MRC Laboratory of Molecular Biology
CUED - Cambridge University Engineering DepartmentNot a member yet
45551 research outputs found
Sort by
A design methodology to reduce the embodied carbon of concrete buildings using thin-shell floors
No full textThis paper explores the potential of thin concrete shells as low-carbon alternatives to floor slabs and beams, which typically make up the majority of structural material in multi-storey buildings. A simple and practical system is proposed, featuring pre-cast textile reinforced concrete shells with a network of prestressed steel tension ties. A non-structural fill is included to provide a level top surface. Building on previous experimental and theoretical work, a complete design methodology is presented. This is then used to explore the structural behaviour of the proposed system, refine its design, and evaluate potential carbon savings. Compared to flat slabs of equivalent structural performance, significant embodied carbon reductions (53–58%) are demonstrated across spans of 6–18 m. Self-weight reductions of 43–53% are also achieved, which would save additional material in columns and foundations. The simplicity of the proposed structure, and conservatism of the design methodology, indicate that further savings could be made with future refinements. These results show that considerable embodied carbon reductions are possible through innovative structural design, and that thin-shell floors are a practical means of achieving this
An Exactly Force-Balanced Boundary-Conforming Arbitrary-Lagrangian-Eulerian Method for Interfacial Dynamics
No full textWe present an interface conforming method for simulating two-dimensional and axisymmetric multiphase flows. In the proposed method, the interface is composed of straight segments which are part of mesh and move with the flow. This interface representation is an integral part of an Arbitrary Lagrangian-Eulerian (ALE) method on an moving adaptive unstructured mesh. Our principal aim is to develop an accurate and robust computational method for interfacial flows driven by strong surface tension and with weak viscous dissipation. We first construct a discrete solutions satisfying the Laplace law on a circular/spherical interfaces exactly, i.e., the balance between the surface tension and the pressure jump across an interface is achieved exactly. The accuracy and stability of these solutions are then investigated for a wide range of Ohnesorge numbers, \Oh. The dimensionless amplitude of the spurious current is reduced to machine zero, i.e., on the order of for \Oh \ge 10^{-3}. Finally, the accuracy and capability of the proposed method are demonstrated through a series of benchmark tests with larger interface deformations. In particular, the method is validated with Prosperitti's analytic results of the bubble/drop oscillations and Peregrine's dripping faucet experiment, in which the values of \Oh are small
A hybrid continuous energy and multi-group Monte Carlo method
No full textThis paper investigates mixing multi-group (MG) and continuous energy (CE) representation of cross-sections depending on location of a particle in a Monte Carlo neutronic eigenvalue calculations in 1D and 2D PWR test cases with UOX and MOX fuel. Different population normalisation needs to be applied to CE and MG region to account for the difference in criticality between CE and MG representation. This normalisation procedure requires a neutron production rate ratio between CE and MG region to be known a priori. A resonance correction in energy spectrum during transition of a particle between the MG and CE region was developed based on the equivalence resonance treatment theory. With these, it was shown that it is possible to accelerate total calculation time, while introducing only a moderate error below 1% in the fission rate distribution. The magnitude of acceleration is heavily dependent on the relative size of CE and MG zones
Quantitative 3D imaging parameters improve prediction of hip osteoarthritis outcome
No full textOsteoarthritis is an increasingly important health problem for which the main treatment remains joint replacement. Therapy developments have been hampered by a lack of biomarkers that can reliably predict disease, while 2D radiographs interpreted by human observers are still the gold standard for clinical trial imaging assessment. We propose a 3D approach using computed tomography—a fast, readily available clinical technique—that can be applied in the assessment of osteoarthritis using a new quantitative 3D analysis technique called joint space mapping (JSM). We demonstrate the application of JSM at the hip in 263 healthy older adults from the AGES-Reykjavík cohort, examining relationships between 3D joint space width, 3D joint shape, and future joint replacement. Using JSM, statistical shape modelling, and statistical parametric mapping, we show an 18% improvement in prediction of joint replacement using 3D metrics combined with radiographic Kellgren & Lawrence grade (AUC 0.86) over the existing 2D FDA-approved gold standard of minimum 2D joint space width (AUC 0.73). We also show that assessment of joint asymmetry can reveal significant differences between individuals destined for joint replacement versus controls at regions of the joint that are not captured by radiographs. This technique is immediately implementable with standard imaging technologies
Incompleteness of Atomic Structure Representations
