Seismic Design of Industrial Facilities

An industrial facility consists of many integrated components and processes. As such, operation of a facility depends upon the performance of its critical components. The greatest risk from an earthquake is that to life safety. However, in large earthquakes, industrial buildings and related machinery and equipment damaged may be costly to repair and there may be additional damage from fire and chemical spills. As such, the design (or seismic retrofit) of industrial facilities should preferably be based on performance-based methodologies with the objective of controlling structural and non-structural damage and the triggered technological disasters. In this paper industrial damages and losses that took place in past important earthquakes, especially in the 1999 Kocaeli earthquake, will be summarized. A general description of industrial-sector and component based earthquake performance and vulnerabilities will be provided.

The wide range of induced effects of earthquakes, from direct damage due to seismic shaking to indirect damage caused by secondary effects (e.g. liquefaction, soil densification and landslides) makes the seismic risk one of the most common cause of structural failures among natural hazards. The degree of vulnerability and the level of exposure of the threatened elements may further amplify such effects. In this sense, the seismic risk induced by an oil-gas storage plant located close to an important commercial harbour in Southern Italy is analyzed. The plant is situated in one of the areas with the highest levels of seismic hazard in Italy, hit in the past by earthquakes as large as 7 in magnitude. Moreover, the plant lies near to the shoreline and the facing seafloor is characterized by the presence of a deep submarine canyon filled by loose, unconsolidated soils coming from the excavation of the harbour channel. Given these conditions the following phenomena have been investigated: local site amplification, liquefaction, submarine landslides and sea-waves run-up. The stability analyses considered both the plant’s structure itself and the site. A vulnerability analysis provided the response to the ground motions of the steel tanks forming the structure, while dynamic analyses gave the response of the soils to the wide range of possible ground failures. Joining all the possible effects that could destabilize the plant, an overall probability that the safety of the plant may be affected was computed. The total risk was then assessed considering the effects, in terms of human life losses, produced by the failure of the plant. This risk was then compared with those deriving from other human activities to provide a reasonable basis for risk the acceptability assessment.

Seismic design loads for standard buildings are given in seismic building codes. Code response spectra are obtained from generalised spectra for different soil classes and reference hazard parameters, like peak ground acceleration, in order to scale the spectrum according to the hazard at the site (e.g. using earthquake zones). For sites of special facilities and constructions that are designed for longer return periods than standard buildings, a site-specific hazard assessment leads to more realistic seismic loads than code spectra scaled by importance factors. The article presents general methodologies, procedures and approaches for a site-specific seismic hazard assessment, taking into account local soil properties.

A generic seismic risk study for critical industrial facilities (CIFs) is presented and discussed in detail. The study is focussed on the residual seismic risk of critical facilities supposed to be correctly designed according to Eurocode (EC) 8. The initial objective was to define a design importance factor γ

I

in order to achieve sufficiently low probabilities of a major accident. The residual seismic risk is dominated by earthquakes for which the probability of occurrence is typically one or two orders of magnitude lower than for the design earthquake. According to Swiss practice, the annual probability of a major accident with more than 100 fatalities outside the industrial facility must not exceed 10-7. In order to achieve this goal, it turned out that a design earthquake with a return period of the order of 100’000 years should be considered, with an associated importance factor around 8! Such a design, however, would be technically and economically unfeasible. Therefore, it is necessary to adopt a risk based view and first explore all possibilities of reducing the largest possible number of fatalities – by other means than just a strong seismic design. At present, it is not yet clear what will be done by the Swiss authorities once all reasonably practicable measures of reducing the size of the largest possible accidents have been put into action and the residual seismic risk is still too high. In any case, however, it is strongly recommended to also look at what could happen if GMs much above design GM occur, instead of simply design for a given GM level.

This paper presents the French regulations for seismic protection of critical industrial facilities. After an overview of the seismic regulations newly enforced to implement the Eurocodes in France, emphasis is put on the scope of the recently published bylaws governing the seismic protection of such installations: scope, definition of the seismic hazard, schedule of implementation. To conclude the guidelines under preparation, defining the technical rules for each type of equipment, are introduced.

Industrial facilities are typically complex systems consisting of a primary loadcarrying structure with multiple technical installations like tanks, vessels and pipes, which are generally designated as secondary structures. Due to high cost of the process engineering components and due to the risk of business interruption and the release of harmful substances into air, water and ground if damages occur, industrial facilities must be designed to safely withstand seismic loading. The design must consider both the primary structure and the secondary structures as well as the dynamic interaction effects between structural and non-structural components. However, in Germany a basis for the seismic design of such facilities is still missing, since the current earthquake code DIN 4149 and the forthcoming code DIN EN 1998-1 are limited to conventional buildings. For this reason a technical guideline for the seismic design of industrial facilities was developed in collaboration with the German Chemical Industry Association (VCI) to close the gap of the design standards. The present paper introduces the guideline with special emphasis on plant specific aspects.

First of all the paper describes the Italian regulatory framework for precast buildings. Then the work focuses on the structural weaknesses most frequently found in existing buildings. It also discusses the changes made to building standards and to the technical specifications following the earthquake that struck the regions of Emilia-Romagna, Veneto and Lombardy in May 2012. Finally, it presents the guidelines developed by the Working Group on the Seismic Conformity of Industrial Buildings for the rapid restoration of accessibility and seismic improvement of existing precast buildings.

