Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Decision making under uncertainty in energy systems: state of the art

Decision making under uncertainty in energy systems: state of the art The energy system studies include a wide range of issues from short term (e.g. real-time, hourly, daily and weekly operating decisions) to long term horizons (e.g. planning or policy making). The decision making chain is fed by input parameters which are usually subject to uncertainties. The art of dealing with uncertainties has been developed in various directions and has recently become a focal point of interest. In this paper, a new standard classification of uncertainty modeling techniques for decision making process is proposed. These methods are introduced and compared along with demonstrating their strengths and weaknesses. The promising lines of future researches are explored in the shadow of a comprehensive overview of the past and present applications. The possibility of using the novel concept of Z-numbers is introduced for the first time. Keywords: Fuzzy arithmetic, info-gap decision theory, probabilistic modeling, robust optimization, interval based analysis, Z-number. 1. Introduction The uncertainty handling has been one of the main concerns of the decision makers (including governors, engineers, managers, and scientists) for many years [1]. Most of the decisions to be made by energy sector decision makers are subject to a significant level of data uncertainty [2]. The uncertain parameters in power system studies can be generally classified into two different categories including (see Fig.1): Corresponding author Email address: University of Tehran, Tehran, Iran, Tel :(Office) +98(21) 66121431 Fax : +98(21) 66124160, North kargar Street, Tehran, Iran (Alireza Soroudi) Preprint submitted to Renewable & Sustainable Energy Reviews November 26, 2019 arXiv:1911.10905v1 [eess.SY] 25 Nov 2019 • Technical parameters: these parameters are generally categorized in two main classes, namely: topological parameters and operational parameters. The topological pa- rameters are those related to network topologies like failure or forced outage of lines, generators or metering devices and etc. The operational parameters are tied with operating decisions like demand or generation values in power systems. • Economical parameters: the parameters which affect the economical indices fall in this category. Microeconomics investigates the decisions of smaller business sectors like aggregators, domestic or industrial consumers while macroeconomics focuses on entire power system industry. For example, uncertainty in fuel supply, costs of production, business taxes, labor are raw materials are analyzed in microeconomics. On the other hand, the issues like regulation or deregulation, environmental policies, economic growth, unemployment rates, gross domestic product (GDP) and inter- est rates are analyzed in macroeconomics. All of these parameters are subject to uncertainties and should be correctly addressed in economical studies. There are various uncertainty handling tools developed for dealing with the aforemen- tioned uncertain parameters as depicted in Fig. 2. The main difference between these methods is in line with the technique used for describing the uncertainty of input pa- rameters. The similarity of them is that all of them try to quantify the effect of input parameters on model’s outputs. These models are described as follows: • Probabilistic approach: one of the earliest works in stochastic programming was done by Dantzig in 1955 [3]. It is assumed that the input parameters of the model are random variables with a known probability density function (PDF). • Possibilistic approach: the fuzzy arithmetic was introduced by Lotfi A. Zadeh in 1965 [4]. The input parameters of the model are described using the membership function (MF) of input parameters. • Hybrid possibilistic-probabilistic approaches: both random and possibilistic param- eters are present in the model. 2 • Information gap decision theory: it was first proposed by Yakov Ben-Haim [5] in 1980. In this method, no PDF or membership function is available for input param- eters. It is based on the difference between what is known and what is vital to be known by quantification of severe lack of information in decision making process. • Robust optimization: it was first proposed by Soyster [6] in 1973. The uncertainty sets are used for describing the uncertainty of input parameters. Using this tech- nique, the obtained decisions remain optimal for the worst-case realization of the uncertain parameter within a given set. • Interval analysis: it was introduced by Ramon E. Moore in 1966 [7]. It is assumed that the uncertain parameters are taking value from a known interval. It is somehow similar to the probabilistic modeling with a uniform PDF. This method finds the bounds of output variables. This paper is to provide a summary of recent techniques used for uncertainty modeling in power system applications. It offers a vision obtained from a relatively large number of previous works. This review serves as a road map to those interested in uncertainty modeling tools in power system studies to find the less explored research areas by standing on the shoulders of giants. The rest of this paper is set out as follows: section 2 presents the Probabilistic ap- proach, the possibilistic methodology is introduced in section 3, the hybrid possibilistic- probabilistic approach is described in section 4, the info-gap decision theory is explained in section 5, the robust optimization technique is described in section 6. Section 7 presents the interval analysis approach. Section 9 describes the promising lines of future researches. Finally, section 10 summarizes the findings of this work. 2. Probabilistic approach In the probabilistic approach, a multivariate function, namely y, y = f(Z) is available. Z is a vector of the form Z = [z , ..., z ], in which z to z are random parameters with 1 m 1 m known PDFs while the PDF of Z is tried to be identified. Three probabilistic uncertainty modeling techniques are described as follows: 3 2.1. Monte Carlo Simulation (MCS) The Monte Carlo simulation is carried out in following steps [8]. It is assumed that the z are uncertain parameters. A sample, z , is generated for each input parameter z , using i i e e e its PDF. The value of y as the outcome variable, is calculated using y = f(Z ) where e e e Z = [z , ..., z ]. The procedure is repeated for a number of iterations, N . Finally, MC 1 m the outcomes are analyzed using statistic criteria, histograms, confidence intervals and etc. There are some methods for reducing the computational burden of MCS like Latin Hypercube Sampling (LHS) [9]. 2.2. Point estimate method The point estimate method (PEM) acts based on the concept of moments of uncertain input parameters. In a problem with n uncertain parameters, the major steps are as follows [10]: Step.1 Set E(Y ) = 0, E(Y ) = 0 and k = 1. Step.2 Determine the locations and probabilities of concentrations, ǫ and P , respec- k,i k,i tively as follows: 1 M (z ) 1 M (z ) 3 k 3 k i+1 ǫ = + (−1) n + ( ) (1) k,i 3 3 2 σ 2 σ z z k k k,3−i P = (−1) q (2) k,i M (z ) 1 3 k 2n n + ( ) 2 σ where M (z ) is the third moment of parameter z . 3 k k Step.3 Determine the concentration points z , as given below. k,i z = µ + ǫ × σ , i = 1, 2 (3) k,i z k,i z k i where, µ and σ are mean and standard deviation of z , respectively. z z k k k Step.4 Calculate the f for both z , as: k,i Z = [z , z , ..., z , ..., z ], i = 1, 2 (4) 1 2 k,i n 4 2 Step.5 Calculate E(Y ) and E(Y ) using: E(Y ) = E(Y ) + P f(z , z , ..., z , ..., z ) (5) k,i 1 2 k,i n i=1 2 2 2 E(Y ) = E(Y ) + P f (z , z , ..., z , ..., z ) k,i 1 2 k,i n i=1 Step.6 k = k + 1 if k < n then go to Step. 2; otherwise continue. Step.7 Calculate the mean and the standard deviation as: µ = E(Y ) (6) 2 2 σ = E(Y ) − E (Y ) 2.3. Scenario based decision making A scenario is defined as a probable realization of an uncertain set of parameters. A list of scenarios are generated using the PDF of each uncertain parameter, Z . The expected value of output variable, y, is calculated as follows: y = π × f(Z ) (7) s s s∈Ω th where π = 1 and π is the probability of s scenario. s s s∈Ω If the number of scenarios are large then it is needed to obtain a small set of scenarios representing the original one. The purpose is to select a small set, Ω , with the cardinality of N , from the original set, Ω [11]. A reasonable trade off must be respected between Ω J the loss of the information and decreasing the computational burden [2]. The scenario reduction technique is carried out via the following steps [12, 13]: step. 1 Construct the probability distance matrix containing the distance between each pair of scenarios c(s, s´) step. 2 Select the fist scenario s as follows: ( ) s = arg min π c(s, s ) (8) 1 s s ∈Ω s∈Ω Ω = {s } , Ω = Ω − Ω S 1 J J S 5 step. 3 Select the next scenario for Ω set, as follows:     ′′ s = arg min π min c(s, s ) (9) n s ′ ′′ s ∈Ω s ∈Ω ∪{s}  s∈Ω −{s } Ω = Ω ∪ {s } , Ω = Ω − Ω S S n J J S step. 4 If the cardinality of Ω is sufficient then go to step 2; else continue. step. 5 Add the probability of each non-selected scenario to its closest scenario in the selected set, End. More details can be found in [2]. 