Frequency-domain Regression Research Articles

High precision nodes localization algorithm based on optimal dual-frequency comparison ranging

ABSTRACTIndoor localization plays an important role in medical treatment and industrial application. Dual-frequency phase-comparison method is widely applied to indoor localization. However, frequency was randomly chosen in most current dual-frequency phase-comparison schemes, which reduced the ranging accuracy. Therefore, an IL-ODFC (indoor localization scheme based on optimal dual-frequency comparison ranging) scheme was proposed. An indoor localization system and a wireless sensor nodes based on dual-phase measurement were designed and established in order to form target nodes and reference nodes and send their communication results to the location engine of indoor localization system. Firstly, reasons for errors of traditions dual-frequency phase-comparison ranging in IL-ODFC scheme was analyzed; then, the concept of frequency domain and one-time linear regression idea was introduced; next, LS (least square) method was used to choose optimal difference frequency in order to obtain accurate ranging data. After obtaining distances among three reference nodes, LS was used to estimate the location of the target nodes. The simulation result shows that the proposed IL-ODFC scheme can effectively improve location accuracy.

An advanced simulation test bed for the stability analysis of variable air volume air-conditioning control system. Part 1: Optimal simplified model of building envelope for room thermal performance prediction

This paper and its future companion papers, were aimed to develop a set of efficient and accurate enough modular tools to build a simulation test bed for the stability analysis of interacting VAV (variable air volume air-conditioning system) control loops. By providing the correlation equations between the heat flow and temperature of the two opposite surfaces, the wall model was expected to link adjacent zones effectively and efficiently. In part 1, the development of an optimal simplified 3R2C (composed of three resistances and two capacitances) thermal network model of building envelope, i.e., the wall model, was descripted. The verification showed that the simplification of one-dimensional heat conduction through a typical wall of heavy construction was acceptable. The accuracy of the optimal simplified 3R2Cmodel was satisfying, and outperformed that of the optimal 4R3C model as well as two 3R2C models which were normally used in practical applications. First and foremost, the room dimensions of the test bed with the boundary conditions in numerical scheme were designed under the principle of similarity. After validation of the numerical scheme with the high-cited scaled experiment, a series of numerical test cases simulating the potential oscillations in practical applications, were conducted. Then, the long neglected concerned frequency range within which heat flow transfer through extern wall, was determined. Finally, the part of solid wall was separated with the part of indoor air, and was developed into a flexible wall module of 3R2C model based on frequency domain regression.

Thermal performance evaluation of a massive brick wall under real weather conditions via the Conduction Transfer function method

The reliable estimation of buildings energy needs for cooling and heating is essential for any eventual thermal improvement of the envelope or the HVAC equipment. This paper presents an interesting method to evaluate the thermal performance of a massive wall by using the frequency-domain regression (FDR) method to calculate CTF coefficients by means of the Fourier transform. The method is based on the EN ISO 13786 (2007) procedure by using the Taylor expansion for the elements of the heat matrix. Numerical results were validated through an experimental heating box with stochastic boundary conditions on one side of the wall representing real weather conditions and constant temperature profile on the other side representing the inside ambiance in real cases. Finally, a frequency analysis was performed in order to assess the validity and accuracy of the method used. The results show that development to the second order is sufficient to predict the thermal behavior of the studied massive wall in the range of frequencies encountered in the building applications (one hour time step). This method is useful for thermal transfer analysis in real weather conditions where the outside temperature is stochastic; it also allows the evaluation of the thermal performance of a wall through a frequency analysis.

An examination of the REIT return–implied volatility relation: a frequency domain approach

This paper examines the relationship between implied volatility (the VIX) and REIT returns using frequency domain approach which allows shocks to vary across frequency bands. The distinguishing feature of the frequency domain method is that it enables the investigator assess quantitatively the impact of independent variables on the dependent variable at different frequencies across the spectra. The estimates of the parameter of interest from the frequency domain analysis may reveal rich policy implications. Specifically, at issue is whether implied volatility can be used to predict movements in REIT returns at different frequencies in the United States. The results from the frequency domain regression show that implied volatility and REIT returns have significantly negative effect on each other at the low-, medium- and long-term frequencies. Furthermore, the empirical findings from the frequency domain causality tests indicate that causality runs from implied volatility to all and equity REIT returns in the short- and medium-term frequencies but not vice versa. However, it is interesting to note that the results show causality running from mortgage REIT returns to implied volatility in the medium term. Taken together, the results from this study suggest that knowledge of implied volatility can help the investors to predict movements in the capital market and hence, to some extent, they can protect their portfolios against uncertainties.