Due to the introduction of the revised German Earthquake design standard DIN 4149 in 2005 [1] including a re-evaluation of earthquake loadings and the forthcoming introduction of the European Earthquake standard DIN EN 1998 (Eurocode 8) [2] the demands on the overall Earthquake structural design increases. As plant operators of Chemical production units governed by the Major Accidents Ordinance (Störfallverordnung) are obligated to operate their facilities in accordance to the latest state of the art safety standards the existing production facilities will need to be evaluated in regards to earthquake resistance. The Evaluation is based on the VCI-guideline [3] which provides in addition to DIN 4149 and DIN EN 1998 a basis of an Earthquake assessment and design principles for chemical production facilities due to Earthquake loading. This paper introduces the Earthquake assessment program of existing chemical production facilities that BASF-SE has undertaken in the past years on their production site in Ludwigshafen, Germany. Based on specific examples the assessment procedure for the initial evaluation of existing Chemical production facilities is presented. Furthermore experiences and results of already finalized assessments of more than 28 production units are summarized and recommendations are derived for further assessments.

This paper presents the application of probabilistic methods for the seismic analysis of existing industrial facilities. First, the main advantages and the rationale for probabilistic (versus deterministic) approaches are discussed for existing structures/facilities. A short overview of existing probabilistic and deterministic seismic analysis approaches follows. Afterwards a simple and efficient probabilistic approach is presented with an example as application on existing industrial facilities. The method involves state-of-the-art probabilistic seismic hazard analysis (PSHA). It covers the whole industrial facility including structures, components, mechanical installations, piping, tanks, etc. In comparison to Monte Carlo Simulation, this method is cost-effective and practical and can be used for risk-informed/performance-based rehabilitation or strengthening.

Developments of modern science and technology have greatly enhanced the ability of engineering community in understanding the phenomena, mechanism and performance of engineering structures and systems. Meanwhile, the defect and inadequate of deterministic methodologies in modelling and analysis of engineering systems also exposed the importance of uncertainty analysis. As a matter of fact, it is recognized more and more clearly that the randomness propagation in a physical system plays an important role in understanding and controlling many phenomena and behaviours of engineering structures and systems, particularly those emerging in nonlinear mechanics and systems. On the basis of the principle of probability preservation and its random event description, a new kind of general probability density evolution equation (GPDEE) is introduced which could capture the randomness propagation in a dynamic system. Then this kind of equation is extended to general physical systems and therefore reveals the secret of randomness propagation in a physical system. Some recent developments using GPDEE are summarized, including: (1) the physical random models for dynamic excitations, especially taking seismic ground motion as an example; (2) the multi-scale stochastic damage model for concrete materials and structures; (3) a physical approach to the global reliability of structures, respectively. Besides, some typical engineering applications are illustrated as well.

Seismic analysis is one of the main steps in the structural design of Nuclear Power Plants (NPPs). Design is usually made by assuming linear structural behavior and using the so-called modal spectrum analysis. This method is based on the calculation of the response peaks for each earthquake direction (X, Y or Z) of several single-degree-of-freedom oscillators representing the modes of the analyzed structure. Then, the modal peaks of each response parameter for each earthquake direction are combined using, for instance, the so-called Complete Quadratic Combination-CQC (Der Kiureghian [1]). The superposed responses are, by definition, positive quantities. Hence, their sign must be defined, according to a fundamental mode shape or another reference structural configuration. Actually, signature of CQC of modal responses is not required for the approach based on the notion of “peak modal response hyper-ellipsoid envelope” (also called “envelope for seismic response vectors” by Menun and Der Kiureghian [2]). This is one of the interesting advantages of this method. For this reason, in this paper we discuss two developments based on the notion of “response envelopes”. The first one is an “equivalent static method” (ASCE [3], Nguyen et al. [4]-[5]) based on the theory of the “response envelopes”. The second development is an improved procedure for the definition of the signs of the CQC of modal peaks. Some of these proposed methods are applied to a NPP building and results are then compared with those coming from a standard modal spectrum analysis.

An industrial heavy structure subjected to seismic action and its response after a few design improvements is presented. The difficulties of modelling of this industrial structure compared with ordinary structures is discussed, especially the effect of free hanged 250 tons of steel pipes used for medium cooling. The effect of long hanger (about 15m) swaying and its possible bouncing on steel casing should be minimalised. A detailed FEM model was prepared. Seismic effects were calculated via time history analyses. Five different alternatives of design improvements were taken into account. They differ by constructing difficulties and costs needed for achieving the desired effects. An introduction of seismic stoppers and dampers is considered too. Gap closing effects and contact forces calculation between different parts of the relatively moving structure are introduced too. The advantages of the best solution are discussed. The ratio of reduction seismic effects with and without appropriate measures is compared.

The International Fusion Reactor – ITER – is being designed and constructed with a high level of safety as an essential requirement. In order to meet the safety and performance objectives of the French regulatory authorities and of the ITER Organization requirements, the Tokomak Complex has been isolated from the potentially highly damaging effects of the hazard seismic loading by employing seismic isolation bearings. The Tokamak Complex seismic base isolation system and the Tokamak Complex structure have been designed by EGIS Industries as a member of the Architect-Engineer team ENGAGE. The design, manufacturing, qualification and installation of the seismic isolation bearings have been carried out by NUVIA Travaux Spéciaux.