3. Possibilistic approach Since the introduction of fuzzy set theory this technique has been used in many power system fields [14]. Suppose y = f(x , . . . , x ) is in hand and X vector contains the un- 1 n certain input parameters described using their associated membership functions. Various membership functions can be used to formulate the degree of membership of a specific uncertain parameter depending on the expert’s opinion. Regardless of the membership function’s shape the questions is “how to determine the MF of y if MFs of X are known?”. The α-cut method can provide an answer to this question [15]. For a given fuzzy set A in U, the crisp set A contains all individuals of U with membership value, A, greater than or equal to α, as calculated in (10). A = {x ∈ U | µ (x) ≥ α} (10) α α α A = (A , A ) The α-cut of each uncertain parameter, x , is determined using (10), then the α-cut of y, y , is calculated as follows: α α α y = (y , y¯ ) (11) α α α y = (min f(X ), max f(X )) (12) α α X X α α α X = (X , X ) (13) 6 α α α α In each α-cut, the upper bound of y , y¯ , and the lower bound of y , y , are maximized and minimized respectively. The final step is defuzzification. The process of translating a fuzzy number to a crisp one is called defuzzification [15]. Many defuzzification tech- niques are available such as maximum defuzzification technique, the centroid method [16], weighted average defuzzification technique and etc. 4. Hybrid possibilistic-probabilistic approach Sometimes, the decision maker is faced with a multivariate objective function, y = f(X, Z), where both possibilistic uncertain parameters (X) and probabilistic uncertain ones (Z) exist. To deal with such cases some methods are developed which are decsribed next. 4.1. Possibilistic-Monte Carlo approach The following steps describe the mixed possibilistic-Monte Carlo approach [17]: • Step.1 : For each z ∈ Z, generate a value using its PDF, z α e α e • Step.2 : Calculate (y¯ ) and (y ) as follows: α e e α (y ) = min f(Z , X ) (14) α e e α (y¯ ) = max f(Z , X ) (15) α α α X = (X , X ) (16) These steps are repeated to obtain the statistical data of the parameters of the output’s MF such as PDF or expected values. 4.2. Possibilistic-scenario based approach The following steps describe this approach [18]: • Step.1 : Generate the scenario set describing the behavior of Z, Ω • Step.2 : Reduce the original scenario set to a small set, Ω 7 α α • Step.3 : Calculate (y ) and (y ) as follows: α α y = min π × f(Z , X ) (17) s s s∈Ω α α y = max π × f(Z , X ) (18) s s s∈Ω X = (X , X ) (19) • Step.4 : Deffuzzify the y. 5. Information Gap Decision Theory The Information Gap Decision Theory (IGDT) is a method to describe the uncertain- ties which can not be described using PDF of MF due to the lack of sufficient information. It is used to make robust decisions against sever uncertainty of input parameters. In IGDT, the robustness is defined as the immunity of satisfaction of a predefined constraint [5]. The constraint satisfaction is defined based on the requirement of the decision maker. It may be defined as the maximum load which a bridge can tolerate or the maximum risk that the decision maker can accept or even the minimum revenue a decision maker is willing to achieve. Here, the main goal is not only optimizing the objective function [19]. Instead, the algorithm tries to find the best possible solution which maximum robustness against the probable forcasting errors. The constraint satisfaction is defined as not violating a predefined critical limit, ζ, for a given cost function, f(x, d), as follows: f(x, d) ≤ ζ (20) H(x, d) = 0 (21) G(x, d) ≥ 0 (22) where, x is the input parameter and d is the vector of decision variables. H and G are the equality and inequality constraints, respectively. The uncertainty of parameters in IGDT method, is usually defined as the envelope 8 bound model [20], as follows: x ∈ U(α, x˜) (23) x − x˜ U(α, x˜) = ≤ α x˜ where, α is the uncertainty level of parameter x, x˜ is the forcasted value of x and U(α, x˜) is the set of all values of x whose deviation from x˜ will never be more than αx˜. The decision maker does not know the values of x and α. ¯ ¯ The robustness of a decision d based on the requirement ζ, αˆ(d, ζ), is defined as the maximum value of α at which the decision maker is sure that the required constraints are always satisfied [5], as follows: αˆ(d, ζ) = max α (24) S.t : Constraints The decision making policy is defined as finding the decision variables, d, which maximizes the robustness, as : max αˆ(d, ζ) (25) ∀x ∈ U(α, x˜) (26) f(x, d) ≤ ζ (27) H(x, d) = 0 (28) G(x, d) ≥ 0 (29) 6. Robust optimization The concept of robust optimization (RO) was first introduced by Soyster [6]. It’s a new approach for solving optimization problems affected by uncertainty specially in case of lack of full information on the nature of uncertainty [21]. The concept of robust optimization is described as follows: consider a function like z = f(x, y) which is linear in x and non-linear in y. The values of x are subject to uncertainty while the values of y are known. In robust optimization, it is assumed that no specified PDF is in hand for describing the uncertain parameter x. The uncertainty of x is modeled with an uncertainty set x ∈ U(x), where 9 U(x) is a set that parameter x can take value from it. The maximization of z = f(x, y) can be formulated via (30) to (31). max z = f(x, z) (30) x ∈ U(x) (31) Since the value of z is linear with respect to x, it can be reformulated as follows: max z (32) z ≤ f(x˜, z) (33) f(x˜, y) = A(y) ∗ x˜ + g(y) (34) x˜ ∈ U(x) = {x| |x − x¯| ≤ xˆ} (35) where x˜, x¯, xˆ are the uncertain value, predicted value and maximum possible deviation of variable x from xˆ, respectively. The robust optimization seeks a solution which not only maximizes the objective function z but also insures the decision maker that if there exist some prediction error about the values of x, the z remains optimum with high probability [22]. To do this, a robust counterpart version of the problem is constructed and solved. The robust counterpart of (31) is defined as follows: max z (36) z ≤ f(x, z) (37) f(x, y) = A(y) ∗ x¯ + g(y) − max a (y) ∗ xˆ ∗ w (38) i i i w ≤ Γ (39) 0 ≤ w ≤ 1 (40) Based on (36), two nested optimization problems are to be solved. The equations (38) to (39) are linear with respect to w and has a dual form as follows: min[Γβ + ξ ] (41) ξ ,β β + ξ ≥ a (y) ∗ xˆ i i i 10 Inserting the (41) into (36) gives : max z (42) y,ξ ,β z ≤ f(x, z) (43) f(x, y) = A(y) ∗ x¯ + g(y) − Γβ − ξ (44) β + ξ ≥ A(y ) ∗ xˆ (45) i i i There are some softwares developed for solving the robust optimization based problems [23]. 7. Interval analysis In this method, the range of values for each uncertain input parameter is defined and it can be represented by an interval. Suppose a multivariate function of the form f = (x , ..., x ) and lb ≤ x ≤ ub where lb , ub are the lower and upper bounds of 1 n i i i i i uncertain parameter x . The goal is finding the lower and upper bounds of objective function f. There are some softwares developed for solving the interval analysis based problems [24]. 8. Applications Context serves to demonstrate the applications of the aforementioned uncertainty modeling techniques. The applications are widely categorized into several fields, as given in Table 1. The summaries of uncertainty modeling attributes are provided in Table 2. • DG impact assessment • Plug-in hybrid electric vehicle (PHEV): (e.g. exploitation of plug in hybrid electric vehicles) • Assessment of available transfer capability (ATC) • Renewable energy (operation and planning)(e.g. hydro power generation manage- ment ) 11 • Load flow/optimal power flow calculations (e.g. probabilistic load flow, fuzzy load flow.) • Reliability evaluation (e.g. reliability-oriented distribution network reconfiguration) • Distribution network operation and planning (e.g. phase balancing, cost-benefit analysis of distribution automation) • Transmission/Generation planning, operation and control: (e.q. self-scheduling of gencos, fault location scheme, dynamic economic dispatch, maintenance scheduling, determination of pilot points for zonal voltage control, small-signal stability ) • State estimation • Electricity market (e.g. real time demand side management, bidding strategy,energy hub management and electricity procurement strategy.) 9. Promising lines of future researches The future trends in uncertainty modeling to investigate and further explore are sum- marized as follows: 9.1. Exploring new uncertain parameters With the increasingly revolutionary changes in power system’s regulatory framework and developing technologies the uncertainty in input data of decision making procedures is increased. These uncertain environment include financial, societal/governmental (the ongoing government policy and the future potential incentive for the renewable energy), environmental (carbon emission and global warming issue) and technical (communication and information architecture in smart grids, demand response, PHEV, energy hubs, smart building) uncertainties, risk preferences in the investment models, fuel prices and market regulations, renewable energy sources and competition among suppliers. 