A new spatiotemporal regularization approach for reconstruction of cardiac transmembrane potential patterns

The single-beat reconstruction of electrical cardiac sources from body-surface electrocardiogram data might become an important issue for clinical application. The feasibility and field of application of noninvasive imaging methods strongly depend on development of stable algorithms for solving the underlying ill-posed inverse problems. We propose a novel spatiotemporal regularization approach for the reconstruction of surface transmembrane potential (TMP) patterns. Regularization is achieved by imposing linearly formulated constraints on the solution in the spatial as well as in the temporal domain. In the spatial domain an operator similar to the surface Laplacian, weighted by a regularization parameter, is used. In the temporal domain monotonic nondecreasing behavior of the potential is presumed. This is formulated as side condition without the need of any regularization parameter. Compared to presuming template functions, the weaker temporal constraint widens the field of application because it enables the reconstruction of TMP patterns with ischemic and infarcted regions. Following the line of Tikhonov regularization, but considering all time points simultaneously, we obtain a linearly constrained sparse large-scale convex optimization problem solved by a fast interior point optimizer. We demonstrate the performance with simulations by comparing reconstructed TMP patterns with the underlying reference patterns.     

A comparison of noninvasive reconstruction of epicardial versus transmembrane potentials in consideration of the null space

We compare two source formulations for the electrocardiographic forward problem in consideration of their implications for regularizing the ill-posed inverse problem. The established epicardial potential source model is compared with a bidomain-theory-based transmembrane potential source formulation. The epicardial source approach is extended to the whole heart surface including the endocardial surfaces. We introduce the concept of the numerical null and signal space to draw attention to the problems associated with the nonuniqueness of the inverse solution and show that reconstruction of null-space components is an important issue for physiologically meaningful inverse solutions. Both formulations were tested with simulated data generated with an anisotropic heart model and with clinically measured data of two patients. A linear and a recently proposed quasi-linear inverse algorithm were applied for reconstructions of the epicardial and transmembrane potential, respectively. A direct comparison of both formulations was performed in terms of computed activation times. We found the transmembrane potential-based formulation is a more promising source formulation as stronger regularization by incorporation of biophysical a priori information is permitted.     

A parallel,load-balanced MHD code on locally-adapted,unstructured grids in 3d

An efficient parallel code for the approximate solution of initial boundary value problems for hyperbolic balance laws is introduced. The method combines three modern numerical techniques: locally-adaptive upwind finite-volume methods on unstructured grids, parallelization based on non-overlapping domain decomposition, and dynamic load balancing. Key ingredient is a hierarchical mesh in three space dimensions.The proposed method is applied to the equations of compressible magnetohydrodynamics (MHD). Results for several testproblems with computable exact solution and for a realistic astrophysical simulation are shown.     

Authentication analysis of caffeine-containing foods via elemental analysis combustion/pyrolysis isotope ratio mass spectrometry (EA-C/P-IRMS)

The δ¹³CVPDB and δ¹⁸OVSMOW values of caffeine isolated from commercial caffeine-containing foods were determined by isotope ratio mass spectrometry (IRMS) using elemental analysis (EA) in the 'combustion' (C) and 'pyrolysis' (P) modes (EA-C/P-IRMS), respectively. In total, 69 samples (32 guarana products, 20 teas, 12 coffee products, and 5 maté teas) were analyzed together with three references of synthetic caffeine. Validation was done by excluding isotope discrimination in the course of sample preparation and determining linear dynamic ranges for EA-P-IRMS measurements. In total, five samples (one caffeine-containing soft drink and four guarana products) under study contained synthetic caffeine.     

Separation of plutonium and neptunium species by capillary electrophoresis-inductively coupled plasma-mass spectrometry and application to natural groundwater samples

Capillary electrophoresis (CE) was coupled to ICPMS in order to combine the good performance of this separation technique with the high sensitivity of the ICPMS for the analysis of plutonium and neptunium oxidation states. The combination of a fused-silica capillary with a MicroMist AR 30-I-FM02 nebulizer and a Cinnabar small-volume cyclonic spray chamber yielded the best separation results. With this setup, it was possible to separate a model element mixture containing neptunium (NpO2(+)), uranium (UO2(2+)), lanthanum (La3+), and thorium (Th4+) in 1 M acetic acid. The same conditions were also suitable for the separation of various oxidation states of plutonium and neptunium in different aqueous samples. All separations were obtained within less than 15 min. A detection limit of 50 ppb identical with 2 x 10(-7) M (3-fold standard deviation of a blank) was achieved. To prove the negligible disturbance of the plutonium and neptunium redox equilibria during the CE separations, plutonium and neptunium speciation by CE-ICPMS in acidic solutions was compared with the results of UV/visible absorption spectroscopy and was found to be in good agreement. The CE-ICPMS system was also applied to study the reduction of Pu(VI) in a humic acid-containing groundwater at different pH values.     

