MRC Laboratory of Molecular Biology
CUED - Cambridge University Engineering DepartmentNot a member yet
45551 research outputs found
Sort by
Thermal Compensation for High Load Spatial Light Modulators in Real-Time
No full textWe introduce a concurrent real-time calibration process for diffractive lithography using liquid crystal spatial light modulators. This allows for increased power handling capabilities for given phase modulation accuracy
Design of a SiC-Based Switched CCM/TCM Inverter for High-speed Machine Drive with Low PWM-Induced Current Ripple
No full textThis paper presents the design of a SiC-based inverter for high-speed machine using a continuous conduction mode (CCM) and triangular conduction mode (TCM) switched switching scheme. The implementation of the switched TCM and CCM on machine drive are explored analytically in the context of PWM-induced current quality and efficiency, based on which, an improved switched switching strategy is employed. According to the difference in uneven current ripple distribution of TCM and CCM, the proposed strategy alternates between TCM and CCM in one line cycle for an enhanced current ripple and efficiency performance
Seismic response of partially saturated soils beneath shallow foundations under sequential ground motions
No full textInduced partial saturation is an innovative soil improvement technique intended to mitigate earthquake-induced liquefaction. Historical records indicate that successive earthquakes may occur in high seismic areas. Therefore, a comprehensive understanding of the response of partially saturated soils to sequential ground motions is of great significance for the rational design and execution of this method in engineering practice. In this study, a series of dynamic centrifuge experiments were conducted to investigate the impacts of sequential ground motions on the behavior of partially saturated soils beneath shallow foundations. Two different shallow foundation models with a bearing pressure of 135 kPa and 50 kPa were examined. Three seismic simulations, in order of increasing amplitudes, were sequentially applied to loosely-packed partially saturated sand models prepared with air injection technique. The assessment of the test results indicated that shallow foundations resting on saturated models of loose sand did suffer excessive settlements with each event, producing a large embedment of the foundation. However, much smaller settlements were recorded for partially saturated ground, and the level of the foundation embedment remained limited in this case. The deformation vector fields also indicated that different displacement mechanisms were observed for each successive event
Asymptotic behaviour of the posterior distribution in approximate Bayesian computation
No full textApproximate Bayesian computation (ABC) is a popular technique for approximating likelihoods and is often used in parameter estimation when the likelihood functions are analytically intractable. In the context of Hidden Markov Models (HMMs), we analyze the asymptotic behavior of the posterior distribution in ABC based Bayesian parameter estimation. In particular we show that Bernstein-von Mises type results still hold but that the resulting posterior is biased in the sense that it concentrates around a point in parameter space that differs from the true parameter value. Furthermore we obtain precise rates for the size of this bias with respect to a natural accuracy parameter of the ABC method. Finally we discuss, via a numerical example, the implications of our results for the practical implementation of ABC
A suboptimal H<inf>2</inf> clustering-based model reduction approach for linear network systems
No full textIn this paper we propose a (suboptimal) H2 model reduction method for a class of linear network systems that describe diffusively coupled networks. To preserve a network structure, we form a reduced-order model by using the characteristic matrix of a graph clustering so that the reduced-order model has less number of vertices. Then, we formulate the model reduction problem as a nonconvex optimization problem with binary variables, aiming for a graph clustering that minimizes the H2-norm of the approximation error. Based on the controllability and the observability Gramians of the error system we derive an optimization problem with mixed-binary variables and then we propose a convex relaxation of the binary variables, leading to a smooth optimization formulation. For this new optimization problem we derive an explicit expression for the gradient of the objective function and then we employ a projected gradient algorithm for solving the optimization problem with mathematical guarantees on its convergence
Reinforcement learning for molecular design guided by quantum mechanics