9.2. Enhancing the existing techniques • Reduce the computational burden specially when applied to large scale power sys- tems and real-time applications 12 Z Z Z Z d b c d 2 2 2 2 −(x−μ) −(x−μ) −(x−μ) −(x−μ) 1 1 x − a x − d 2 2 2 2 2σ 2σ 2σ 2σ √ √ P rob = A e = [ e + e + e ] b − a c − d σ 2π σ 2π a a b c (46) G(P rob) = µ (P rob) (47) • Choosing the appropriate uncertainty handling technique • Hybridizing the existing techniques to better describe the uncertain environment • Using the heuristic methods to soften the computation procedures 9.3. Exploring the new uncertainty handling methods The taxonomy of the uncertainty modeling methods in past, present and future is as depicted in Fig.3. In 2011, Zadeh introduced a new class of uncertain numbers called “Z-numbers” [25]. The Z-numbers are expressed as a pair in form of Z = (A, B), in which, A, B are restrictions describing the behavior of Z. A is usually a fuzzy set while B describes the certainty degree. The certainty degree may be expressed as a PDF or a fuzzy set. In this context, Z = {x|x ∈ A with certainty degree equal to B}. In classic fuzzy numbers decision maker just has A and it is quit sure that Z belongs to A. However in Z-numbers, Z is described using the set A with a certainty (reliability) degree of information called B. Examples for Z-numbers are provided in Table 3. The normal PDF, as a function of µ, σ , is a reasonable choice for modeling the ran- domness of the load variable. In order to disambiguate this concept, a simple two-bus network is used as shown in Fig.4. The value of load can be described in various way as described in Table 4. For example, we are almost certain (set B ) that the demand value in a given bus (Z number) is low (set A ) as depicted in Fig.5. The probability that the load value is low can be calculated as (46). In (46), G(P rob) indicates the degree to which P rob belongs to A . Now, the information of Z-number expressed as L = (A , B ) 1 1 2 for load parameter is represented as a possibility distribution (G(P rob)) over the space of probability distributions (various values of µ, σ ). 13 10. Conclusion This paper proposed a standard classification of uncertainty handling methods along with the promising lines of future researches. The possibility of using Z-numbers for un- certainty modeling of load values was introduced for the first time. The assessed method- ologies include probabilistic, possibilistic, hybrid methods, robust optimization, interval based analysis as well as Z-numbers. These models are compared and their strength and shortcomings are investigated. Based on the proposed comprehensive classification, it is deduced that each method is suitable for a specific type of uncertainty. The severity of uncertainty dictates choosing the appropriate uncertainty modeling technique. Addition- ally, according to the carried out taxonomy of the methodologies, it was revealed that some research areas are still remained untouched. References [1] N. Attoh-Okine, B. Ayyub, Applied research in uncertainty modeling and analysis, Vol. 20, Springer Verlag, 2005. [2] A. J. Conejo, M. Carrion, J. M. Morales, Decision Making Under Uncertainty in Electricity Markets, Springer, New York, 2010. [3] G. Dantzig, Linear programming under uncertainty, Stochastic Programming (2011) 1–11. [4] L. Zadeh, Fuzzy sets, Information and control 8 (3) (1965) 338–353. [5] Y. B. Haim, Info-Gap Decision Theory (Second Edition), Academic Press, California, [6] A. L. Soyster, Convex programming with set-inclusive constraints and applications to inexact linear programming, Journal of Operations Research 21 (2) (1973) 1154– [7] R. Moore, Interval analysis, Prentice-Hall, Englewood Cliff, New Jersey, 1966. 14 [8] M. H. Kalos, P. A. Whitlock, Monte Carlo Methods, WILEY-VCH Verlag GmbH and Co. KGaA, 2004. [9] H. Yu, C. Chung, K. Wong, H. Lee, J. Zhang, Probabilistic load flow evaluation with hybrid latin hypercube sampling and cholesky decomposition, IEEE Transactions on Power Systems, 24 (2) (2009) 661 –667. [10] H. P. Hong, An efficient point estimate method for probabilistic analysis, Reliability Engineering and System Safety 59 (3) (1998) 261 – 267. [11] J. Morales, S. Pineda, A. Conejo, M. Carrion, Scenario reduction for futures market trading in electricity markets, IEEE Trans. on Power Sys., 24 (2) (2009) 878 –888. [12] S. Pineda, A. Conejo, Scenario reduction for risk-averse electricity trading, Genera- tion, Transmission Distribution, IET 4 (6) (2010) 694 –705. [13] T. Amraee, A. Soroudi, A. Ranjbar, Probabilistic determination of pilot points for zonal voltage control, Generation, Transmission Distribution, IET 6 (1) (2012) 1 –10. [14] J. Momoh, X. Ma, K. Tomsovic, Overview and literature survey of fuzzy set theory in power systems, IEEE Transactions on Power Systems, 10 (3) (1995) 1676 –1690. [15] H. Zhang, D. Liu (Eds.), Fuzzy Modeling and Fuzzy Control, Birkhuser, 2006. [16] T. Ross (Ed.), Fuzzy logic with engineering applications, Wiley, 2004. [17] A. Soroudi, M. Ehsan, A possibilisticprobabilistic tool for evaluating the impact of stochastic renewable and controllable power generation on energy losses in distribu- tion networksa case study, Renewable and Sustainable Energy Reviews 15 (1) (2011) 794 – 800. [18] A. Soroudi, Possibilistic-scenario model for dg impact assessment on distribution networks in an uncertain environment, IEEE Transactions on Power Systems, 27 (3) (2012) 1283 – 1293. 15 [19] A. Soroudi, M. Ehsan, Igdt based robust decision making tool for dnos in load pro- curement under severe uncertainty, IEEE Transactions on Power Systems, PP (99) (2013) 1–10. [20] B. Mohammadi-Ivatloo, H. Zareipour, N. Amjady, M. Ehsan, Application of information-gap decision theory to risk-constrained self-scheduling of gencos, IEEE Transactions on Power Systems, PP (99) (2012) 1–10. [21] A. Ben-Tal, L. E. Ghaoui, A. Nemirovski, Robust Optimization, 1st Edition, Prince- ton Series in Applied Mathematics, 2009. [22] D. Bertsimas, M. Sim, The price of robustness, Journal of Operations Research 52 (1) (2004) 35–53. [23] M. S. Joel Goh, Robust Optimization Made Easy with ROME, Vol. 59, Informs, Dordrecht, 2011, p. 973985. [24] S. Rump, INTLAB - INTerval LABoratory, in: T. Csendes (Ed.), Develop- ments in Reliable Computing, Kluwer Academic Publishers, Dordrecht, 1999, pp. 77–104, http://www.ti3.tu-harburg.de/rump/. [25] L. A. Zadeh, A note on z-numbers, Information Sciences 181 (14) (2011) 2923 – 2932. [26] W. El-Khattam, Y. Hegazy, M. Salama, Investigating distributed generation systems performance using monte carlo simulation, IEEE Transactions on Power Systems, 21 (2) (2006) 524 – 532. [27] A. Soroudi, R. Caire, N. Hadjsaid, M. Ehsan, Probabilistic dynamic multi-objective model for renewable and non-renewable distributed generation planning, Generation, Transmission Distribution, IET 5 (11) (2011) 1173 –1182. [28] C.-L. Su, Stochastic evaluation of voltages in distribution networks with distributed generation using detailed distribution operation models, IEEE Transactions on Power Systems, 25 (2) (2010) 786 –795. 16 [29] Z. Liu, F. Wen, G. Ledwich, Optimal siting and sizing of distributed generators in distribution systems considering uncertainties, IEEE Transactions on Power Delivery, 26 (4) (2011) 2541 –2551. [30] V. Martins, C. Borges, Active distribution network integrated planning incorporating distributed generation and load response uncertainties, IEEE Transactions on Power Systems, 26 (4) (2011) 2164 –2172. [31] A. Soroudi, M. Ehsan, R. Caire, N. Hadjsaid, Possibilistic evaluation of distributed generations impacts on distribution networks, IEEE Transactions on Power Systems, 26 (4) (2011) 2293 –2301. [32] A. Lojowska, D. Kurowicka, G. Papaefthymiou, L. van der Sluis, Stochastic modeling of power demand due to evs using copula, IEEE Transactions on Power Systems, PP (99) (2012) 1. [33] M. Pantos, Exploitation of electric-drive vehicles in electricity markets, IEEE Trans- actions on Power Systems, 27 (2) (2012) 682 –694. [34] A. Hajimiragha, C. Canizares, M. Fowler, S. Moazeni, A. Elkamel, A robust op- timization approach for planning the transition to plug-in hybrid electric vehicles, IEEE Transactions on Power Systems, 26 (4) (2011) 2264 –2274. [35] A. B. Rodrigues, M. G. Da Silva, Probabilistic assessment of available transfer capa- bility based on monte carlo method with sequential simulation, IEEE Transactions on Power Systems, 22 (1) (2007) 484 –492. [36] C.-L. Su, C.-N. Lu, Two-point estimate method for quantifying transfer capability uncertainty, IEEE Transactions on Power Systems, 20 (2) (2005) 573 – 579. [37] A. Khairuddin, S. Ahmed, M. Mustafa, A. Zin, H. Ahmad, A novel method for atc computations in a large-scale power system, IEEE Transactions on Power Systems, 19 (2) (2004) 1150 – 1158. 