No full textMany-body descriptors are widely used to represent atomic environments in the construction of machine-learned interatomic potentials and more broadly for fitting, classification, and embedding tasks on atomic structures. There is a widespread belief in the community that three-body correlations are likely to provide an overcomplete description of the environment of an atom. We produce several counterexamples to this belief, with the consequence that any classifier, regression, or embedding model for atom-centered properties that uses three- (or four)-body features will incorrectly give identical results for different configurations. Writing global properties (such as total energies) as a sum of many atom-centered contributions mitigates the impact of this fundamental deficiency - explaining the success of current "machine-learning"force fields. We anticipate the issues that will arise as the desired accuracy increases, and suggest potential solutions
Shear behaviour of fabric formed T beams reinforced using W-FRP
No full textA combination of flexible moulds as external formwork and bespoke robotically fabricated fibre reinforced polymer cages as tensile reinforcement offers a new opportunity for the manufacture of structural concrete components that have been optimised to minimise material use. This technology could potentially help in our quest to reduce carbon emissions in the construction industry, yet there remain technical issues to overcome if such flexibly formed concrete structures are to become a reality. This paper presents experimental research on fabric-formed T beams reinforced with wound fibre-reinforced polymer (W-FRP) to quantify the shear contribution of this novel system. It is shown that, depending on the geometry of the beam, carefully chosen flexural and shear reinforcement can resist shear in a predictable manner. Because of geometric variation along the length of the beam, shear resistance is found to move from being provided by both the W-FRP reinforcement and the sloping longitudinal reinforcement to being provided predominantly by the longitudinal FRP reinforcement as the W-FRP gradually ruptures. In turn, this demands higher anchorage capacity of the longitudinal bars than that might have been expected by design codes of practice. By overcoming such issues, this paper shows that savings in concrete of up to 64% can be made in the webs in such structures, compared with conventional T-beams
A simple implicit coupling scheme for Monte Carlo neutronics and isotopic depletion
No full textThe stochastic implicit Euler scheme and its variants can be used to prevent non-physical behaviour that may emerge when coupling Monte Carlo neutron transport and isotopic depletion solvers for spatially-decoupled reactor problems. However, stochastic implicit methods tend to require many iterations to obtain a stable solution. This paper demonstrates that this is due to using a sub-optimal relaxation scheme: rather than using a variable relaxation factor, a fixed relaxation factor can give a more stable solution in fewer iterations. Furthermore, like stochastic implicit schemes, even though multiple transport solutions are required, using a fixed relaxation factor allows computational effort to be reduced by lowering the number of particles simulated during the corrector step while still providing stable results. This shows that using a fixed relaxation factor to stabilise Monte Carlo burn-up calculations can be more effective than applying the stochastic approximation
Plasma production of nanomaterials for energy storage: Continuous gas-phase synthesis of metal oxide CNT materials: Via a microwave plasma
No full textIn this work we show for the first time that a continuous plasma process can synthesize materials from bulk industrial powders to produce hierarchical structures for energy storage applications. The plasma production process's unique advantages are that it is fast, inexpensive, and scalable due to its high energy density that enables low-cost precursors. The synthesized hierarchical material is comprised of iron oxide and aluminum oxide aggregate particles and carbon nanotubes grown in situ from the iron particles. New aerosol-based methods were used for the first time on a battery material to characterize aggregate and primary particle morphologies, while showing good agreement with observations from TEM measurements. As an anode for lithium ion batteries, a reversible capacity of 870 mA h g-1 based on metal oxide mass was observed and the material showed good recovery from high rate cycling. The high rate of material synthesis (∼10 s residence time) enables this plasma hierarchical material synthesis platform to be optimized as a means for energetic material production for the global energy storage material supply chain
Silicon Photonics Codesign for Deep Learning
No full textDeep learning is revolutionizing many aspects of our society, addressing a wide variety of decision-making tasks, from image classification to autonomous vehicle control. Matrix multiplication is an essential and computationally intensive step of deep-learning calculations. The computational complexity of deep neural networks requires dedicated hardware accelerators for additional processing throughput and improved energy efficiency in order to enable scaling to larger networks in the upcoming applications. Silicon photonics is a promising platform for hardware acceleration due to recent advances in CMOS-compatible manufacturing capabilities, which enable efficient exploitation of the inherent parallelism of optics. This article provides a detailed description of recent implementations in the relatively new and promising platform of silicon photonics for deep learning. Opportunities for multiwavelength microring silicon photonic architectures codesigned with field-programmable gate array (FPGA) for pre- and postprocessing are presented. The detailed analysis of a silicon photonic integrated circuit shows that a codesigned implementation based on the decomposition of large matrix-vector multiplication into smaller instances and the use of nonnegative weights could significantly simplify the photonic implementation of the matrix multiplier and allow increased scalability. We conclude this article by presenting an overview and a detailed analysis of design parameters. Insights for ways forward are explored