12 February 2003
York University Biosciences Research Facility
How an innovative use of steel and a vibration evaluation strategy produced a sensitive and flexible building
The effect of vibration on the serviceability of structures is becoming increasingly well documented and understood. However, Matthew Lovell, Associate at Buro Happold, argues that designers need practical guidance on both the performance standards to be achieved and the modelling techniques to be adopted.
Modelling human-induced vertical excitation of building floors is particularly important when they accommodate the ultra-sensitive equipment used in the rapidly developing micro- and nano-technologies in the UK. The problem is particularly acute for buildings accommodating research and development activities which require the use of electron microscopes and other sensitive equipment, or which are used for the production of microchips where the highest precision is required. Hospital operating theatres also require a high standard of resistance to vibration.
In 1999, Buro Happold and Anshen Dyer Architects were commissioned to design a £20million bio-science research facility at the University of York, funded by the Joint Infrastructure Fund, the Higher Education Funding Council, and Yorkshire Cancer Research.
During the design process, it became evident that equipment inside the building was sensitive to vibrations caused by the movement of people around the facilities. Buro Happold and Anshen Dyer proposed a design solution where the most vulnerable equipment, such as electron microscopes, would be mounted on concrete isolation plinths on the ground floor of the building. For the upper floors, they proposed a concrete frame solution, based on the standard laboratory space planning parameters.
However, contractors HBG Construction Ltd advised that a cost saving of between £200-250k could be achieved if a steel frame solution was adopted instead of a concrete frame. The Buro Happold structural group undertook feasibility studies to assess whether a steel frame could provide a similar level of performance to that of a concrete frame. Careful consideration was given to a steel frame proposal that would have similar performance characteristics to the concrete frame solution. As the fundamental dynamic performance of a structure depends on mass and stiffness, the traditional “slimflor” slab solution was modified to have a similar mass to the concrete frame. This resulted in a 400mm deep precast plank with a 75mm structural topping, with the benefit of a large reserve of strength, reduced cost and ease of erection. This solution also had the advantage of maintaining separate structure/servicing zones due to the absence of downstand beams, thus maintaining flexibility within the building height constraints. Finite element modelling clearly showed that the proposed steel frame was at least as good as the original concrete framed solution and that its response could satisfy existing guidelines for the design of laboratories.
A vibration evaluation strategy was proposed for the steel frame solution which sought to benchmark the dynamic performance of the steel frame against that of the concrete. This evaluation was followed by more detailed analysis of the steel frame to estimate the quantitative performance against the laboratory usage classes of ISO 2631. Input forces for the analysis were representative of one or more people walking along the floor at the most onerous positions for exciting vibration, and the analyses were performed parametrically to give a prediction of the range of performance that might be expected from the structure.
The area of human-induced excitation of structures has traditionally been little appreciated by British engineering consultants, as the vibration response of a structure is rarely deemed to be critical in its design. Often in the design office, simple hand calculations are performed to provide an estimate of the natural frequency and then the design proceeds according to vague rules. However, the effect that vibrations have on the serviceability of structures is becoming increasingly well reported and understood.
There are three main types of structure, namely footbridges, grandstands and floors, which are susceptible to dynamic excitation by humans, and should be checked for this type of loading prior to construction. For all three, there is a lack of research and understanding of the nature of human-structure dynamic interaction, and the past research has produced varying results difficult to use in design practice. Designers need practical guidance on both the performance standards to be achieved and the modelling techniques to be adopted.
When Buro Happold’s original analytical work was undertaken the building had not been constructed, yet its vibration response was already under scrutiny. This provided a unique opportunity to test the actual laboratory floor in its bare state and compare it to improved, and increasingly detailed, analytical models of the structure. The result was significant information on the response of buildings to vibrations.
Buro Happold commissioned Professor Alex Pavic, Head of the Vibration Engineering Research Section in the Department of Civil and Structural Engineering at the University of Sheffield, to assist with a second phase of research. Together they experimentally determined the response of the floor and proposed relevant input forces using modal testing of the floor, repeated measurements of walking-induced vibrations, finite element modelling and model updating to match the measured modal properties. The vibration responses to the three walking models were used in harmonic, transient and spectral analyses, and the results were then compared with their experimental counterparts. A literature review found several methods to assess the levels of RMS velocities caused by walking across a floor, and it is hoped that the comparison of results will give future designers confidence in the use of these methods.
The research team has suggested the following sequence for designers and consultants who feel that a building might be susceptible to human-induced vibrations.
Consult with the client to determine whether they think vibrations would have a significant effect on people or equipment.
Using industry guidelines, determine the minimum criteria acceptable for the structure.
If the client requires more stringent vibration criteria due to the nature of the work or the type of structure, explain the cost implications.
Using simple hand calculations, determine the natural frequency of the building components and classify as high or low frequency. Use this information to determine velocity and acceleration responses.
Compare these results to the required vibration criteria. If this analysis does not give adequate results, consider modelling the structure more carefully using a finite element program.
Use finite elements to determine the natural frequencies of vibration and the modal mass of the structure.
Make a realistic estimate of the damping and apply a realistic input force.
Check that the results generated by this modelling are within the structural vibrations criteria set by the client.
If not, seek advice on the modelling techniques employed and consider a different structural scheme.
Press Office and Practice information at www.burohappold.com
Buro Happold is a multi-disciplinary international practice of consulting engineers established in 1976 offering civil and structural engineering, mechanical and electrical engineering, quantity surveying, building services and environmental engineering, infrastructure and traffic engineering, geotechnical engineering, façade engineering, fire engineering, computational fluid dynamics analysis, access consultancy, project management, urban design and a range of specialist CAD services.