17 [38] S. Bu, W. Du, H. Wang, Z. Chen, L. Xiao, H. Li, Probabilistic analysis of small-signal stability of large-scale power systems as affected by penetration of wind generation, IEEE Transactions on Power Systems, 27 (2) (2012) 762 –770. [39] Q. Wang, Y. Guan, J. Wang, A chance-constrained two-stage stochastic program for unit commitment with uncertain wind power output, IEEE Transactions on Power Systems, 27 (1) (2012) 206 –215. [40] A. Soroudi, M. Aien, M. Ehsan, A probabilistic modeling of photo voltaic modules and wind power generation impact on distribution networks, Systems Journal, IEEE 6 (2) (2012) 254 –259. [41] L. Escudero, J. de la Fuente, C. Garcia, F. Prieto, Hydropower generation man- agement under uncertainty via scenario analysis and parallel computation, IEEE Transactions on Power Systems, 11 (2) (1996) 683 –689. [42] Y. Makarov, P. Etingov, J. Ma, Z. Huang, K. Subbarao, Incorporating uncertainty of wind power generation forecast into power system operation, dispatch, and unit commitment procedures, IEEE Transactions on Sustainable Energy 2 (4) (2011) 433 –442. [43] A. Saber, G. Venayagamoorthy, Resource scheduling under uncertainty in a smart grid with renewables and plug-in vehicles, Systems Journal, IEEE 6 (1) (2012) 103 –109. [44] C. Baslis, A. Bakirtzis, Mid-term stochastic scheduling of a price-maker hydro pro- ducer with pumped storage, IEEE Transactions on Power Systems, 26 (4) (2011) 1856 –1865. [45] S. Rebennack, B. Flach, M. Pereira, P. Pardalos, Stochastic hydro-thermal scheduling under emissions constraints, IEEE Transactions on Power Systems, 27 (1) (2012) 58 –68. [46] B. Venkatesh, P. Yu, H. Gooi, D. Choling, Fuzzy milp unit commitment incorporating wind generators, IEEE Transactions on Power Systems, 23 (4) (2008) 1738 –1746. 18 [47] R. Jiang, J. Wang, Y. Guan, Robust unit commitment with wind power and pumped storage hydro, IEEE Transactions on Power Systems, PP (99) (2011) 1. [48] H. Zhang, P. Li, Probabilistic analysis for optimal power flow under uncertainty, Generation, Transmission Distribution, IET 4 (5) (2010) 553 –561. [49] X. Li, Y. Li, S. Zhang, Analysis of probabilistic optimal power flow taking account of the variation of load power, IEEE Transactions on Power Systems, 23 (3) (2008) 992 –999. [50] E. M. Gouveia, M. A. Matos, Symmetric ac fuzzy power flow model, European Journal of Operational Research 197 (3) (2009) 1012 – 1018. [51] M. Matos, E. Gouveia, The fuzzy power flow revisited, IEEE Transactions on Power Systems, 23 (1) (2008) 213 –218. [52] A. Romero, H. Zini, G. Ratta? and, R. Dib, Harmonic load-flow approach based on the possibility theory, Generation, Transmission Distribution, IET 5 (4) (2011) 393 –404. [53] Z. Wang, F. Alvarado, Interval arithmetic in power flow analysis, IEEE Transactions on Power Systems, 7 (3) (1992) 1341 –1349. [54] R. Karki, P. Hu, R. Billinton, Reliability evaluation considering wind and hydro power coordination, IEEE Transactions on Power Systems, 25 (2) (2010) 685 –693. [55] L. Wu, M. Shahidehpour, Y. Fu, Security-constrained generation and transmission outage scheduling with uncertainties, IEEE Transactions on Power Systems, 25 (3) (2010) 1674 –1685. [56] A. Papavasiliou, S. Oren, R. O’Neill, Reserve requirements for wind power integra- tion: A scenario-based stochastic programming framework, IEEE Transactions on Power Systems, 26 (4) (2011) 2197 –2206. [57] L. Wu, M. Shahidehpour, T. Li, Stochastic security-constrained unit commitment, IEEE Transactions on Power Systems, 22 (2) (2007) 800 –811. 19 [58] H. Ge, S. Asgarpoor, Reliability evaluation of equipment and substations with fuzzy markov processes, IEEE Transactions on Power Systems, 25 (3) (2010) 1319 –1328. [59] P. Zhang, W. Li, S. Wang, Reliability-oriented distribution network reconfiguration considering uncertainties of data by interval analysis, International Journal of Elec- trical Power & Energy Systems 34 (1) (2012) 138 – 144. [60] C. yang Li, X. Chen, X. shan Yi, J. yong Tao, Interval-valued reliability analysis of multi-state systems,, Reliability, IEEE Transactions on 60 (1) (2011) 323 –330. [61] P. Carvalho, L. Ferreira, F. Lobo, L. Barruncho, Distribution network expansion planning under uncertainty: a hedging algorithm in an evolutionary approach, IEEE Transactions on Power Delivery, 15 (1) (2000) 412 –416. [62] I. Ramirez-Rosado, J. Dominguez-Navarro, Possibilistic model based on fuzzy sets for the multiobjective optimal planning of electric power distribution networks, IEEE Transactions on Power Systems, 19 (4) (2004) 1801 – 1810. [63] H. E. Shaalan, R. P. Broadwater, Using interval mathematics in cost-benefit analysis of distribution automation, Electric Power Systems Research 27 (2) (1993) 145 – 152. [64] A. Kazerooni, J. Mutale, Transmission network planning under security and envi- ronmental constraints, IEEE Transactions on Power Systems, 25 (2) (2010) 1169 –1178. [65] S. Kazempour, A. Conejo, Strategic generation investment under uncertainty via benders decomposition, IEEE Transactions on Power Systems, 27 (1) (2012) 424 –432. [66] T. Li, M. Shahidehpour, Risk-constrained generation asset arbitrage in power sys- tems,, IEEE Transactions on Power Systems, 22 (3) (2007) 1330 –1339. [67] A. Saric, A. Stankovic, Model uncertainty in security assessment of power systems,, IEEE Transactions on Power Systems, 20 (3) (2005) 1398 – 1407. [68] H. Yamin, Fuzzy self-scheduling for gencos, IEEE Transactions on Power Systems, 20 (1) (2005) 503 – 505. 20 [69] P. Attaviriyanupap, H. Kita, E. Tanaka, J. Hasegawa, A fuzzy-optimization approach to dynamic economic dispatch considering uncertainties, IEEE Transactions on Power Systems, 19 (3) (2004) 1299 – 1307. [70] D. K. Mohanta, P. K. Sadhu, R. Chakrabarti, Fuzzy reliability evaluation of captive power plant maintenance scheduling incorporating uncertain forced outage rate and load representation, Electric Power Systems Research 72 (1) (2004) 73 – 84. [71] Y.-Y. Hong, P.-H. Chen, Genetic-based underfrequency load shedding in a stand- alone power system considering fuzzy loads, Power Delivery, IEEE Transactions on 27 (1) (2012) 87 –95. [72] Z. Xu, M. Liu, G. Yang, N. Li, Application of interval analysis and evidence theory to fault location, Electric Power Applications, IET 3 (1) (2009) 77 –84. [73] Y. Zhu, Y. Li, G. Huang, Planning municipal-scale energy systems under functional interval uncertainties, Renewable Energy 39 (1) (2012) 71 – 84. [74] Y. Wang, Q. Xia, C. Kang, Unit commitment with volatile node injections by using interval optimization, IEEE Transactions on Power Systems, 26 (3) (2011) 1705 – [75] A. Street, F. Oliveira, J. Arroyo, Contingency-constrained unit commitment with n - k security criterion: A robust optimization approach, IEEE Transactions on Power Systems, 26 (3) (2011) 1581 –1590. [76] R. Singh, B. Pal, R. Vinter, Measurement placement in distribution system state estimation, IEEE Transactions on Power Systems, 24 (2) (2009) 668 –675. [77] E. Caro, J.-M. Morales, A. Conejo, R. Minguez, Calculation of measurement correla- tions using point estimate, IEEE Transactions on Power Delivery, 25 (4) (2010) 2095 –2103. [78] K.-R. Shih, S.-J. Huang, Application of a robust algorithm for dynamic state esti- mation of a power system, IEEE Transactions on Power Systems, 17 (1) (2002) 141 –147. 21 [79] A. Saric, R. Ciric, Integrated fuzzy state estimation and load flow analysis in distri- bution networks, IEEE Transactions on Power Delivery, 18 (2) (2003) 571 – 578. [80] Y. Wang, W. Li, P. Zhang, B. Wang, J. Lu, Reliability analysis of phasor mea- surement unit considering data uncertainty, IEEE Transactions on Power Systems, PP (99) (2012) 1 –8. [81] C. Rakpenthai, S. Uatrongjit, S. Premrudeeprechacharn, State estimation of power system considering network parameter uncertainty based on parametric interval lin- ear systems,, IEEE Transactions on Power Systems, 27 (1) (2012) 305 –313. [82] M. Perninge, F. Lindskog, L. Soder, Importance sampling of injected powers for electric power system security analysis, IEEE Transactions on Power Systems, 27 (1) (2012) 3 –11. [83] P. Duenas, J. Reneses, J. Barquin, Dealing with multi-factor uncertainty in electricity markets by combining monte carlo simulation with spatial interpolation techniques, Generation, Transmission Distribution, IET 5 (3) (2011) 323 –331. [84] A. Saric, A. Stankovic, An application of interval analysis and optimization to electric energy markets, IEEE Transactions on Power Systems, 21 (2) (2006) 515 – 523. [85] A. Conejo, J. Morales, L. Baringo, Real-time demand response model, IEEE Trans- actions on Smart Grid 1 (3) (2010) 236 –242. [86] M.-P. Cheong, D. Berleant, G. Sheble, Information gap decision theory as a tool for strategic bidding in competitive electricity markets, in: Probabilistic Methods Applied to Power Systems, 2004 International Conference on, 2004, pp. 421 –426. [87] A. Parisio, C. D. Vecchio, A. Vaccaro, A robust optimization approach to energy hub management, International Journal of Electrical Power & Energy Systems 42 (1) (2012) 98 – 104. [88] K. Zare, M. Moghaddam, M. Sheikh-El-Eslami, Risk-based electricity procurement for large consumers, IEEE Transactions on Power Systems, 26 (4) (2011) 1826 –1835. 