The present paper shall give some ideas to protect power plant machinery against seismic demands. The elastic support of turbine foundations, fans, boiler feed pumps and coal mills is a well-accepted strategy for the dynamic uncoupling from their substructures and for the vibration isolation. If the corresponding bearing systems are combined with certain strategies an efficient earthquake protection for the important machinery can be achieved. Seismic control may be obtained by increasing the fundamental period or increasing the damping or changing the shape of the fundamental mode of a structure. A combination of these measures could lead to an optimum seismic protection system as described in this contribution. Here, the first step consists of the choice of the required stiffness properties of the flexible support. Helical steel springs possess the possibility of providing a three-dimensional flexibility. Thus, it is possible to obtain a vertically and horizontally acting protection system. Depending on the seismic input the spring properties could be chosen in a specific range. The system frequency can be decreased and simultaneously, the damping ratio can be increased by incorporating viscous dampers at different locations of the spring supported structure. Internal stresses of important members, acceleration amplification as well as deformations due to seismic excitation can be decreased compared to a structure without any precautions. The possible damage after a severe earthquake can be reduced significantly, and the behaviour of the structural members could remain in the elastic range. Details of executed projects and corresponding results of numerical analyses document the effectiveness of the presented seismic protection strategies. Selected pictures demonstrate the general applicability of the applied systems.

The MARMOT seismic monitoring and trip system perfectly responds to the increasing safety demand in vulnerable industries such as Nuclear Power Plants (NPP), Nuclear Storage Facilities, Liquid Natural Gas Storage (LNG), Refineries and many more. The system measures and analyses systematically tremors that occur at different locations in a facility and quickly recognizes dangerous patterns. With its distributed intelligence it guarantees dependable alarms for automatic shutdown (trip) information impacted by earthquakes on the structures. MARMOT complies with all relevant standards (e.g. IEC 61508, IEC 60780, and IEC 60880) applicable in these industries, fully tested and certified by the “TÜV-Rheinland” organization. This paper presents requirements and the corresponding MARMOT solution regarding seismic monitoring for industrial facilities.

The French regulation has been updated in 2010, and now explicitly requires that equipment of high hazard industrial facilities (outside nuclear field) do not lead to unacceptable consequences under the highest earthquake of the seismic zone where the facility is located. As well the seismic zones have been re-evaluated, considering four levels for the French metropolitan territory. To meet this new requirement AFPS has been asked to draft a guide that defines a strategy to stop the facility on detection of the earthquake. This specific guide is part of a set of documents that will support operators in the different design requirements that could be implemented to demonstrate compliance with the regulation. The guide explains how automatic or manual actuators could isolate the dangerous inventory inside the facility to prevent or limit the impact. As well mitigation of indirect effects is considered. Earthquake phenomenology, threshold to trigger the safe shutdown, principle to demonstrate compliance with the regulation, logic, hardware & software requirements, qualification and in service inspection are described together with real case study. The present paper focuses on earthquake phenomenology, detection strategies and threshold to trigger the safe shutdown. As far as the threshold level is concerned a possible -very low- default threshold that would prevent long diagnostic analysis of the weaknesses of the equipment will be discussed.

In order to research the seismic behaviour and effective vibration control strategy for the wind turbine tower and electrical transmission tower, shake table tests on reduced-scale wind turbine tower model and electrical transmission tower model were carried out at the State Key Laboratory of Disaster Reduction in Civil Engineering, Tongji University. Tuned Mass Damper (TMD) systems were applied for reducing the seismic responses of the model towers. Experimental results of both model tests are presented in this paper. The test results indicate that the TMD systems are remarkable in seismic responses reduction for the wind turbine tower and electrical transmission tower, and can be widely used for engineering application.

The basic concepts and some representative results of the work carried out within the European collaborative research project SYNER-G (http://www.syner-g.eu) are presented in this paper. The overall goal is to develop an integrated methodology for systemic seismic vulnerability and risk analysis of urban systems, transportation and utility networks and critical facilities. SYNER-G developed an innovative methodological framework for the assessment of physical as well as socio-economic seismic vulnerability and loss assessment at urban and regional level. The built environment is modeled according to a detailed taxonomy into its components and sub-systems, grouped into the following categories: buildings, transportation and utility networks, and critical facilities. Each category may have several types of components. The framework encompasses in an integrated way all aspects in the chain, from regional hazard to vulnerability assessment of components to the socioeconomic impacts of an earthquake, accounting for relevant uncertainties within an efficient quantitative simulation scheme, and modeling interactions between the multiple component systems in the taxonomy. The prototype software (OOFIMS) together with several complementary tools are implemented in the SYNER-G platform, which provides several pre and postprocessing capabilities. The methodology and software tools are applied and validated in selected sites and systems in urban and regional scale. Representative results of the application in the city of Thessaloniki are presented here.

Seismic design forces of nonstructural components are commonly obtained by application of floor response spectra. This method is usually applied using estimated modal shapes and periods of the main structure; it allows for a separated design of components and their anchorages by the producers of equipment. Simplified formulas for determination of floor response spectra are provided by current codes such as Eurocode 8. All of them follow the assumption of the first fundamental elastic mode governing the acceleration values at the floors. These approaches do not take into account effects of higher modes, topology, ground response spectrum and plastification of supporting structures. Floor response spectra of four different building frames, one typical for an industrial 5-storey steel supporting structure and other three representing 5-, 10- and 15-storey regular steel buildings, were investigated using nonlinear incremental dynamic analyses. The results were compared to current code provisions revealing large discrepancies which have impact on safety as well as on economy of the design. Three aspects were identified and qualified: Application of ground response spectrum values instead of peak ground acceleration as basic input variable; Importance of higher modes; Impact of plastification of the main structure and the components. It could be shown that all three parameters have a significant influence on the acceleration values, on the dimensioning of the anchorages and on the ductility demand for components designed to dissipate energy.