Frequency domain regression method to predict thermal behavior of brick wall of existing buildings

The prediction of the thermal behavior of the envelope of old buildings (built before 1948), subjected to various boundary conditions, is very useful for the implementation of an effective thermal renovation strategy. For this purpose, the present work aims at studying the transient thermal behavior of a brick wall and determinate the optimum insulation thickness by using a theoretical model based on the Frequency-Domain Regression (FDR) model. The brick wall of 34cm thick (main characteristics of Northern European old buildings) has been used in this work. The numerical results agree well with experimental results obtained on a specially developed experimental setup. Subsequently, the model has been used to investigate the effects of the insulation thickness on the energy requirement and total cost.Simulation results indicated that thermal insulation is able to enhance the thermal behavior of massive wall, and that insulation thickness has influence on the profile of heat flux. The optimum insulation thicknesses for internal insulation vary between 5.5 and 12.4cm, vary between 4.9 and 8.4cm for external insulation depending on the fuel types.

An improved frequency-domain regression method for structural damage detection in wireless sensor network

Distributed computing strategy over a network of wireless sensors has emerged as a promising application in order to enable real-time structural health monitoring. Most previous studies of distributed computing strategy consider relatively simple operations like modal parameters identification (e.g. frequencies and modal shapes) that cannot directly assess structural health condition demanded by many practical engineering applications. This article focuses on developing distributed computing strategy techniques in wireless sensor networks used for direct stiffness estimation of structural damage locations and extent. An improved frequency-domain regression method is proposed to detect and quantify damage in frame structures. In the proposed method, the static condensation technique is adopted to reconstruct the rotational responses on joints, which is considered as inputs to estimate stiffness. Therefore, this improved method provides a more practical and accurate tool for structural damage detection. A numerical simulation of a five-story two-bay frame structure demonstrates that the proposed damage detection method using a limited number of wireless sensors can identify, locate, and quantify the structural stiffness changes accurately.

Discouraged or added worker effect: Which one prevails in the Polish labour market?

We test whether the discouraged worker or added worker effect prevails in the Polish labour market. The discouraged worker effect implies that the participation rate is procyclical with respect to the GDP and countercyclical with respect to the unemployment rate. The added worker effect yields contrary findings. We analyse the period 1994–2014 with quarterly data. We focus on the working age population, both males and females. We apply a range of methods to obtain robust results, some of which have never been employed to resolve this problem. They include ad hoc filtration, spectral analysis, unobserved component model, time-varying parameter model, and frequency domain regression. The results indicate that the added worker effect prevails in most of the business cycle frequencies. It is significant and varies over time. It is true for both males and females. It is considerably stronger in contractions than in expansions. In low business cycle frequencies, the discouraged worker effect prevails. Although the last case is rare, it proves the heterogeneity of labour force behaviour over the business cycle.

Comparative Analysis of Two Methods for Conduction Transfer Functions of Building Constructions

Conduction transfer function (CTF) is widely used to calculate conduction heat transfer in building cooling/heating load and energy calculations. It can conveniently fit into any load and energy calculation techniques to perform conduction heat transfer calculations. There are two popular methods, direct root-finding (DRF) method and frequency-domain regression (FDR) method to calculate CTF coefficients. The limitation of methodology possibly results in imprecise or false CTF coefficients. This paper analyzes the errors of two methods to the material properties of a multilayer heavyweight building construction. The results show that the calculation error of DRF method becomes increasing as the wall characteristic parameter becomes increasing. The maximal error even reaches almost 100%. However, the calculation error of FDR method always remains within 1% no matter how the wall characteristic parameters are varied. Thus, FDR method is more robust and reliable than DRF method.

Testing for seasonal unit roots by frequency domain regression

This paper develops univariate seasonal unit root tests based on spectral regression estimators. An advantage of the frequency domain approach is that it enables serial correlation to be treated non-parametrically. We demonstrate that our proposed statistics have pivotal limiting distributions under both the null and near seasonally integrated alternatives when we allow for weak dependence in the driving shocks. This is in contrast to the popular seasonal unit root tests of, among others, Hylleberg et al. (1990) which treat serial correlation parametrically via lag augmentation of the test regression. Our analysis allows for (possibly infinite order) moving average behaviour in the shocks. The size and power properties of our proposed frequency domain regression-based tests are explored and compared for the case of quarterly data with those of the tests of Hylleberg et al. (1990) in simulation experiments.

