Method of Fitting a Nonlinear Function to Measurement Data, Its Uncertainty and as Example the Application in Thermistor Circuits

A new direct parameter extraction method of small-signal equivalent circuit for radio frequency laterally-diffused Metal Oxide Semiconductor Field Effect Transistor (RF LDMOSFET) with Faraday shield biased in cut-off operation is presented in this paper. A series of analytical equations are derived for non-linear rational function fitting as well as linear regression to measured device frequency response characteristics. The method is successfully applied to a set of fabricated RF LDMOSFETs with different geometry scales. As a result, all of the cut-off small-signal equivalent circuit elements are determined. Validation of the extraction method is verified by good agreement between the simulation results and the corresponding measurement data. In addition, reasonable physical meaningfulness of the method is further demonstrated by characterizing the device geometrical dependences of the extracted gate-drain capacitance (C gd ) and drain-source capacitance (C ds ).

This paper presents a parameter estimation algorithm for dynamical systems based on frequency responses. By means of the discrete measurement data, a nonlinear cost function is established for deriving the parameter estimation algorithm. In order to achieve good convergence and high stability, we present a self-regulation factor algorithm for estimating system parameters. Moreover, the gradient parameter estimation algorithm is derived to compare the effectiveness of the proposed self-regulation factor algorithm effectiveness. The simulation results indicate that the proposed method works efficiently.

This paper considers the identification problem of network structures for networked dynamical systems. We define the network structure as a coupling function describing the network connectivity and the nonlinear data exchange functions in the network, and attempt to identify the coupling function from potentially noisy measurement data. We develop an identification method of network structures applicable even to network systems consisting of nonlinear systems and nonlinear coupling functions by using the Koopman operator theory. First, we design observable functions as basis functions of a functional space, and determine the Koopman operators associated with the dynamics of the network. Then, the coupling function is identified as a projection on the span of the observables. Also, we make use of the sparse identification techniques to reduce requirements on data amounts and improve robustness with respect to measurement noise. Numerical examples show that the proposed method is applicable to a wide range of nonlinear systems, including chaotic systems with nonlinear coupling functions, and yields better performance than some existing methods. Identification results for two different nonlinear network systems with nonlinear coupling functions show the usefulness of the proposed method.

A new solution to the problem of nonlinear parameter estimation is considered. The peculiarity of this problem is that the arguments of a nonlinear function are not only the measurement data and the required parameters, but also supplementary parameters. Supplementary parameters are a priori unknown, but they are necessary only to find optimal estimates of the required parameters. The estimate of the desired parameters is formed according to the point-mass method as a weighted sum of partial estimates obtained for the specified values of supplementary parameters. This solution allows us to eliminate a priori probabilities for supplementary parameters by taking into account additional covariance of weight coefficients and (or) specified partial estimates. The considered approach is effective in solving specified nonlinear estimation problems characterized by low accuracy of available measurement data and (or) their few number.

The working of three-way catalytic converters (TWCs) is based on chemical reactions whose rates are nonlinear functions of temperature and reactant concentrations all along the device. The choice of suitable expressions and the tuning of their parameters is particularly difficult in dynamic conditions. In this paper, we introduce a hybrid modeling technique which allows one to preserve the most important features of an accurate, distributed parameter TWC model, while it circumvents both the structural and the parameter uncertainties of classical reaction kinetics models, and saves the computational time; in particular, we compute the rates within the TWC dynamic model by a neural network which becomes a static nonlinear component of a larger dynamic system. A purposely designed genetic algorithm, in conjunction with a fast ad hoc partial differential equation integration procedure, allows one to train the neural network, embedded in the whole model structure, using currently available measurement data and without computing gradient information.

What characteristics and risk factors are associated with Chronic Obstructive Pulmonary Disease (COPD) in non-smokers?

SummaryAccurate prediction of the solubility of chemical substances in solvents remains a challenge. The sparsity of high-quality solubility data is recognized as the biggest hurdle in the development of robust data-driven methods for practical use. Nonetheless, the effects of the quality and quantity of data on aqueous solubility predictions have not yet been scrutinized. In this study, the roles of the size and the quality of data sets on the performances of the solubility prediction models are unraveled, and the concepts of actual and observed performances are introduced. In an effort to curtail the gap between actual and observed performances, a quality-oriented data selection method, which evaluates the quality of data and extracts the most accurate part of it through statistical validation, is designed. Applying this method on the largest publicly available solubility database and using a consensus machine learning approach, a top-performing solubility prediction model is achieved.