The generation of synthetic earthquakes is an important point in earthquake design in order to have representative earthquake time histories for a given response spectrum. Different possibilities like amplitude modification and wavelet modification exist to match the synthetic earthquake with the target response spectrum as closely as possible and hence to allow for a design of secondary systems using direct time integration methods. However, for the design of secondary systems within a building against earthquake excitation, other methods are also applicable. Besides a calculation using direct time integration with a complete FE-model of the building including the secondary system, the floor response spectrum method or a simplified method, as given e.g. within the EC 8, may also be used. The applicability of these approaches, however, depends on their validity compared to the direct time integration method. This paper compares and discusses the known methods for the generation of earthquake time histories and checks the design methods for secondary systems within a large reinforced concrete structure to enable a reliable design of secondary systems against earthquake excitation.

The KTA safety standards not only apply to nuclear power plants but also to other nuclear facilities. The experience gained from retrofitting of structural and non-structural components in nuclear power plants can be applied to other areas where KTA standards are required. When designing and executing according to these standards best practices taken from conventional design often cannot be used. In many cases engineers and contractors are not aware of the additional expenditures involved. Differences between conventional design and design according to KTA standards are shown in the following areas: planning objectives, sources of infor-mation, changes to basis of design, probability levels, and, as an example, anchor-ing of a cable tray support to concrete.

This paper summarizes general aspects for the seismic qualification of equipment in industrial facilities. In the first part of the paper a guideline for the seismic qualification of equipment is described. The purpose of the guideline is to assist engineers in addressing the seismic design requirements for the equipment. Important points that need to be addressed are design philosophy, seismic performance goals, scope of application, methods of qualification, applicable codes and the description of seismic input. In the second part of the paper the seismic qualification of equipment for a facility located in China is described. The development of the seismic loads according to the applicable Chinese code is shown and for selected components the seismic qualification using numerical analysis and qualification by analogy is demonstrated.

For understanding the seismic behaviour of extra-large scale cooling tower with dimension of 220 meters high and 188 meters in diameter, the shake table tests for its’ 1:30 (length ratio) tower model were carried out to simulate the structural response to potential earthquake impacts. The model structure was excited by three dimensional white noise and different intensity of earthquake motions from PGA=0.04g to PGA=0.40g in considering of four different site conditions from soft soil to hard rock (I~IV). Through the tests, the dynamic responses and damage patterns of the cooling tower under different three-dimensional seismic excitations were studied.

The seismic qualification of electrical cabinets can be established by different methods like analysis, test and proof by analogy. This contribution gives two examples of seismic qualification; the first example shows the qualification of a cabinet with respect to stability and functionality. Stability is proved by analysis using a finite element model of the cabinet and performing an RSMA-calculation. By comparing the von-Mises comparative stress against permissible values, the stability of the cabinet is assessed. Functionality is proved by separate component tests. To define the test loading for the uniaxial component tests, the calculated maximum accelerations are used. The second example shows how to make use of a successful seismic qualification of a reference cabinet to qualify a similar cabinet with respect to stability. For this purpose the method of ‘proof by analogy’ (similarity) is used.

For the connection of steel structures and mechanical components like steel platforms, piping systems or vent pipes to concrete structures fastenings with metal anchors will be used. In nuclear power plants safety related fastenings require an adequate seismic design which is based on nuclear specific standards like the German safety standard series KTA 2201 “Design of Nuclear Power Plants against Seismic Events”. This safety standard series defines demands on determining the design basis earthquake as well as the design requirements of components and building structures including dynamic analysis procedures. The different demands on safety-related fastenings with anchors have been established in the German DIBt-guideline with the specifications for the technical approval of metal anchors and for the design of anchor connections. This guideline considers extraordinary action effects like earthquake actions. For example the load bearing of anchors has to be guaranteed in cracked concrete structures with large crack openings considering cyclic loading typically for seismic events. In addition to the DIBt-guideline the status report KTA-GS-80 presents a review about safety related fastenings in nuclear power plants. This report comprehends the essential information about the design and safety concept of those anchor fastenings regarding the interface between mechanical and constructional engineering. In this context the high demands on the limitation of deformations represents an important design criterion for the components with the assumption of a rigid connection to concrete structures.

Under seismic loading, the performance of a connection in a structure is crucial either to its stability or in order to avoid casualties and major economic impacts, due to the collapse of non-structural elements. In the United States the anchor seismic resistance shall be evaluated in accordance with ACI 318 Appendix D. Created in accordance with the ACI 355.2 regulated testing procedures and acceptance criteria ICC-ES AC193 and AC308, pre-qualification reports provide sound data in a proper design format. With the release of the ETAG Annex E at the beginning of 2013, the seismic pre-qualification of anchors became regulated in Europe. Anchors submitted to these new test procedures will now also incorporate in the ETA (European Technical Approval) all the required technical data for seismic design. Until the release of the EN 1992-4, planned for 2015, an EOTA TR (Technical Report) will set the standard for the seismic design of steel to concrete connections. Therefore, the design framework for the seismic design of anchors is already available through both the U.S. and European regulations.

Fastenings like headed studs and post-installed mechanical or chemical anchors for use in concrete are often used in Industrial Facilities. This paper deals with fastenings that are used to transmit seismic actions by means of tension, shear, or a combination of tension and shear,

- between connected structural elements

- between non-structural attachments and structural elements

Although the majority of fastenings up to now are designed and tested for use in non-seismic environments, they are commonly used for applications in structures in earthquake regions. Fasteners will be subjected to both crack cycling and load cycling at dynamic rates during an earthquake (Fig 1 [1, 2]). Therefore in future special requirements for the use of fasteners in seismic regions are demanded.