A toolset for building the virtual enterprise

Much research has been undertaken to define what a virtual enterprise is and how it should work. This paper addresses the specific question, how a virtual enterprise can be designed to have the agility to support short-term business opportunities. A framework is presented for the organizational design and the changing business roles of the business architect who constructs the various phases of the virtual enterprises lifecycle. This infrastructure for creating virtual enterprises is referred to as the value system designer; a set of methods and tools to select partners, re-engineer business- and logistic processes and to set up an information and communication platform for the virtual enterprise. The methods and tools have been developed in two longitudial research projects TELEflow and the Virtuelle Fabrik between 1995 and 1999. Focusing on the experiences gained from numerous cases, a summary on crucial success factors for designing virtual enterprises shall be presented. Thus, this paper gives insights and applicable know-how for companies and managing engineers in their role as virtual enterprise architects, for example leaders of project consortia or joint ventures or as first-tier suppliers co-ordinating supplier (sub-) nets.     

Rolling and recrystallisation textures in Cu–Al,Cu–Mn and Cu–Ni alloys

The development of strongly cube textured Cu based substrates is important in the cost effective production of long lengths of high temperature superconducting cables. The present paper reports textures (deformation and recrystallisation) development in pure Cu, Cu–Al, Cu–Mn (with a solute content of 1–3 at.%) and Cu–35 at.% Ni alloys.     

State/event fault trees—A safety analysis model for software-controlled systems

Safety models for software-controlled systems should be intuitive, compositional and have the expressive power to model both software and hardware behaviour. Moreover, they should provide quantitative results for failure or hazard probabilities. Fault trees are an accepted and intuitive model for safety analysis, but they are incapable of expressing state dependencies or temporal order of events. We propose to combine fault trees with an explicit State/Event semantics, using a graphical notation that is similar to Statecharts. Our new model, named State/Event Fault Trees (SEFTs), subsumes both deterministic state machines suited to describe software behaviour, and Markov chains that model probabilistic failures, while keeping the visualisation of causal chains known from fault trees. We allow exponentially distributed probabilistic events, deterministic delays, and triggered events. The model provides a component concept, where components are connected by typed ports. Quantitative evaluation is achieved by translating the component models to Deterministic and Stochastic Petri Nets (DSPNs) and using an existing tool for analysis or simulation. This paper, which is an extended version of [Kaiser B, Gramlich C. State-Event-Fault-Trees—a safety analysis model for software controlled systems. Computer safety, reliability, and security. Proceedings of the 23rd international conference, SAFECOMP 2004, Potsdam, Germany, September 21st–24th. Lecture Notes in Computer Science, vol. 3219, 2004.p. 195–209], revisits the model elements and the analysis procedure and provides a small case study of a fire alarm system, completed by an outlook on our tool project ESSaRel.     

变频器中间直流总线供电的开关电源

“带有360～900V直流输入,可以直接由变频器中间直流总线供电“,这样的电源听起来非常令人心动。但是这样的电源却有着特别的设计要求。传统上,用于控制系统和其他应用的电源都是直接由单相或三相电网供电的。然而,由于越来越多的变频器以及伺服电机放大器的应用,出现了新的供电可能:由变频器的中间直流总线供电(参见图1)。这种供电的优点在于,可以利用运转的电机中储存的“免费“动能来为控制系统供电。如果这种可能成为现实,将大大提高电源相对电网波动的稳健性,而不必使用需要经常性维护的蓄电池缓冲系统。     

Process identification with multiple neural network models

This paper describes the identification of chemical processes with multiple neural network models. This concept is called a 'multimodel' approach. The multimodel approach presents a flexible framework which allows the integration of other model paradigms. Three different methods to construct neural network multimodels are presented. First, a priori knowledge is used to decompose the input domain into operating regimes. Each regime is modelled by a different neural network. In the second multimodel approach, unsupervised learning in form of clustering and SOMs is used to split the input domain. The third approach uses a gating neural network to divide the input space. In contrast to the first two approaches, the non-linear gated network approach allows the multim odel to simultaneously learn a suitable decomposition and the mapping in each regime. All three approaches are evaluated for a fed-batch fermentation.