Search

This paper presents a novel approach to passivity enforcement of multiport frequency-dependent rational models. The strategy employs graph theory, and convex optimization to extend the positive-definite matrix completion problem to passive system completions. The main distinguishing feature of the proposed technique relative to those in the literature is that passivity compensations are only applied to those parameters contributing to passivity violations, in a completion manner. Numerical experiments using actual transformer data reveals that the proposed method results in models with increased accuracy.

When it comes to estimating dynamics (modes) of interconnected power systems through transient (ringdown) data, one can expect more realistic estimates for the compound system if multiple signals distributed over different locations can be used simultaneously during modal estimation. In this context, the contributions of this paper are twofold. First, it is proposed a multi-signal ringdown time-domain Vector Fitting method (‘multi-signal RTD-VF’) which is capable of modeling multiple ringdown signals simultaneously. This method appropriately generalizes the ‘single-signal RTD-VF’ method, which processes ringdown signals independently, i.e., one at a time. Second, it is proposed an instrumental variable (IV) extension of multi-signal RTD-VF. By guaranteeing local optimums after algorithm convergence, such a ‘multi-signal IV-RTD-VF’ formulation may be used for further refining estimates provided by regular multi-signal RTD-VF. Two numerical examples are used to validate the proposed methods. The first example considers actual ringdown data sets extracted from North American Eastern Interconnection (NAEI) and Brazilian Interconnected Power (BIP) systems, whereas the second one considers mode identification of a synthetic signal corrupted by colored noise.

This paper addresses the problem of obtaining optimal black-box models of n-terminal power transformers for transient studies. We use actual transformer data to demonstrate that a transformer model endowed with both passivity and symmetry can be optimally achieved only if the corresponding constraints are formulated and simultaneously included in the parameter estimation problem as opposed to enforcing them independently as two distinct post-processing model adjustments.

This paper proposes an optimal and unifying instrumental variable (IV)-based Vector Fitting (VF) method for frequency-domain (FD) identification of models formed by rational basis function (RBF) expansions. The proposed method is denoted by IV-FD-VF and can be similarly applied for estimating models formed by both continuous- and discrete-time RBF sets. The key advantage of IV-FD-VF lies in the fact that, differently from standard FD-VF approaches, it guarantees an optimal solution after convergence. This important optimality property is proved to be independent of the nature of the noise that corrupts the data (for instance, if it is white or coloured). Two case studies are used to validate the proposed IV-FD-VF method. One of these case studies considers actual frequency-domain data sets extracted from two different single-phase power transformers.

In this paper, a methodology using Differential Evolution algorithm is proposed for the definition of the discourse universe of the fuzzy PID controller structure variables, for the load/frequency control in power systems, with this eliminating the need of the expert in systems speed control of hydraulic turbines. This controller uses the fuzzy system to tune the PID controller in real time. A mathematical model, based upon constructive information and where nonlinearities are taken into account, of a hydroelectric plant belonging to the Brazilian Interconnected System is presented as a test benchmark to analysis the proposed controller performance under various loading conditions. Simulation results are introduced to show the performance of the developed controller in comparison with a PID controller tuned by a Differential Evolution Algorithm and with a PID algorithm similar to the one actually used in the field. These results denote that the proposed controller offer better performance over others in terms of settling times and oscillation of the frequency.

This paper proposes a novel method, called Ringdown Time-Domain Vector Fitting (RTD-VF), for estimating electromechanical modes in interconnected power systems. Such a method authentically extends, to the context of ringdown analysis, the well known Time-Domain Vector Fitting (TD-VF) method, which has already been successfully applied within other power systems areas. The proposed method is based on a state-space discretization framework which enables ringdown events to be effectively estimated when described as artificial unit impulse responses. Moreover, RTD-VF completely avoids the necessity to perform discrete Fourier transforms (DFTs) of ringdown data sequences. Three case studies are used to validate the proposed method. One of the examples considers a synthetic test signal, whereas the other two case studies consider actual ringdown data sets extracted from the North American Eastern Interconnection (NAEI) system and from the Brazilian Interconnected Power (BIP) system.