Metal anchors used to resist seismic actions shall meet the requirements of ETAG 001 Annex E: “Assesment of Metal Anchors under seismic action” [3]. The document deals with the preconditions, assumptions, required tests and assessment for metal anchors under seismic action. The design value of the effect of seismic actions acting on the fixture shall be determined according to Eurocode 8 [4]. Furthermore the EOTA “Technical report TR 045 – Design of Metal Anchors under Seismic Actions” [5] and the draft of Eurocode 2 part 4 “Design of Fasteners for Use in Concrete” [6] gives further requirements regarding the design of fasteners under seismic actions in addition to Eurocode 8.

Capacity design rules are applied to ensure the intended plastic behaviour of structures subjected to earthquakes. They need to cover scattering due to the seismic action and material strength. Recent evaluations of data of structural steel from European producers show for low steel grade an overstrength, which is higher than covered by the recommended value in DIN EN 1998-1. In this paper results of reliability analysis performed on this topic are presented. For this purpose a stochastic model for the seismic action was derived, which enables to perform the investigations via push-over analysis instead of time-consuming non-linear time history analysis. The main findings are discussed in the view of adjustments of design rules in DIN EN 1998-1 and European product standards, respectively.

The paper presents the development of a study on low cost seismic protection devices to put in place at the joints of prefabricated structural systems with the aim of improving their seismic response. In particular, this phase of the research focuses on the optimisation of protection devices used on two-dimensional mono-and multi-storey frames. A comparative analysis of the seismic response of the systems varying the mechanical characteristics of the devices was developed. The friction-type protective devices adopted were installed at the beam-column and column-foundation interfaces. The performed analyses show a significant improvement in seismic response, in terms of both reduction of stresses and increase of dissipative capacity.

Concrete-filled steel tube (CFST) consists of outer steel tube and concrete in-filled, which combines the merits of steel and concrete. This kind of composite member has various advantages, i.e., high strength and high ductility, favorable cyclic behaviour, high fire resistance and excellent constructability, have been recognized all over the world. Nowadays CFST has been widely used in construction, including many industrial facilities. This paper gives a brief review on the investigations of seismic behaviour of CFST members, joints, planar frames, hybrid walls and high-rise buildings, especially in China. The development of concrete-filled steel tubular members’ family is also introduced. Some industrial projects utilizing CFST members are also presented.

This paper presents a case study for identification of the dynamic characteristics of an industrial building with flexible steel moment resisting frame system based on ambient vibration measurements and short time monitoring. The field tests were conducted after detection of damage on non-structural separation masonry wall in the building with the intention to identify the properties of the building and detect possible sources of extreme operational conditions that lead to appearance of cracks in the walls. The accelerations, displacements and variations of temperature inside the building were monitored for 5 days via real time online monitoring system. The results revealed presence of continuous, but low level accelerations with different intensities throughout the building, significant variations of the relative displacements of the cracked wall in relation to the floor system and negligible variation of temperature.

Research development has demonstrated that numerical simulation is becoming one of the most powerful tools for collapse analysis of building structures in addition to the conventional laboratory model tests and post-earthquake investigations. In this paper, a finite element (FE) method based numerical model encompassing fiber-beam element model, multi-layer shell model and elemental deactivation technique is proposed to predict the collapse process of buildings subjected to extreme earthquake. The potential collapse processes are simulated for several different types of buildings. The analysis results indicate that the proposed numerical model is capable of simulating collapse process of buildings by identifying potentially weak components of the structure that may induce collapse. The study outcome will be beneficial to aid further development of optimal design philosophy.

The seismic analysis and design of liquid-filled storage tanks is an engineering problem connected with a significant degree of complexity, due to the liquid-structure-soil interaction that defines the dynamic response and determines the design of the tank. The implementation of adequate design rules is essential for ensuring the continuous operation of tanks after strong earthquakes and avoiding significant property loss and environmental damage. The European Standard Norms provide guidelines for steel shells, which can be implemented to investigate seismically excited liquid-filled tanks against buckling. This paper addresses the thematic area of shell buckling for the condition of seismic loading relevant for liquid-filled tanks, as defined in Eurocode 8 Part 4. The stress design approach described in Eurocode 3 Part 1-6 is examined, discussing the estimation and influence of the buckling relevant boundary conditions, geometrical tolerances and resistances. The concept is thereafter implemented by means of a parameterized design tool applicable for cylindrical anchored tanks based on rigid foundations. The investigation of a typical storage tank, the characteristic damage forms and their influence on the design are finally presented and evaluated. As the stress design procedure proposed in Eurocode 3 Part 1-6 remains the main option for the engineering practice, its implementation and results are examined in regard to their contribution to a cost-effective and earthquake resistant design.

Liquid filled tanks play an important role in the infrastructure of many industrial facilities assuring the supply with raw material needed for the production process or serving as storage for intermediate products. Due to their oftentimes large dimensions in diameter and height the stored fluid develops high seismic loads to the tank shell induced by the vibration of the liquid and the interaction of shell and liquid. In the design of tank shells the determination of the seismically induced pressure to the tank shell and the resulting overturning moments pose some challenges in engineering practice, especially with respect to the impulsive load component (interaction of shell and liquid). The following paper presents two different methods to calculate the eigenperiod, the eigenmode and the associated hydrodynamic pressure distribution for thin cylindrical liquid storage tanks for the circumferential wave number m=1 (lateral ground excitation). The first method includes an improved variation of the added-mass-iteration scheme: It employs a Rayleigh quotient’s formulation of the liquid-shell free vibration and repetitively manipulates the distribution of the kinetic energy of the fluid until convergence occurs. The second method involves the calculation of the added mass matrix directly from the appropriate expression for the work done by the liquid-shell interface forces on the basis of the radial displacement shape functions. The closed form solution of the governing matrix equation of motion of the shell enables the computation of higher eigenmodes and no iterative procedure is required.