22 List of Figure Captions: • Figure 1. General classification of uncertain parameters in energy system studies • Figure 2. Uncertainty modeling tools • Figure 3. Uncertainty modeling trends: past, present and future • Figure 4. Simple two-bus illustrative network • Figure 5. Concept of Z-number 23 Table 1: Summaries of uncertainty modeling applications Probabilistic Possibilistic Hybrid Interval RO IDGT Applications MC PEM Scenario DG units [26, 27] [28] [29, 30] [31] [17, 18] PHEV [29, 32] [33] [33] [34] Available transfer capability (ATC) [35] [36] [37] Renewable energy (operation and planning) [38, 39] [40] [41, 42, 43, 44, 45] [46] [47] Load flow/Optimal power flow [48] [49] [50, 51, 52] [53] Reliability evaluation [54, 55] [56] [57, 55] [58] [59, 60] Distribution operation and planning [27] [61] [62] [63] Transmission/generation planning and operation/control [64] [65, 66, 13, 67] [68, 69, 70, 71] [69] [72, 73, 74] [75] State estimation [76] [77] [78] [79, 80] [79] [81] Electricity market [82, 83] [68] [84] [85, 86, 87] [88, 86] Unexplored research directions —— 24 Table 2: Summaries of uncertainty modeling attributes Input repre- Method Output attributes Advantages Disadvantages sentation Computationally ex- pensive, needs a large Statistics like expecta- Probabilistic PDF Easy to implement amount of historic tion, variance, etc. data, approximate result Converting linguistic Complex implementa- Possibilistic MF MF knowledge to numer- tion ical values Membership function Dealing with both Computationally ex- Hybrid MF & PDF with probabilistic pa- uncertainty types si- pensive rameters multaneously Decision variables Forecasted Useful for severe un- IGDT satisfying the require- Too conservative values certainties ments Robust Op- Controlled conserva- Useful when just an Difficult to use in non- Intervals timization tiveness interval is available linear models The correlations among intervals Interval Useful when just an Intervals Bounds of the outputs are neglected this Analysis interval is available would make it too conservative Table 3: Examples for Z-numbers Parameter A B Demand value High Very sure Wind speed Weibul PDF Normally Voltage magnitude Uniform distribution in [0.951.05] In most cases 25 Table 4: Describing the load values as Z-numbers A B Load Not sure L = (A , B ) 1 1 Low Almost certain L = (A , B ) 1 2 Quit sure L = (A , B ) 1 3 Not sure L = (A , B ) 2 1 Medium Almost certain L = (A , B ) 2 2 Quit sure L = (A , B ) 2 3 Not sure L = (A , B ) 3 1 High Almost certain L = (A , B ) 3 2 Quit sure L = (A , B ) 3 3 26 Uncertain Parameters !"#$%& yz{|}~ ’()*+,-./01<=>?@ABCDEF 23456789:;GHIJKLMNOP ¡¢£⁄¥ƒ §¤'“«‹›fifl–†‡·¶• klmnQRSTUVWXYZ[ ‚„”»…‰¿`´ˆ˜ ƪŁØŒºæı opqrstuvwx \]^_‘abcdefghij¯˘˙ łøœß ¨˚¸˝˛ˇ— Figure 1: General classification of uncertain parameters in energy system studies "#$%&’( RSTUVW ‰¿`´ˆ˜¯˘˙¨ ¢£⁄¥ƒ§¤'“«‹›fiXYZ[\]^_‘abcde IGDT ˚¸˝˛ˇ— fghijklmnopqr =>?@ABCD )*+,-./012stuvwxyz{|} ¡EFGHIfl–†‡·¶•‚„”»… 3456789:;<~ JKLMNOPQ Figure 2: Uncertainty modeling tools ıłø ƪŁØŒ ºæ Intervals Real Value Numbers Figure 3: Uncertainty modeling trends: past, present and future 28 V <δ Xj E <0 P+Qj Figure 4: Simple two-bus illustrative network efghij <=>?@ABopqrstuv klmn Ad Ab Ac 9;8 œß !"#$% xyz{|} CDEFGHIJ Q[ N B Bw BK Figure 5: Concept of Z-number 29 1 0.8 0.6 0.4 0.2 a a a min L U max Predicted Value of A Universe of discourse Membership Degree http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Electrical Engineering and Systems Science arXiv (Cornell University)

Decision making under uncertainty in energy systems: state of the art

Loading next page...
 
/lp/arxiv-cornell-university/decision-making-under-uncertainty-in-energy-systems-state-of-the-art-krwD0e0g2P

References (91)

ISSN
1364-0321
eISSN
ARCH-3348
DOI
10.1016/j.rser.2013.08.039
Publisher site
See Article on Publisher Site

Abstract

The energy system studies include a wide range of issues from short term (e.g. real-time, hourly, daily and weekly operating decisions) to long term horizons (e.g. planning or policy making). The decision making chain is fed by input parameters which are usually subject to uncertainties. The art of dealing with uncertainties has been developed in various directions and has recently become a focal point of interest. In this paper, a new standard classification of uncertainty modeling techniques for decision making process is proposed. These methods are introduced and compared along with demonstrating their strengths and weaknesses. The promising lines of future researches are explored in the shadow of a comprehensive overview of the past and present applications. The possibility of using the novel concept of Z-numbers is introduced for the first time. Keywords: Fuzzy arithmetic, info-gap decision theory, probabilistic modeling, robust optimization, interval based analysis, Z-number. 1. Introduction The uncertainty handling has been one of the main concerns of the decision makers (including governors, engineers, managers, and scientists) for many years [1]. Most of the decisions to be made by energy sector decision makers are subject to a significant level of data uncertainty [2]. The uncertain parameters in power system studies can be generally classified into two different categories including (see Fig.1): Corresponding author Email address: University of Tehran, Tehran, Iran, Tel :(Office) +98(21) 66121431 Fax : +98(21) 66124160, North kargar Street, Tehran, Iran (Alireza Soroudi) Preprint submitted to Renewable & Sustainable Energy Reviews November 26, 2019 arXiv:1911.10905v1 [eess.SY] 25 Nov 2019 • Technical parameters: these parameters are generally categorized in two main classes, namely: topological parameters and operational parameters. The topological pa- rameters are those related to network topologies like failure or forced outage of lines, generators or metering devices and etc. The operational parameters are tied with operating decisions like demand or generation values in power systems. • Economical parameters: the parameters which affect the economical indices fall in this category. Microeconomics investigates the decisions of smaller business sectors like aggregators, domestic or industrial consumers while macroeconomics focuses on entire power system industry. For example, uncertainty in fuel supply, costs of production, business taxes, labor are raw materials are analyzed in microeconomics. On the other hand, the issues like regulation or deregulation, environmental policies, economic growth, unemployment rates, gross domestic product (GDP) and inter- est rates are analyzed in macroeconomics. All of these parameters are subject to uncertainties and should be correctly addressed in economical studies. There are various uncertainty handling tools developed for dealing with the aforemen- tioned uncertain parameters as depicted in Fig. 2. The main difference between these methods is in line with the technique used for describing the uncertainty of input pa- rameters. The similarity of them is that all of them try to quantify the effect of input parameters on model’s outputs. These models are described as follows: • Probabilistic approach: one of the earliest works in stochastic programming was done by Dantzig in 1955 [3]. It is assumed that the input parameters of the model are random variables with a known probability density function (PDF). • Possibilistic approach: the fuzzy arithmetic was introduced by Lotfi A. Zadeh in 1965 [4]. The input parameters of the model are described using the membership function (MF) of input parameters. • Hybrid possibilistic-probabilistic approaches: both random and possibilistic param- eters are present in the model. 2 • Information gap decision theory: it was first proposed by Yakov Ben-Haim [5] in 1980. In this method, no PDF or membership function is available for input param- eters. It is based on the difference between what is known and what is vital to be known by quantification of severe lack of information in decision making process. • Robust optimization: it was first proposed by Soyster [6] in 1973. The uncertainty sets are used for describing the uncertainty of input parameters. Using this tech- nique, the obtained decisions remain optimal for the worst-case realization of the uncertain parameter within a given set. • Interval analysis: it was introduced by Ramon E. Moore in 1966 [7]. It is assumed that the uncertain parameters are taking value from a known interval. It is somehow similar to the probabilistic modeling with a uniform PDF. This method finds the bounds of output variables. This paper is to provide a summary of recent techniques used for uncertainty modeling in power system applications. It offers a vision obtained from a relatively large number of previous works. This review serves as a road map to those interested in uncertainty modeling tools in power system studies to find the less explored research areas by standing on the shoulders of giants. The rest of this paper is set out as follows: section 2 presents the Probabilistic ap- proach, the possibilistic methodology is introduced in section 3, the hybrid possibilistic- probabilistic approach is described in section 4, the info-gap decision theory is explained in section 5, the robust optimization technique is described in section 6. Section 7 presents the interval analysis approach. Section 9 describes the promising lines of future researches. Finally, section 10 summarizes the findings of this work. 2. Probabilistic approach In the probabilistic approach, a multivariate function, namely y, y = f(Z) is available. Z is a vector of the form Z = [z , ..., z ], in which z to z are random parameters with 1 m 1 m known PDFs while the PDF of Z is tried to be identified. Three probabilistic uncertainty modeling techniques are described as follows: 3 2.1. Monte Carlo Simulation (MCS) The Monte Carlo simulation is carried out in following steps [8]. It is assumed that the z are uncertain parameters. A sample, z , is generated for each input parameter z , using i i e e e its PDF. The value of y as the outcome variable, is calculated using y = f(Z ) where e e e Z = [z , ..., z ]. The procedure is repeated for a number of iterations, N . Finally, MC 1 m the outcomes are analyzed using statistic criteria, histograms, confidence intervals and etc. There are some methods for reducing the computational burden of MCS like Latin Hypercube Sampling (LHS) [9]. 