Errors Analysis of Calculation Methods for Conduction Transfer Functions of Building Constructions

Conduction transfer function (CTF) is widely used to calculate conduction heat transfer in building cooling/heating load and energy calculations. It can conveniently fit into any load and energy calculation techniques to perform conduction heat transfer calculations. There are two popular methods, state-space (SS) method and frequency-domain regression (FDR) method to calculate CTF coefficients. The limitation of methodology possibly results in imprecise or false CTF coefficients. This paper analyzes the errors of two methods to the material properties of a single-layer and a multilayer heavyweight building construction. The results show that the calculation error of SS method becomes increasing as the wall characteristic parameter becomes increasing. The maximal error even reaches almost 100%. However, the calculation error of FDR method always remains within 1% no matter how the wall characteristic parameters are varied. Thus, FDR method is more robust and reliable than SS method.

A frequency-domain regression method for estimating building foundation heat transfer

In this article, the Interzone Temperature Profile Estimation solutions are first utilized to carry out a frequency analysis for the foundation heat transfer to determine the effect of indoor air and ambient air temperature fluctuations on the variation of the foundation heat loss/gain. Then, a frequency-domain regression-based method is used to develop transfer functions for ground-coupled surfaces. The regression frequency-domain method is tested and validated using one-dimensional heat transfer for above-grade walls. It is found that the proposed method provides a more effective alternative than numerical methods to estimate conduction transfer function coefficients for building foundations and thus can be easily implemented in whole-building simulation programs.

A simplified dynamic model of building structures integrated with shaped-stabilized phase change materials

The use of phase change materials (PCM) to enhance the building energy performance has attracted increasing attention of researchers and practitioners over the last few years. Thermodynamic models of building structures using PCMs are essential for analyzing their impacts on building energy performance at different conditions and using different control strategies. There are few PCM models of detailed physics providing good accuracy in simulating thermodynamic behavior of building structures integrated with PCM layers. However, simplified models with acceptable accuracy and good reliability are preferable in many practical applications concerning computation speed and program size particularly when involving large buildings or models are used for online applications. A simplified physical dynamic model of building structures integrated with SSPCM (shaped-stabilized phase change material) is developed and validated in this study. The simplified physical model represents the wall by 3 resistances and 2 capacitances and the PCM layer by 4 resistances and 2 capacitances respectively while the key issue is the parameter identification of the model. The parameters of the simplified model are identified using genetic algorithm (GA) on the basis of the basic physical properties of the wall and PCM layer. Two GA-based preprocessors are developed to identify the optimal parameters (resistances and capacitances) of the model by frequency-domain regression and time-domain regression respectively. Validation results show that the simplified model can represent light walls and median walls integrated with SSPCM with good accuracy.

Short time step heat flow calculation of building constructions based on frequency-domain regression method

Short time step heat flow calculation of building constructions is often needed for practical applications. Conventional methods such as state-space method and root-finding method may produce unstable conduction transfer function (CTF) coefficients at short time steps, and thus result in unstable heat flow calculation through building constructions. Frequency-domain regression (FDR) method is a newly developed method for computing CTF coefficients efficiently by representing the real building construction system with equivalent polynomial s-transfer functions. Previous studies on this method mainly addressed CTF coefficients at the conventional time step of 3600 s and the performance of heat flow calculation using these coefficients. This paper presents the investigation on the performance of CTF coefficients at various short time steps based on FDR method, and the performance of the heat flow calculation using these coefficients. The results show that FDR method can produce stable CTF coefficients at various time steps for most building constructions, and the calculated heat flows using these coefficients are of high accuracy.

Applicability of calculation methods for conduction transfer function of building constructions

Conduction transfer function (CTF) is widely used to calculate conduction heat transfer in building cooling/heating loads and energy calculations. It can conveniently fit into any load and energy calculation techniques to perform conduction heat transfer calculations. There are three popular methods, direct root-finding (DRF) method, state-space (SS) method and frequency-domain regression (FDR) method to calculate CTF coefficients. The limitation of a methodology possibly results in imprecise or false CTF coefficients. This paper investigates the applicability of three methods as Fourier number and thermal structure factor are varied and in detail explains the sources that introduce error in the CTF solutions. The results show that the calculation error of SS and DRF methods becomes increasing as the reciprocal of the product of Fourier number and thermal structure factor becomes increasing. The maximal error even reaches almost 100%. However, the calculation error of FDR method always remains within 1% no matter how Fourier number and thermal structure factor are varied. Thus, FDR method is more robust and reliable than SS and DRF methods. And it is more practical to calculate CTF coefficients and may be a better choice to calculate the cooling/heating loads for building structures for the architect/designer.

An improvement to frequency-domain regression method for calculating conduction transfer functions of building walls

An improvement is proposed for the frequency-domain regression (FDR) method to calculate CTF (conduction transfer function) coefficients of multilayer constructions. This improvement aims at providing complete CTF coefficients with a unique set of d values. In the improvement, the external and internal polynomial s-transfer functions of heat conduction of a wall are identified based on equivalent frequency response characteristics by constraining their denominators equal to that of the cross polynomial s-transfer function while this cross polynomial s-transfer function of this wall is identified using FDR method in advance. The complete CTF coefficients of this wall (i.e., external CTF values a, cross CTF values b, internal CTF values c and a unique set of heat flow rate history coefficients d) can be easily calculated from these identified polynomial s-transfer functions. Examples and comparisons show that this improvement of the frequency-domain regression method works well.