Spherical pressure vessels are globally used for storage of pressurized liquids or gases of different hazard classes. An adequate seismic design of these structures must consider their particular structural behaviour and consequences of possible damage or failure. A study of the current standard situation for seismic design of pressure vessels revealed significant gaps and missing design rules, in particular for spherical pressure vessels. Within the European Research Project INDUSE the seismic performance and applicability of existing European and American codes to pressure vessels with cylindrical and spherical shape were investigated. This paper describes the results of a study on different examples of spherical pressure vessels which were selected to be representative for the current practice. The study comprised numerical investigations as well as simplified models for the estimation of the dynamic properties of the vessel structures. It is shown, which failure modes and stress concentrations areas are crucial in the event of an earthquake. In addition engineering calculation methods to determine fundamental periods and internal forces for braced and non-braced spherical pressure vessels were developed and compared to results of numerical simulations. The applicability of behaviour factors is discussed based on proposals made by European and American codes in comparison to own results. Recommendations for the behaviour factor of spherical pressure vessels with different dimensions were developed based on push over analyses and non-linear incremental dynamic analyses. Furthermore the influence of sloshing effects in spherical vessels, for which no specific rules are given in the codes, was investigated according to the current state of the art.

Seismic excited liquid filled tanks are subjected to extreme loading due to hydrodynamic pressures, which can lead to nonlinear stability failure of the thinwalled cylindrical tanks, as it is known from past earthquakes. A significant reduction of the seismically induced loads can be obtained by the application of base isolation systems, which have to be designed carefully with respect to the modified hydrodynamic behaviour of the tank in interaction with the liquid. For this reason a highly sophisticated fluid-structure interaction model has to be applied for a realistic simulation of the overall dynamic system. In the following, such a model is presented and compared with the results of simplified mathematical models for rigidly supported tanks. Finally, it is examined to what extent a simple mechanical model can represent the behaviour of a base isolated tank in case of seismic excitation

For design of industrial plants like LNG (liquefied natural gas) terminal the earthquake engineering for piping design is one of the most important design criteria [1]. The required calculation approaches in analyzing reactions of piping systems due to seismic events are specified in a variety of international and European codes and standards (e.g. in [2], [3] and [4]). Within these methods the simplified static equivalent method and the modal response spectra analysis are the most used in practice. From the engineering’s point of view the simplified static analysis has obviously its advantages. This is why it is often used to perform some preliminary or final stress calculations. But in practice it also can be seen that this approach is even extended to the piping connected to the storage tank, where the modal response spectra analysis shall be applied according to the codes [3] and [4]. Furthermore there is no precise prediction about the results of the simplified static method in the area of piping design, neither in aspect of reliability nor in aspect of economy. This article, based on a calculation of a typical unloading line for a new LNG storage tank – carried out by means of the CAESAR II program [7], compares the simplified static equivalent method and the modal response spectra analysis. The aim of this article is trying to set a general evaluation criterion and to give an answer to questions, under which conditions the simplified calculation method can be used. How big are the differences of the results between the two approaches?

The DIN EN 1998 standard and the VCI-Guideline provide comprehensive information for the design and verification of industrial facilities concerning earthquakes. For the fabrication and distribution of pressure equipment in the European Union the Pressure Equipment Directive (directive 97/23/EG) defines a basic framework. In Germany it was enacted by the Pressure Equipment Regulations. The requirements of the Pressure Equipment Regulations are included in harmonised standards, such as EN 13445. The application of other technical regulation is also possible in order to fulfil the Pressure Equipment Regulations. In Germany especially the AD-regulations are applied for the design of pressure vessels. They were developed by the chemical industry over many decades for a reliable and economical operation of the pressure equipment. The accordance of the AD-regulations with the Pressure Equipment Regulations has to be approved by a

Certification Body

when used in design practice. In the present paper the application of the AD-regulations in accordance with the specifications of the VCIGuideline is explained by means of a current project for the

Evonik Industries GmbH Rheinfelden

in the earthquake evaluation of pressure vessels. The consequences for the structural models and the design procedures will be shown. The influences on safety factors and mechanical properties are discussed in detail. With the suggested procedure it is possible to achieve a consistent earthquake design for pressure equipment in accordance with the current standards and guidelines.

Industrial facilities contain a large number of constructions and structural components. Both building and non‐building structures typically can be found in an industrial/chemical plant. Above ground pressurised tanks are typical examples of non‐building structures of such sites. These equipment are typically used for the storage of gas and liquid materials, e.g. chlorium, ammonia etc. The overall design of such structures, especially in low to moderate seismicity areas, has neglected any seismic loading in the past, basically due to the absence of relevant seismic requirements in previous codes. The seismic security of above ground pressurised tanks is of great importance, since failure of these structures can lead to negative impact for the environment and to economic losses. Recent codes for seismic design and construction of horizontal cylindrical reservoirs provide tools which can serve also to assess existing tanks. From experience, the seismic deficiencies of reservoirs of this type are in general concentrated in some strategic points. This paper describes the main deficiencies of such structures and the simplified methodology used for their assessment based on the guidelines presented in Eurocode 8. In addition, typical cost effective solutions for the retrofit of the tanks with these shortcomings are presented and critically discussed. The above assessment and retrofit methodology is illustrated for some examples of typical equipment.