2.2. Point estimate method The point estimate method (PEM) acts based on the concept of moments of uncertain input parameters. In a problem with n uncertain parameters, the major steps are as follows [10]: Step.1 Set E(Y ) = 0, E(Y ) = 0 and k = 1. Step.2 Determine the locations and probabilities of concentrations, ǫ and P , respec- k,i k,i tively as follows: 1 M (z ) 1 M (z ) 3 k 3 k i+1 ǫ = + (−1) n + ( ) (1) k,i 3 3 2 σ 2 σ z z k k k,3−i P = (−1) q (2) k,i M (z ) 1 3 k 2n n + ( ) 2 σ where M (z ) is the third moment of parameter z . 3 k k Step.3 Determine the concentration points z , as given below. k,i z = µ + ǫ × σ , i = 1, 2 (3) k,i z k,i z k i where, µ and σ are mean and standard deviation of z , respectively. z z k k k Step.4 Calculate the f for both z , as: k,i Z = [z , z , ..., z , ..., z ], i = 1, 2 (4) 1 2 k,i n 4 2 Step.5 Calculate E(Y ) and E(Y ) using: E(Y ) = E(Y ) + P f(z , z , ..., z , ..., z ) (5) k,i 1 2 k,i n i=1 2 2 2 E(Y ) = E(Y ) + P f (z , z , ..., z , ..., z ) k,i 1 2 k,i n i=1 Step.6 k = k + 1 if k < n then go to Step. 2; otherwise continue. Step.7 Calculate the mean and the standard deviation as: µ = E(Y ) (6) 2 2 σ = E(Y ) − E (Y ) 2.3. Scenario based decision making A scenario is defined as a probable realization of an uncertain set of parameters. A list of scenarios are generated using the PDF of each uncertain parameter, Z . The expected value of output variable, y, is calculated as follows: y = π × f(Z ) (7) s s s∈Ω th where π = 1 and π is the probability of s scenario. s s s∈Ω If the number of scenarios are large then it is needed to obtain a small set of scenarios representing the original one. The purpose is to select a small set, Ω , with the cardinality of N , from the original set, Ω [11]. A reasonable trade off must be respected between Ω J the loss of the information and decreasing the computational burden [2]. The scenario reduction technique is carried out via the following steps [12, 13]: step. 1 Construct the probability distance matrix containing the distance between each pair of scenarios c(s, s´) step. 2 Select the fist scenario s as follows: ( ) s = arg min π c(s, s ) (8) 1 s s ∈Ω s∈Ω Ω = {s } , Ω = Ω − Ω S 1 J J S 5 step. 3 Select the next scenario for Ω set, as follows:     ′′ s = arg min π min c(s, s ) (9) n s ′ ′′ s ∈Ω s ∈Ω ∪{s}  s∈Ω −{s } Ω = Ω ∪ {s } , Ω = Ω − Ω S S n J J S step. 4 If the cardinality of Ω is sufficient then go to step 2; else continue. step. 5 Add the probability of each non-selected scenario to its closest scenario in the selected set, End. More details can be found in [2]. 3. Possibilistic approach Since the introduction of fuzzy set theory this technique has been used in many power system fields [14]. Suppose y = f(x , . . . , x ) is in hand and X vector contains the un- 1 n certain input parameters described using their associated membership functions. Various membership functions can be used to formulate the degree of membership of a specific uncertain parameter depending on the expert’s opinion. Regardless of the membership function’s shape the questions is “how to determine the MF of y if MFs of X are known?”. The α-cut method can provide an answer to this question [15]. For a given fuzzy set A in U, the crisp set A contains all individuals of U with membership value, A, greater than or equal to α, as calculated in (10). A = {x ∈ U | µ (x) ≥ α} (10) α α α A = (A , A ) The α-cut of each uncertain parameter, x , is determined using (10), then the α-cut of y, y , is calculated as follows: α α α y = (y , y¯ ) (11) α α α y = (min f(X ), max f(X )) (12) α α X X α α α X = (X , X ) (13) 6 α α α α In each α-cut, the upper bound of y , y¯ , and the lower bound of y , y , are maximized and minimized respectively. The final step is defuzzification. The process of translating a fuzzy number to a crisp one is called defuzzification [15]. Many defuzzification tech- niques are available such as maximum defuzzification technique, the centroid method [16], weighted average defuzzification technique and etc. 4. Hybrid possibilistic-probabilistic approach Sometimes, the decision maker is faced with a multivariate objective function, y = f(X, Z), where both possibilistic uncertain parameters (X) and probabilistic uncertain ones (Z) exist. To deal with such cases some methods are developed which are decsribed next. 4.1. Possibilistic-Monte Carlo approach The following steps describe the mixed possibilistic-Monte Carlo approach [17]: • Step.1 : For each z ∈ Z, generate a value using its PDF, z α e α e • Step.2 : Calculate (y¯ ) and (y ) as follows: α e e α (y ) = min f(Z , X ) (14) α e e α (y¯ ) = max f(Z , X ) (15) α α α X = (X , X ) (16) These steps are repeated to obtain the statistical data of the parameters of the output’s MF such as PDF or expected values. 4.2. Possibilistic-scenario based approach The following steps describe this approach [18]: • Step.1 : Generate the scenario set describing the behavior of Z, Ω • Step.2 : Reduce the original scenario set to a small set, Ω 7 α α • Step.3 : Calculate (y ) and (y ) as follows: α α y = min π × f(Z , X ) (17) s s s∈Ω α α y = max π × f(Z , X ) (18) s s s∈Ω X = (X , X ) (19) • Step.4 : Deffuzzify the y. 5. Information Gap Decision Theory The Information Gap Decision Theory (IGDT) is a method to describe the uncertain- ties which can not be described using PDF of MF due to the lack of sufficient information. It is used to make robust decisions against sever uncertainty of input parameters. In IGDT, the robustness is defined as the immunity of satisfaction of a predefined constraint [5]. The constraint satisfaction is defined based on the requirement of the decision maker. It may be defined as the maximum load which a bridge can tolerate or the maximum risk that the decision maker can accept or even the minimum revenue a decision maker is willing to achieve. Here, the main goal is not only optimizing the objective function [19]. Instead, the algorithm tries to find the best possible solution which maximum robustness against the probable forcasting errors. The constraint satisfaction is defined as not violating a predefined critical limit, ζ, for a given cost function, f(x, d), as follows: f(x, d) ≤ ζ (20) H(x, d) = 0 (21) G(x, d) ≥ 0 (22) where, x is the input parameter and d is the vector of decision variables. H and G are the equality and inequality constraints, respectively. The uncertainty of parameters in IGDT method, is usually defined as the envelope 8 bound model [20], as follows: x ∈ U(α, x˜) (23) x − x˜ U(α, x˜) = ≤ α x˜ where, α is the uncertainty level of parameter x, x˜ is the forcasted value of x and U(α, x˜) is the set of all values of x whose deviation from x˜ will never be more than αx˜. The decision maker does not know the values of x and α. ¯ ¯ The robustness of a decision d based on the requirement ζ, αˆ(d, ζ), is defined as the maximum value of α at which the decision maker is sure that the required constraints are always satisfied [5], as follows: αˆ(d, ζ) = max α (24) S.t : Constraints The decision making policy is defined as finding the decision variables, d, which maximizes the robustness, as : max αˆ(d, ζ) (25) ∀x ∈ U(α, x˜) (26) f(x, d) ≤ ζ (27) H(x, d) = 0 (28) G(x, d) ≥ 0 (29) 6. Robust optimization The concept of robust optimization (RO) was first introduced by Soyster [6]. It’s a new approach for solving optimization problems affected by uncertainty specially in case of lack of full information on the nature of uncertainty [21]. The concept of robust optimization is described as follows: consider a function like z = f(x, y) which is linear in x and non-linear in y. The values of x are subject to uncertainty while the values of y are known. In robust optimization, it is assumed that no specified PDF is in hand for describing the uncertain parameter x. The uncertainty of x is modeled with an uncertainty set x ∈ U(x), where 9 U(x) is a set that parameter x can take value from it. The maximization of z = f(x, y) can be formulated via (30) to (31). max z = f(x, z) (30) x ∈ U(x) (31) Since the value of z is linear with respect to x, it can be reformulated as follows: max z (32) z ≤ f(x˜, z) (33) f(x˜, y) = A(y) ∗ x˜ + g(y) (34) x˜ ∈ U(x) = {x| |x − x¯| ≤ xˆ} (35) where x˜, x¯, xˆ are the uncertain value, predicted value and maximum possible deviation of variable x from xˆ, respectively. The robust optimization seeks a solution which not only maximizes the objective function z but also insures the decision maker that if there exist some prediction error about the values of x, the z remains optimum with high probability [22]. To do this, a robust counterpart version of the problem is constructed and solved. The robust counterpart of (31) is defined as follows: max z (36) z ≤ f(x, z) (37) f(x, y) = A(y) ∗ x¯ + g(y) − max a (y) ∗ xˆ ∗ w (38) i i i w ≤ Γ (39) 0 ≤ w ≤ 1 (40) Based on (36), two nested optimization problems are to be solved. The equations (38) to (39) are linear with respect to w and has a dual form as follows: min[Γβ + ξ ] (41) ξ ,β β + ξ ≥ a (y) ∗ xˆ i i i 10 Inserting the (41) into (36) gives : max z (42) y,ξ ,β z ≤ f(x, z) (43) f(x, y) = A(y) ∗ x¯ + g(y) − Γβ − ξ (44) β + ξ ≥ A(y ) ∗ xˆ (45) i i i There are some softwares developed for solving the robust optimization based problems [23]. 