Improved fault detection employing hybrid memetic fuzzy modeling and adaptive filters

This paper presents a kernel-based adaptive filter that is applied for the digital domain self-interference cancellation (SIC) in a transceiver operating in full-duplex (FD) mode. In FD, the benefit of simultaneous transmission and receiving of signals comes at the price of strong self-interference (SI). In this work, we are primarily interested in suppressing the SI using an adaptive filter namely adaptive projected subgradient method (APSM) in a reproducing kernel Hilbert space (RKHS) of functions. Using the projection concept as a powerful tool, APSM is used to model and consequently remove the SI. A low-complexity and fast-tracking algorithm is provided taking advantage of parallel projections as well as the kernel trick in RKHS. The performance of the proposed method is evaluated on real measurement data. The method illustrates the good performance of the proposed adaptive filter, compared to the known popular benchmarks. They demonstrate that the kernel-based algorithm achieves a favorable level of digital SIC while enabling parallel computation-based implementation within a rich and nonlinear function space, thanks to the employed adaptive filtering method.

Due to the widespread installation of Distributed Energy Resources (DER) and Demand Side Management resources (DSM) into Distribution Systems (DS), the usage pattern of the DS has changed remarkably in recent times. As a result, the DS has the potential to lead its network operational parameters, such as bus voltage profiles and line flows, outside of their allowable limits. Therefore, they increase the variability and range of variation of loads in the network. However, the DS generally are not designed for such conditions. Thus, it becomes necessary to reassess and redesign the network capacity. A standard strategy would be to increase the capacity of the network, however, this is costly. A cost-effective strategy could be to improve the management of such a system by improving the determination of its actual capacity. Recent advances in ICT equipment make it appear possible to manage the critical loading conditions instead of increasing capacity of the network. Thus, an energy management system is required to perform that function during critical loading. In this thesis, a framework for controlling a network-oriented autonomous energy management system on the distribution network using network capacity constrained is proposed. The framework is based on a State Estimation (SE) technique, which is able to determine the most probable system state of a network. Practically, the system’s operational parameters may be unobservable due to measurement deficiency, loss of communication etc. In such a situation, the observability analysis needs to be performed on the available measurement data to investigate whether the operational parameters are observable or unobservable. Such an analysis is taken into account in this thesis. The proposed framework is akin to an Optimal Power Flow (OPF) problem, where SE gives a starting point to the simulation layer. The simulation layer supports the optimization under the network constraints. In this framework, the challenging task is to formulate the network constraints. To this end, a linearization based map is generated from the functional relationships between an arbitrary set of control variables and an extensive variety of constrainable operational parameters. This map permits the formulation of linear inequality constraints to build a flexible system to manage the network within existing capacity. The feasibility of the proposed constrained generation is demonstrated using an extensive variety of cost functions from a simple Linear Programming (LP) to non-linear functions like absolute value. The significant feature of this approach is that it is flexible both in terms of constrained parameters and control variables. Along with observability, controllability is another crucial factor regarding capacity constrained state optimization. The controllability analysis would allow the decision to be made of whether or not available control variables allow changing the value of operational parameters to specific values without other influences. With this in mind, a technique to analyze the controllability has been developed in this thesis. In addition, some technical applications either already implemented (e.g. phase balances or back-to-back systems) or expected in the near future (e.g. market mechanism) introduce the interdependencies between control variables within the system. The ability to handle the interdependence will be important elements in future distribution system management. Therefore, an approach for modeling of interdependencies introduced by energy transaction and a back-to-back system is developed. In this approach, all control variables are assumed to be independent and are usually chosen to be nodal parameters describing the interaction of equipment connected, but not an element of the network, is the ability to model interdependencies amongst these nodal variables. This approach can originate in a range of different scenarios. One is phase balancing, another is back-to-back systems between networks or feeder, and another one is energy transactions. Furthermore, this framework also facilitated the future smart applications in market structure, namely price bidding scheme and the peer-to-peer energy transfer platform to transact energy between consumers. The proposed framework is applied to two networks namely, an IEEE 13 bus test system and a 29 bus reference distribution network.