Site effects can significantly modify the seismic motion in certain frequency domains, due to the resonance of soft deposits and subsequent amplification of the motion and / or due to the shape of the bedrock surface under soft deposits. Consequently, the shape of an appropriate elastic response spectrum might significantly differ from those proposed in building codes like EC8 based on a few soil classes. A site specific elastic response spectrum can either be lower or higher than the corresponding code spectrum or even both together, depending on the considered frequency band. Especially in the framework of assessment and reinforcement of existing industrial facilities, it might be of great importance to determine a site specific spectrum, much more adapted to account for local site effects. In some cases, such a specific spectrum makes it possible to save millions of unnecessary reinforcements. Some brief methodological aspects are presented, followed by real case examples, showing the importance of specific site effect studies and their consequences in terms of elastic response spectra for a more appropriate assessment or design of industrial facilities. In particular, the soil classification in EC8 is essentially based on Vs30 whereas site specific studies also account for the velocity contrast between the bottom of loose soil deposits and the bedrock, a parameter that can have a great influence on the amplitude of the resulting response spectrum.

The basic mechanisms of earthquake-induced soil liquefaction are introduced by considering the shaking of a block on a thin granular layer, which mechanical behaviour is modelled with a hypoplastic constitutive model. If the block is founded on a dry cohesionless soil or drainage of the granular layer is fully allowed, the soil densifies and the block settles step-wise. On the other hand, if drainage is impeded pore pressure develops and effective pressure decays with increasing number of shaking cycles, until, depending on the initial density, either a quasi-stationary cyclic state is reached or the effective pressure vanishes (liquefaction). The coupled nature of dynamic problems involving soils is also shown by the results of the analyses, i.e. the motion of the block causes changes of the soil state which in turn affect the block motion. Investigations of soil liquefaction under dynamic earthquake-like excitation with a 1-g laminar box confirm the predicted behaviour. The same constitutive equation is applied to the numerical simulation of the propagation of plane waves in homogeneous and layered level soil deposits induced by a wave coming from below. Welldocumented sites during strong earthquakes are used to verify the adequacy of the hypoplasticity-based numerical model for the prediction of soil response during strong earthquakes. It is concluded that liquefaction susceptibility during strong earthquakes can be reliably assessed with the proposed method. The influence of local site conditions, seismic excitation and nonlinearity of the soil behaviour on the ground response can be realistically taken into account by the model.

Seismic design and qualification of safety and/or radiological relevant structures for Swiss NPPs are subjected to rigorous procedures. Structures have to meet high safety standards, be robustly designed and therefore cover a wide range of parameter uncertainties both on seismic action and capacity side. In the framework of a wide Facility Power Retrofit at the Swiss Leibstadt NPP the Project ZENT (acronym of radioactive waste storage building) aims to erect two new structures on the site. Due to operational and radiological reasons, these new structures have to be built very close respectively above to the already mentioned existing structures. One structure has to be founded on piles above the existing NPP main cooling pipelines. Approval of this new seismic foundation system by Swiss Federal Nuclear Safety Inspectorate (ENSI) required extended seismic design. This paper attempts to describe the analyses performed by the Owner for seismic qualification and verification of structural integrity at planning stage.

The structural behavior of modern wind turbines has reached a very high complexity and many factors are involved: slenderness of the structure, excitation environment and operational controls. Moreover, if a project is located at sites with relevant seismic hazard, the wind unit must be designed considering a reasonable likelihood of earthquake occurrence during the operational state or an emergency shutdown. The influence of the subsoil on the seismic response of a wind turbine can be crucial during the seismic design phase and need to be properly included into the computational model. Norms and guidelines need to keep up with technological developments and structural peculiarities. However the dynamic soilstructure interaction is often neglected or roughly mentioned. It is usually suggested to represent the soil through springs. The proposed investigation estimates the seismic response of a soil-turbine system and involves a 1.5-MW, 3- blade wind turbine, grounded on a layered half space. The wind turbine system is modeled by means of

F

inite

E

lement

M

ethod (FEM). The effects of the layering are investigated. The soil is simply idealized as a generalized spring, according to the majority of standard codes. In parallel, the same investigation is performed with a more accurate method, a coupling between finite element and Boundary Element Method (BEM). This allows assessing the applicability and accuracy of the simplified soil representation.

Dynamic soil-structure interaction (SSI) effects have always been important in the context of assessing the seismic safety and vulnerability of large and complex infrastructures such as bridges, dams, power plants, industrial units etc. Although SSI problem has been under intensive investigations in the past several decades, relatively little is known about the SSI in the case of multi-layered half-space. Majority of previous researches dealing with SSI problems were conducted for rather simple soil profiles: elastic half-space, a soil layer bonded to rigid base, or a soil layer overlying elastic half-space. Very few papers appeared in the literature tackle the SSI problem having two or more soil layers overlying elastic half-space. This is probably due to the substantial computational effort required. Advantages and limitations of widely used current approaches such as finite element method (FEM), boundary element method (BEM) and thin-layer method (TLM) are discussed. Sponsored by the Science Foundation of Sino-German Center researches into SSI on layered soil carried out at Dalian University of Technology are briefly introduced. An advanced approach for dynamic SSI analysis of structures on multilayered half-space is proposed, which circumvent difficulties encountered by FEM, BEM and TLM with relative ease. The governing wave motion equations are solved in the frequency-wave-number domain analytically. The precise integration method (PIM) is employed to perform integration to obtain numerical results. Very high accuracy can be achieved. Analytical solution of wave motion equation is written in dual vector form, which enables efficient and convenient assembling of two adjacent layers into a new one without losing effective digits. Formulations dealing with dynamic impedance of arbitrary-shaped foundation on isotropic as well as arbitrary anisotropic multi-layered soil are presented. The solution is not difficult to extend to problems dealing with foundation embedment and throughthe- soil coupling of two or more foundations. Numerical results validate efficiency and accuracy of the proposed approach.

