisthesis summarizes the environmental factors which may cause reservoir-induced seismicity using quantitative analysis, which are the basis of hierarchy assessment systemof reservoir-induced seismicity.(2) Due to the complexity of the hydrogeology and environmental geology in thelarge submerged area of reservoir, base on the risk identification in reservoir-inducedseisimicity combine environmental factors, association rules are developed to illustrate therelationship between environmental factor and possible earthquake magnitude and thenthis thesis analyse the importance of environmental factor using quantitative method. Thenthe hierarchy assessment system of reservoir-induced seismicity based on association rulesis constructed, which supports the comparison among environmental factors.(3) Due to the uncertainty of the seismogenic environment of reservoir-inducedseismicity, this thesis considers the combination of envirionmental factors inreservoir-induced seisimicity as a fuzzy system and carries out weight calculation andfuzzy mapping from low layer to high layer. A fuzzy comprehensive evaluation modelbased on the multi-hierarchy risk index system of reservoir-induced seismicity is proposed,and then the advantages and disadvantages of this method are discussed.(4) Considering that the-decision of important scales and membership degree is ofteninfluenced by subjectivity, this thesis proposes a cloud improved multi-hierarchy fuzzycomprehensive evaluation model to assess the reservoir-induced seismicity. With thenumerical characteristic index described by expectation value (Ex), entropy (En) andhyper entropy (He), this model includes not only the medium value of the concept, butalso the fuzziness and the discreteness, causing the results of reservoir-induced seismicityassessment more objective and more accurate.The results in this thesis help to illustrate the objective law of seismogenicenvironment and seismic activities, and this thesis is also an extension of previousresearch in the prediction and assessment method of reservoir-induced seismicity.Moreover, this thesis contributes to improve the technology of disaster assessment inhydraulic engineering, and provides a reliable evidence of risk assessment inreservoir-induced seismicity.