7. Interval analysis In this method, the range of values for each uncertain input parameter is defined and it can be represented by an interval. Suppose a multivariate function of the form f = (x , ..., x ) and lb ≤ x ≤ ub where lb , ub are the lower and upper bounds of 1 n i i i i i uncertain parameter x . The goal is finding the lower and upper bounds of objective function f. There are some softwares developed for solving the interval analysis based problems [24]. 8. Applications Context serves to demonstrate the applications of the aforementioned uncertainty modeling techniques. The applications are widely categorized into several fields, as given in Table 1. The summaries of uncertainty modeling attributes are provided in Table 2. • DG impact assessment • Plug-in hybrid electric vehicle (PHEV): (e.g. exploitation of plug in hybrid electric vehicles) • Assessment of available transfer capability (ATC) • Renewable energy (operation and planning)(e.g. hydro power generation manage- ment ) 11 • Load flow/optimal power flow calculations (e.g. probabilistic load flow, fuzzy load flow.) • Reliability evaluation (e.g. reliability-oriented distribution network reconfiguration) • Distribution network operation and planning (e.g. phase balancing, cost-benefit analysis of distribution automation) • Transmission/Generation planning, operation and control: (e.q. self-scheduling of gencos, fault location scheme, dynamic economic dispatch, maintenance scheduling, determination of pilot points for zonal voltage control, small-signal stability ) • State estimation • Electricity market (e.g. real time demand side management, bidding strategy,energy hub management and electricity procurement strategy.) 9. Promising lines of future researches The future trends in uncertainty modeling to investigate and further explore are sum- marized as follows: 9.1. Exploring new uncertain parameters With the increasingly revolutionary changes in power system’s regulatory framework and developing technologies the uncertainty in input data of decision making procedures is increased. These uncertain environment include financial, societal/governmental (the ongoing government policy and the future potential incentive for the renewable energy), environmental (carbon emission and global warming issue) and technical (communication and information architecture in smart grids, demand response, PHEV, energy hubs, smart building) uncertainties, risk preferences in the investment models, fuel prices and market regulations, renewable energy sources and competition among suppliers. 9.2. Enhancing the existing techniques • Reduce the computational burden specially when applied to large scale power sys- tems and real-time applications 12 Z Z Z Z d b c d 2 2 2 2 −(x−μ) −(x−μ) −(x−μ) −(x−μ) 1 1 x − a x − d 2 2 2 2 2σ 2σ 2σ 2σ √ √ P rob = A e = [ e + e + e ] b − a c − d σ 2π σ 2π a a b c (46) G(P rob) = µ (P rob) (47) • Choosing the appropriate uncertainty handling technique • Hybridizing the existing techniques to better describe the uncertain environment • Using the heuristic methods to soften the computation procedures 9.3. Exploring the new uncertainty handling methods The taxonomy of the uncertainty modeling methods in past, present and future is as depicted in Fig.3. In 2011, Zadeh introduced a new class of uncertain numbers called “Z-numbers” [25]. The Z-numbers are expressed as a pair in form of Z = (A, B), in which, A, B are restrictions describing the behavior of Z. A is usually a fuzzy set while B describes the certainty degree. The certainty degree may be expressed as a PDF or a fuzzy set. In this context, Z = {x|x ∈ A with certainty degree equal to B}. In classic fuzzy numbers decision maker just has A and it is quit sure that Z belongs to A. However in Z-numbers, Z is described using the set A with a certainty (reliability) degree of information called B. Examples for Z-numbers are provided in Table 3. The normal PDF, as a function of µ, σ , is a reasonable choice for modeling the ran- domness of the load variable. In order to disambiguate this concept, a simple two-bus network is used as shown in Fig.4. The value of load can be described in various way as described in Table 4. For example, we are almost certain (set B ) that the demand value in a given bus (Z number) is low (set A ) as depicted in Fig.5. The probability that the load value is low can be calculated as (46). In (46), G(P rob) indicates the degree to which P rob belongs to A . Now, the information of Z-number expressed as L = (A , B ) 1 1 2 for load parameter is represented as a possibility distribution (G(P rob)) over the space of probability distributions (various values of µ, σ ). 13 10. Conclusion This paper proposed a standard classification of uncertainty handling methods along with the promising lines of future researches. The possibility of using Z-numbers for un- certainty modeling of load values was introduced for the first time. The assessed method- ologies include probabilistic, possibilistic, hybrid methods, robust optimization, interval based analysis as well as Z-numbers. These models are compared and their strength and shortcomings are investigated. Based on the proposed comprehensive classification, it is deduced that each method is suitable for a specific type of uncertainty. The severity of uncertainty dictates choosing the appropriate uncertainty modeling technique. Addition- ally, according to the carried out taxonomy of the methodologies, it was revealed that some research areas are still remained untouched. References [1] N. Attoh-Okine, B. Ayyub, Applied research in uncertainty modeling and analysis, Vol. 20, Springer Verlag, 2005. [2] A. J. Conejo, M. Carrion, J. M. Morales, Decision Making Under Uncertainty in Electricity Markets, Springer, New York, 2010. [3] G. Dantzig, Linear programming under uncertainty, Stochastic Programming (2011) 1–11. [4] L. Zadeh, Fuzzy sets, Information and control 8 (3) (1965) 338–353. [5] Y. B. Haim, Info-Gap Decision Theory (Second Edition), Academic Press, California, [6] A. L. Soyster, Convex programming with set-inclusive constraints and applications to inexact linear programming, Journal of Operations Research 21 (2) (1973) 1154– [7] R. Moore, Interval analysis, Prentice-Hall, Englewood Cliff, New Jersey, 1966. 14 [8] M. H. Kalos, P. A. Whitlock, Monte Carlo Methods, WILEY-VCH Verlag GmbH and Co. KGaA, 2004. [9] H. Yu, C. Chung, K. Wong, H. Lee, J. Zhang, Probabilistic load flow evaluation with hybrid latin hypercube sampling and cholesky decomposition, IEEE Transactions on Power Systems, 24 (2) (2009) 661 –667. [10] H. P. Hong, An efficient point estimate method for probabilistic analysis, Reliability Engineering and System Safety 59 (3) (1998) 261 – 267. [11] J. Morales, S. Pineda, A. Conejo, M. Carrion, Scenario reduction for futures market trading in electricity markets, IEEE Trans. on Power Sys., 24 (2) (2009) 878 –888. [12] S. Pineda, A. Conejo, Scenario reduction for risk-averse electricity trading, Genera- tion, Transmission Distribution, IET 4 (6) (2010) 694 –705. [13] T. Amraee, A. Soroudi, A. Ranjbar, Probabilistic determination of pilot points for zonal voltage control, Generation, Transmission Distribution, IET 6 (1) (2012) 1 –10. [14] J. Momoh, X. Ma, K. Tomsovic, Overview and literature survey of fuzzy set theory in power systems, IEEE Transactions on Power Systems, 10 (3) (1995) 1676 –1690. [15] H. Zhang, D. Liu (Eds.), Fuzzy Modeling and Fuzzy Control, Birkhuser, 2006. [16] T. Ross (Ed.), Fuzzy logic with engineering applications, Wiley, 2004. [17] A. Soroudi, M. Ehsan, A possibilisticprobabilistic tool for evaluating the impact of stochastic renewable and controllable power generation on energy losses in distribu- tion networksa case study, Renewable and Sustainable Energy Reviews 15 (1) (2011) 794 – 800. [18] A. Soroudi, Possibilistic-scenario model for dg impact assessment on distribution networks in an uncertain environment, IEEE Transactions on Power Systems, 27 (3) (2012) 1283 – 1293. 