A high-order time-domain approach for wave propagation in bounded and unbounded domains is developed based on the scaled boundary finite element method. The dynamic stiffness matrices of bounded and unbounded domains are expressed as continued-fraction expansions. The coefficient matrices of the expansions are determined recursively. This approach leads to accurate results with only about 3 terms per wavelength. A scheme for coupling the proposed high-order time-domain formulation for bounded domains with a high-order transmitting boundary suggested previously is also proposed. In the time-domain, the coupled model corresponds to equations of motion with symmetric, banded and frequencyindependent coefficient matrices, which can be solved efficiently using standard time-integration schemes. A numerical example is presented.

A two-dimensional model based on thin layer method (TLM) has been used to analyze the attenuation of ground-borne vibration induced by the subway in Shanghai. The subway’s tunnel was simulated by the finite element method (FEM), and the nodes of FEM are coincident with the layer division of TLM. The frequency response functions on ground surface under action of the acceleration at ballast bed near rail track were calculated. By using the Fourier transform, the vibration acceleration and the vibration level (VL) induced by the subway on ground surface with deferent distance away from subway central line was analyzed. Comparison between the calculated and the measured VL at ground surface in Shanghai showed good agreement. Then the VL on ground surface induced by the subway in Shanghai with deferent distances and with deferent tunnel depths has been calculated. Finally the empirical attenuation equation of ground-borne vibration induced by subway in Shanghai has been proposed.

Taking a long-span arch bridge as an example, the characteristics of dynamic responses of the arch bridge under the horizontal and vertical travelling seismic wave excitations and the effect of the wave travelling velocity on the responses are discussed based on numerical analysis results. According to the structure symmetry, a simplified computation method is proposed for the travelling seismic response analysis of the bridge in time domain by using the seismic response of two half arch bridges under the uniform excitation. The time history analysis results indicate that the simplified method can achieve a good precision. The comparison of the numerical results describes the phenomena that the seismic response under travelling wave excitation does not linearly change with the seismic wave velocity. Through some further analysis, this article proposes a new concept of the resonant under travelling wave excitation, and the mechanism of the resonant effect in travelling seismic wave excitation is being expatiated.

It is generally recognized that soils and rocks in nature invariably exhibit some degree of anisotropy in their response to static or dynamic loads. An approach based on precise integration method (PIM) and the dual vector formulation (DVF) of wave motion equation is proposed for the evaluation of Green’s influence function of anisotropic stratified half-space. Then the problem of an arbitraryshaped foundation on a multi-layered subsoil is studied. The wave motion equation for a typical horizontally anisotropic layer and the half-space is solved analytically and the integration is performed by PIM. Any desired accuracy can be achieved. The DVF of wave motion equation is suggested for assembling the layers. As a result, the Green’s influence foundations for anisotropic stratified half-space are found based on the standard method in matrix algebra with the size of matrices not greater than (3×3). The computational effort is reduced to a great extent. Special treatment has been taken to preserve the effective digits. The computation is always stable. There is no limit of the thickness and number of soil layers to be considered. To satisfy the mixed boundary condition at the surface of arbitraryshaped foundation, the interface between the foundation and the multi-layered soils is discretized into a number of uniformly spaced sub-disks as in the usual manner. Numerical examples validate the efficiency and accuracy of the proposed approach.

Two parameters are proposed to improve the accuracy of the Green’s functions for a layered half space modelled with the thin layer method (TLM). The parameters, which define the thickness of the thin sub-layer and the buffer layer in the thin layer method, rely on the observation of the Green’s functions for a homogeneous half space. Based on them, the convergence of the Green’s functions at both highfrequency and low-frequency range can be ensured; and the efficiency of the thin layer method is improved.

Soil-structure interaction is widely recognized as a very important issue that should be considered in dynamic analysis and design of structures subjected to various dynamic disturbances such as earthquake or wind forces. An approach for timedomain response analysis of three-dimensional rigid surface foundations of arbitrary shape bonded to multi-layered soil is presented. The formulation consists of two parts: (a) frequency-spatial domain solution to the dynamic impedance of rigid surface foundation and (b) time-domain analysis by employing interpolating discrete values of dynamic impedance matrices by means of a continued matrix valued rational function. Practical applications compared with the analytical solutions or existing classical results dealing with rigid surface foundations of arbitrary shape demonstrate the accuracy and applicability of the proposed approach.

The target spectrum which has been used most frequently for the seismic analysis of structures is the Uniform Hazard Response Spectrum (UHRS). The joint occurrence of the spectral values in different periods, in the development of UHRS, is a key assumption which remains questionable. The Conditional Mean Spectrum (CMS) has been recently developed by Baker et al. as an alternative for UHRS. The CMS provides the expected response spectrum conditioned on the occurrence of the target spectral acceleration value in the period of interest which can be accounted as an improvement of the UHRS. In order to enhance the CMS, the correlation between the Peak Ground Velocity (PGV) and the spectral acceleration values has been investigated in the current study, and finally, a newer form of target spectrum has been proposed. It is shown that the emerged new spectrum, named Eta-based Conditional Mean Spectrum (E-CMS), is more efficient than the conventional CMS in order to enhance the UHRS, especially in the case of industrial facilities.