Optimal simplified thermal models of building envelope based on frequency domain regression using genetic algorithm

Simple and effective building energy models are essentially needed for many applications, such as building performance diagnosis and optimal control, etc. The energy model involves two important parts of building components, i.e., building envelopes and building internal mass. This paper presents a methodology for parameter optimization of 3R2C thermal network model of building envelopes (composed of three resistances and two capacitances) based on frequency domain regression using genetic algorithm (GA). First, the theoretical frequency characteristics of heat transfer through building envelope are calculated using detailed physical description within the frequency range of concern. Second, the frequency characteristics of the simplified 3R2C model are calculated with random values of individual resistances and capacitances which constrain to total thermal resistance and capacitance. Then, the errors between the theoretical frequency characteristics and the frequency characteristics of the simplified model are calculated. Finally, GA estimator is developed to optimize the parameters of the simplified model, allowing the frequency responses of the simplified model match the actual heat transfer through building envelope the best. Various case studies are conducted also to validate the parameter optimization method of the simplified 3R2C model. The accuracy of simplified models for constructions of different weights is studied.

Second order optimality for estimators in time series regression models

We consider the second order asymptotic properties of an efficient frequency domain regression coefficient estimator β ^ proposed by Hannan [Regression for time series, Proc. Sympos. Time Series Analysis (Brown Univ., 1962), Wiley, New York, 1963, pp. 17–37]. This estimator is a semiparametric estimator based on nonparametric spectral estimators. We derive the second order Edgeworth expansion of the distribution of β ^ . Then it is shown that the second order asymptotic properties are independent of the bandwidth choice for residual spectral estimator, which implies that β ^ has the same rate of convergence as in regular parametric estimation. This is a sharp contrast with the general semiparametric estimation theory. We also examine the second order Gaussian efficiency of β ^ . Numerical studies are given to confirm the theoretical results.

A new procedure for calculating periodic response factors based on frequency domain regression method

The radiant time series method (RTSM) takes advantages of the fact that design cooling load calculations are based on steady periodic excitations. The main difference between the RTSM and the other cooling load calculation methods is that the periodic response factors of the RTSM are restricted to calculating the conduction heat gain through building elements under periodic outdoor conditions, which simplifies the computational procedure significantly. It is vital to have a reliable method or procedure to accurately calculate the periodic response factors of various types of walls and roofs. In this study, a procedure, based on the frequency-domain regression (FDR) method, is developed to directly and accurately calculate the outside, across and inside periodic response factors of a multilayer wall or roof from its geometric and thermal properties. At first, a polynomial s-transfer function is established from the frequency characteristics of the wall or roof using the FDR method. The periodic response factors are then generated from the poles and residues of the polynomial s-transfer function. Computational tests show that the FDR method provides an accurate and hopefully better alternative procedure to calculate periodic response factors. Using this procedure, the periodic response factors of various representative wall and roof types are calculated and compared with those calculated by other conventional methods. Some results, particularly of the periodic response factors whose CTF coefficients tabulated in the ASHRAE Handbook are inaccurate, are presented and evaluated.

Bootstrap techniques in semiparametric estimation methods for ARFIMA models: A comparison study

This paper considers different bootstrap procedures for investigating the estimation of the fractional parameter d in a particular case of long memory processes, i.e. for ARFIMA models withd in (0.0, 0.5). We propose two bootstrap techniques to deal with semiparametric estimation methods of d. One approach consists of the local bootstrap method for time frequency initially suggested for the ARMA case by Paparoditis and Politis (1999), and the other consists of the bootstrapping in the residuals of the frequency-domain regression equation. Through Monte Carlo simulation, these alternative bootstrap methods are compared, based on the mean and the mean square error of the estimators, with the well-known parametric and nonparametric bootstrap techniques for time series models.

A procedure for calculating transient thermal load through multilayer cylindrical structures

This paper presents the transmission matrix and frequency characteristics of transient heat conduction through a multilayer cylindrical structure. The frequency-domain regression method [Appl. Thermal Eng. 21 (6) (2000) 683] is introduced to estimate some simple polynomial s-transfer functions from the frequency characteristics and further calculate the transient heat flow of the cylindrical structure including its thermal response factors and Z-conduction transfer function coefficients. The calculation examples and comparisons have fully demonstrated that the developed procedure is not only easy and simple, but also has high computation accuracy and efficiency.