Estimating electromechanical oscillations of power systems plays a crucial role to infer about their stability. In this paper it is proposed a new method in which dominant dynamics of power systems are inferred by analyzing pole estimates produced by several regularized robust recursive least squares (R3LS) implementations operating individually — each one with a different set of autoregressive moving average exogenous (ARMAX) model orders. The proposed method is named “Clustered R3LS”, since it properly adapts the so-called K-means clustering algorithm to automatically model regions of pole agglomeration produced by these individual R3LS implementations. An additional strategy for ignoring (discarding) poles identified as spurious is also part of the proposed Clustered R3LS algorithm. From a practical point of view, Clustered R3LS is here shown to more effectively track dominant dynamics of power systems when compared to Classical R3LS, which is based on analyzing only pole estimates produced by a single R3LS implementation operating with a fixed set of ARMAX model orders that is usually hard to tune. Three case studies are here used to validate the proposed Clustered R3LS method for both ambient and ringdown data. The first and second case studies are based on two well known benchmark models proposed for the analysis of oscillatory dynamics in power systems, whereas the third case study considers actual power system data extracted during a large-scale event from the Brazilian Interconnected Power (BIP) system.

This paper presents a subspace-based frequency identification method with passivity enforcement via Linear-Matrix Inequalities to achieve a globally optimal and passive solution. These inequalities are formulated and solved as a convex optimization problem using a state-of-the-art numerical solver. We also apply a weighting methodology to the LMI formulation to enhance the approximation between the model and the measured frequency-domain data. A case study including a coaxial cable experimental setup is presented to showcase and validate the proposed approach. Our results reveal that the proposed algorithm solves passivity enforcement problems in a single step without interactions and is able to successfully estimate passive models within the subspace framework.

Vector fitting (VF) algorithms have become popular and powerful tools for estimating models formed by rational basis function (RBF) expansions. In this paper, we first translate the well-known continuous time-domain VF method (cTD-VF) to a discrete time-domain framework. We denote this new domain VF method by dTD-VF. Differently from the cTD-VF, the dTD-VF formulation does not rely on a numerical approximation of convolution integrals and, as a result, it can be easily implemented with a variety of RBF sets. The second part of this paper shows that the proposed dTD-VF can also be transformed into a novel instrumental variable (IV)-dTD-VF technique, which is shown to have a guaranteed optimal solution at convergence. Moreover, this important optimality property does not depend on the nature of the noise that corrupts the data (for instance, if it is white or colored). Two case studies highlight the advantages of using the proposed methods. One of these examples consists of modeling the admittance characteristics of a power system implemented as a frequency-dependent network equivalent (FDNE).

This paper presents a nonlinear model for a hydraulic amplifier, a component of the governor in the speed control loop of hydroelectric power plants. The amplifier transforms the electrical signal of the controller into mechanical movement of the turbine components, including the switching characteristics. This model is used to propose, on the one hand, a Piecewise Affine (PWA) representation for the hydraulic amplifier, and, on the other hand, a methodology for estimating the model parameters using field measurements, which facilitates its practical implementation. This representation is referred to as semi-physical because the model parameters are closely related with the physical construction of the hydraulic amplifier components. Among the advantages of this PWA representation are the appropriate structuring for use in system identification methods, for estimating its parameters, and the existence of advanced control techniques that use this structure in controller design, thereby improving the load-frequency control performance. The paper concludes with a description of the results, including the parameters that were estimated by using the hydraulic amplifier model with PWA structure.

The passivity enforcement problem is always a key issue when estimating passive black-box models. Passivity has been broadly scrutinized in the literature and is known to yield stable time-domain simulation; thus, passivity is a much-desired model property which cannot remain an unmodeled effect. Many successful algorithms exist that accomplish the enforcement task by model parameter perturbation, with the caveat that the passivity violations to be suppressed are not large. This paper proposes a novel approach to address this issue. We propose four different mathematical formulations to preprocess the measurement of admittance parameters. Preprocessed data serve as inputs to parameter estimation algorithms to obtain passive models for power system applications. This approach is particularly adequate to use with model parameter perturbation methods. Three examples containing data of actual measurements illustrate the advantage of using the proposed methodology.

Power system transient analysis relies on the deployment of models with ever-increasing accuracy that represents the system's frequency-dependent components with respect to the phenomena being scrutinized. A key issue in this context is the construction of a model yielding passive properties for each system component. Such properties are commonly compromised by corrupting noise in measurements that serve as input to the computation of blackbox models. In addition, model passivity could be further compromised during parameter estimation. This paper undertakes the task of developing a proposal to build an optimization methodology based on linear matrix inequalities (LMI-based) which is intended for obtaining an optimal, passive, blackbox model for a $n$ -terminal power transformer. The resulting model can be further realized as a circuit representation in order to simulate the transformer dynamics in an Eletromagnetic Transient program (EMT type). The results derived from two operating power transformers are later provided to confirm how effective the procedure herein proposed is.