No full textAutomating molecular design using deep reinforcement learning (RL) holds the promise of accelerating the discovery of new chemical compounds. Existing approaches work with molecular graphs and thus ignore the location of atoms in space, which restricts them to 1) generating single organic molecules and 2) heuristic reward functions. To address this, we present a novel RL formulation for molecular design in Cartesian coordinates, thereby extending the class of molecules that can be built. Our reward function is directly based on fundamental physical properties such as the energy, which we approximate via fast quantum-chemical methods. To enable progress towards de-novo molecular design, we introduce MOLGYM, an RL environment comprising several challenging molecular design tasks along with baselines. In our experiments, we show that our agent can efficiently learn to solve these tasks from scratch by working in a translation and rotation invariant state-action space
Effect of catalyst preparation and storage on chemical looping epoxidation of ethylene
No full textChemical looping epoxidation (CLE) of ethylene to ethylene oxide (EO) presents an exciting alternative to the incumbent technology of direct epoxidation of ethylene with O2(g). In CLE, the reaction is still catalysed by Ag but oxygen is provided as Olattice from a solid metal oxide, eliminating the need and limitations of using O2(g). Here, the influence of catalyst preparation in CLE is investigated. The Ag catalyst was impregnated on a solid oxide, SrFeO3, which was used as the donor of Olattice in CLE. Temperature programmed reduction in H2 indicated that O-species participating in CLE can be attributed to the removal of the first monolayer of oxygen in SrFeO3. By changing the temperature of calcination in Ag-SrFeO3 preparation, the size of Ag particles was varied, resulting in a simultaneous increase of selectivity for EO (up to 60%) and conversion of C2H4 (up to 10%). Then, an assessment of the effects of impurities (carbonates, hydroxides), which deposit over time on the Ag-SrFeO3 surface, was performed, showing that impurities deteriorate the catalyst performance in CLE. Finally, the doping of the SrFeO3 with Ce to the A-site of the perovskite was carried out, and a substantial improvement of the conversion of C2H4 was achieved, reaching 15%, while maintaining the 60% selectivity for EO
Technology progress on quantum dot light-emitting diodes for next-generation displays.
No full textQuantum dot light-emitting diodes (QD-LEDs) are widely recognised as great alternatives to organic light-emitting diodes (OLEDs) due to their enhanced performances. This focus article surveys the current progress on the state-of-the-art QD-LED technology including material synthesis, device optimization and innovative fabrication processes. A discussion on the material synthesis of core nanocrystals, shell layers and surface-binding ligands is presented for high photoluminescence quantum yield (PLQY) quantum dots (QDs) using heavy-metal free materials. The operational principles of several types of QD-LED device architectures are also covered, and the recent evolution of device engineering technologies is investigated. By exploring the fabrication process for pixel-patterning of QD-LEDs on an active-matrix backplane for full-colour display applications, we anticipate further improvement in device performance for the commercialisation of next-generation displays
Thermoacoustic stabilization of a longitudinal combustor using adjoint methods
No full textWe construct a low-order thermoacoustic network model that contains the most influential physical mechanisms of a thermoacoustic system. We apply it to a laboratory-scale longitudinal combustor that has been found to be thermoacoustically unstable in experiments. We model the flame, which is behind a bluff body, by a geometric level-set method. We obtain the thermoacoustic eigenvalues of this configuration and examine a configuration in which six eigenmodes are unstable. We then derive the adjoint equations of this model and use the corresponding adjoint eigenmodes to obtain the sensitivities of the unstable eigenvalues to modifications of the model geometry. These sensitivities contain contributions from changes to the steady base flow and changes to the fluctuating flow. We find that these two contributions have similar magnitudes, showing that both contributions need to be considered. We then wrap these sensitivities within a gradient-based optimization algorithm and stabilize all six eigenvalues by changing the geometry. The required geometry changes are well approximated by the first step in the optimization process, showing that this sensitivity information is useful even before it is embedded within an optimization algorithm. We examine the acoustic energy balance during the optimization process and identify the physical mechanisms through which the algorithm is stabilizing the combustor. The algorithm works by, for each mode, reducing the work done by the flame while simultaneously increasing the work done by the system on the outlet boundary. We find that only small geometry changes are required in order to stabilize every mode. The network model used in this study deliberately has the same structure as one used in the gas turbine industry in order to ease its implementation in practice