This article presents a novel approach to analysing a power transformer’s internal response to system transients. In this approach a hybrid modelling methodology is adopted which leverages the distinct advantages offered by both Black and Grey Box modelling techniques. The Black Box model of the transformer is used within the EMTP system study environment in order to take advantage of its mathematical flexibility and modelling accuracy. Transients derived from network switching operations within the study can then be used for injection tests within the Grey Box modelling environment. The Grey Box model, which is based upon the physical structure of the transformer, will facilitate analysis of the transformer’s internal voltage response to the external stimulus. A fundamental difference between the approach described in this paper and more traditional approaches is that it does not require prior knowledge of the internal geometry of the transformer. All of the modelling parameters are derived from external tests, nameplate details and an intrinsic understanding of common transformer design principles. This can be a distinct advantage since in most cases a transformer’s design specifications are not readily available outside of the laboratory due to the manufacturer’s intellectual property restrictions. A study of a gas insulated substation within a hydroelectric power plant in Brazil is used to demonstrate the proposed technique.

Modeling dynamic systems from frequency-domain data is a relevant issue for system analysis in many fields of engineering. Many techniques to approach such problems are based on models built by using a series of rational functions with fixed structure. As far as rational basis function models are concerned, the problem of selecting the function’s poles or dynamics is an issue and the Sanathanan–Koerner (SK) iterations represent a successful approach that can be applied to solve it. Frequency localizing basis function (FLBF) models are a rational function which can be used in such a context. However FLBF, as originally proposed, are not suitable for application with SK-iterations because they have not been designed to incorporate complex poles. To overcome this issue and to make the application of SK-iterations possible, this paper proposes a new set of basis functions that are based on the FLBF. In this paper, they will be referred to as complex frequency localizing basis functions. This new set of basis functions preserves the FLBF behavior but real and complex poles can be incorporated in the model and so it allows the use of SK-iterations. The proposed model is validated through simulation.

Automated test systems (ATS) are application-specific systems that help in enhancing reliability and productivity of testing activities in different stages of product development cycle and manufacturing. Different tests are necessary to anticipate and validate a product and its sub-systems performances along the development cycle. Moreover, in the production phase, quality control systems that perform up to 100% inspection may be present and the integration of test results with company's management systems may be required. In this paper, the design of ATS is analyzed under the light of design methodologies, which can basically be understood as procedures stating stages and intermediary results to discipline the development process of technical systems. Usually, design methodologies are presented as a set of prescribed actions and recommendations, kept with some degree of generality to allow their compatibility with a wide range of industrial products, based on different technology branches. In this paper, an adaptation of Pahl and Beitz's approach allowed the derivation of a reference model to ATS development. Proposed stages are explained considering ATS typical division, i.e.: physical system, hardware and software. The adapted methodology, in which prescriptions are made with an intermediary degree and directed to one particular kind of system, can make easier the design of a highly context-dependant kind of system, helping the design team to consider important aspects in decisions taken along the development process and keeping these decisions linked to the requirements defined in early stages. Through the use of the proposed methodology, shorter development cycles are expected, since less design loops tend occur in a higher-systematization environment.

The present work is focused on the study of indoor thermal comfort control problem in buildings equipped with HVAC (heating, ventilation and air conditioning) systems. The occupants’ thermal comfort sensation is addressed here by the well-known comfort index known as PMV (predicted mean vote) and by a comfort zone defined in a psychrometric chart. In this context, different strategies for the control algorithms are proposed by using an only-one-actuator system that can be associated to a cooling and/or heating system. The first set of strategies is related to the thermal comfort optimization and the second one includes energy consumption minimization, while maintaining the indoor thermal comfort criterion at an adequate level. The methods are based on the model predictive control scheme and simulation results are presented for two case studies. The results validate the proposed methodology in terms of both thermal comfort and energy savings.

Regression equations can be used for predicting indoor air temperature, relative humidity and energy consumption in an easier and more rapid way than building energy simulation tools. The independent variables, that is, the input data, are heating, ventilation and air conditioning (HVAC) power, outdoor temperature, relative humidity and total solar radiation. The present methodology for obtaining the regression equations is based on defining a couple of linear Multiple-Input/Single-Output (MISO) models, since two main outputs are involved, that is, indoor temperature and relative humidity. The methodology has been tested for the low- and high-thermal mass cases of the BESTest model (cases 600 and 900) and the output data is generated by using a building hygrothermal simulation tool. Validation procedures have shown very good agreement between the regression equations and the simulation tool for both winter and summer periods.