15 [19] A. Soroudi, M. Ehsan, Igdt based robust decision making tool for dnos in load pro- curement under severe uncertainty, IEEE Transactions on Power Systems, PP (99) (2013) 1–10. [20] B. Mohammadi-Ivatloo, H. Zareipour, N. Amjady, M. Ehsan, Application of information-gap decision theory to risk-constrained self-scheduling of gencos, IEEE Transactions on Power Systems, PP (99) (2012) 1–10. [21] A. Ben-Tal, L. E. Ghaoui, A. Nemirovski, Robust Optimization, 1st Edition, Prince- ton Series in Applied Mathematics, 2009. [22] D. Bertsimas, M. Sim, The price of robustness, Journal of Operations Research 52 (1) (2004) 35–53. [23] M. S. Joel Goh, Robust Optimization Made Easy with ROME, Vol. 59, Informs, Dordrecht, 2011, p. 973985. [24] S. Rump, INTLAB - INTerval LABoratory, in: T. Csendes (Ed.), Develop- ments in Reliable Computing, Kluwer Academic Publishers, Dordrecht, 1999, pp. 77–104, http://www.ti3.tu-harburg.de/rump/. [25] L. A. Zadeh, A note on z-numbers, Information Sciences 181 (14) (2011) 2923 – 2932. [26] W. El-Khattam, Y. Hegazy, M. Salama, Investigating distributed generation systems performance using monte carlo simulation, IEEE Transactions on Power Systems, 21 (2) (2006) 524 – 532. [27] A. Soroudi, R. Caire, N. Hadjsaid, M. Ehsan, Probabilistic dynamic multi-objective model for renewable and non-renewable distributed generation planning, Generation, Transmission Distribution, IET 5 (11) (2011) 1173 –1182. [28] C.-L. Su, Stochastic evaluation of voltages in distribution networks with distributed generation using detailed distribution operation models, IEEE Transactions on Power Systems, 25 (2) (2010) 786 –795. 16 [29] Z. Liu, F. Wen, G. Ledwich, Optimal siting and sizing of distributed generators in distribution systems considering uncertainties, IEEE Transactions on Power Delivery, 26 (4) (2011) 2541 –2551. [30] V. Martins, C. Borges, Active distribution network integrated planning incorporating distributed generation and load response uncertainties, IEEE Transactions on Power Systems, 26 (4) (2011) 2164 –2172. [31] A. Soroudi, M. Ehsan, R. Caire, N. Hadjsaid, Possibilistic evaluation of distributed generations impacts on distribution networks, IEEE Transactions on Power Systems, 26 (4) (2011) 2293 –2301. [32] A. Lojowska, D. Kurowicka, G. Papaefthymiou, L. van der Sluis, Stochastic modeling of power demand due to evs using copula, IEEE Transactions on Power Systems, PP (99) (2012) 1. [33] M. Pantos, Exploitation of electric-drive vehicles in electricity markets, IEEE Trans- actions on Power Systems, 27 (2) (2012) 682 –694. [34] A. Hajimiragha, C. Canizares, M. Fowler, S. Moazeni, A. Elkamel, A robust op- timization approach for planning the transition to plug-in hybrid electric vehicles, IEEE Transactions on Power Systems, 26 (4) (2011) 2264 –2274. [35] A. B. Rodrigues, M. G. Da Silva, Probabilistic assessment of available transfer capa- bility based on monte carlo method with sequential simulation, IEEE Transactions on Power Systems, 22 (1) (2007) 484 –492. [36] C.-L. Su, C.-N. Lu, Two-point estimate method for quantifying transfer capability uncertainty, IEEE Transactions on Power Systems, 20 (2) (2005) 573 – 579. [37] A. Khairuddin, S. Ahmed, M. Mustafa, A. Zin, H. Ahmad, A novel method for atc computations in a large-scale power system, IEEE Transactions on Power Systems, 19 (2) (2004) 1150 – 1158. 17 [38] S. Bu, W. Du, H. Wang, Z. Chen, L. Xiao, H. Li, Probabilistic analysis of small-signal stability of large-scale power systems as affected by penetration of wind generation, IEEE Transactions on Power Systems, 27 (2) (2012) 762 –770. [39] Q. Wang, Y. Guan, J. Wang, A chance-constrained two-stage stochastic program for unit commitment with uncertain wind power output, IEEE Transactions on Power Systems, 27 (1) (2012) 206 –215. [40] A. Soroudi, M. Aien, M. Ehsan, A probabilistic modeling of photo voltaic modules and wind power generation impact on distribution networks, Systems Journal, IEEE 6 (2) (2012) 254 –259. [41] L. Escudero, J. de la Fuente, C. Garcia, F. Prieto, Hydropower generation man- agement under uncertainty via scenario analysis and parallel computation, IEEE Transactions on Power Systems, 11 (2) (1996) 683 –689. [42] Y. Makarov, P. Etingov, J. Ma, Z. Huang, K. Subbarao, Incorporating uncertainty of wind power generation forecast into power system operation, dispatch, and unit commitment procedures, IEEE Transactions on Sustainable Energy 2 (4) (2011) 433 –442. [43] A. Saber, G. Venayagamoorthy, Resource scheduling under uncertainty in a smart grid with renewables and plug-in vehicles, Systems Journal, IEEE 6 (1) (2012) 103 –109. [44] C. Baslis, A. Bakirtzis, Mid-term stochastic scheduling of a price-maker hydro pro- ducer with pumped storage, IEEE Transactions on Power Systems, 26 (4) (2011) 1856 –1865. [45] S. Rebennack, B. Flach, M. Pereira, P. Pardalos, Stochastic hydro-thermal scheduling under emissions constraints, IEEE Transactions on Power Systems, 27 (1) (2012) 58 –68. [46] B. Venkatesh, P. Yu, H. Gooi, D. Choling, Fuzzy milp unit commitment incorporating wind generators, IEEE Transactions on Power Systems, 23 (4) (2008) 1738 –1746. 18 [47] R. Jiang, J. Wang, Y. Guan, Robust unit commitment with wind power and pumped storage hydro, IEEE Transactions on Power Systems, PP (99) (2011) 1. [48] H. Zhang, P. Li, Probabilistic analysis for optimal power flow under uncertainty, Generation, Transmission Distribution, IET 4 (5) (2010) 553 –561. [49] X. Li, Y. Li, S. Zhang, Analysis of probabilistic optimal power flow taking account of the variation of load power, IEEE Transactions on Power Systems, 23 (3) (2008) 992 –999. [50] E. M. Gouveia, M. A. Matos, Symmetric ac fuzzy power flow model, European Journal of Operational Research 197 (3) (2009) 1012 – 1018. [51] M. Matos, E. Gouveia, The fuzzy power flow revisited, IEEE Transactions on Power Systems, 23 (1) (2008) 213 –218. [52] A. Romero, H. Zini, G. Ratta? and, R. Dib, Harmonic load-flow approach based on the possibility theory, Generation, Transmission Distribution, IET 5 (4) (2011) 393 –404. [53] Z. Wang, F. Alvarado, Interval arithmetic in power flow analysis, IEEE Transactions on Power Systems, 7 (3) (1992) 1341 –1349. [54] R. Karki, P. Hu, R. Billinton, Reliability evaluation considering wind and hydro power coordination, IEEE Transactions on Power Systems, 25 (2) (2010) 685 –693. [55] L. Wu, M. Shahidehpour, Y. Fu, Security-constrained generation and transmission outage scheduling with uncertainties, IEEE Transactions on Power Systems, 25 (3) (2010) 1674 –1685. [56] A. Papavasiliou, S. Oren, R. O’Neill, Reserve requirements for wind power integra- tion: A scenario-based stochastic programming framework, IEEE Transactions on Power Systems, 26 (4) (2011) 2197 –2206. [57] L. Wu, M. Shahidehpour, T. Li, Stochastic security-constrained unit commitment, IEEE Transactions on Power Systems, 22 (2) (2007) 800 –811. 19 [58] H. Ge, S. Asgarpoor, Reliability evaluation of equipment and substations with fuzzy markov processes, IEEE Transactions on Power Systems, 25 (3) (2010) 1319 –1328. [59] P. Zhang, W. Li, S. Wang, Reliability-oriented distribution network reconfiguration considering uncertainties of data by interval analysis, International Journal of Elec- trical Power & Energy Systems 34 (1) (2012) 138 – 144. [60] C. yang Li, X. Chen, X. shan Yi, J. yong Tao, Interval-valued reliability analysis of multi-state systems,, Reliability, IEEE Transactions on 60 (1) (2011) 323 –330. [61] P. Carvalho, L. Ferreira, F. Lobo, L. Barruncho, Distribution network expansion planning under uncertainty: a hedging algorithm in an evolutionary approach, IEEE Transactions on Power Delivery, 15 (1) (2000) 412 –416. [62] I. Ramirez-Rosado, J. Dominguez-Navarro, Possibilistic model based on fuzzy sets for the multiobjective optimal planning of electric power distribution networks, IEEE Transactions on Power Systems, 19 (4) (2004) 1801 – 1810. [63] H. E. Shaalan, R. P. Broadwater, Using interval mathematics in cost-benefit analysis of distribution automation, Electric Power Systems Research 27 (2) (1993) 145 – 152. [64] A. Kazerooni, J. Mutale, Transmission network planning under security and envi- ronmental constraints, IEEE Transactions on Power Systems, 25 (2) (2010) 1169 –1178. [65] S. Kazempour, A. Conejo, Strategic generation investment under uncertainty via benders decomposition, IEEE Transactions on Power Systems, 27 (1) (2012) 424 –432. [66] T. Li, M. Shahidehpour, Risk-constrained generation asset arbitrage in power sys- tems,, IEEE Transactions on Power Systems, 22 (3) (2007) 1330 –1339. [67] A. Saric, A. Stankovic, Model uncertainty in security assessment of power systems,, IEEE Transactions on Power Systems, 20 (3) (2005) 1398 – 1407. [68] H. Yamin, Fuzzy self-scheduling for gencos, IEEE Transactions on Power Systems, 20 (1) (2005) 503 – 505. 20 [69] P. Attaviriyanupap, H. Kita, E. Tanaka, J. Hasegawa, A fuzzy-optimization approach to dynamic economic dispatch considering uncertainties, IEEE Transactions on Power Systems, 19 (3) (2004) 1299 – 1307. [70] D. K. Mohanta, P. K. Sadhu, R. Chakrabarti, Fuzzy reliability evaluation of captive power plant maintenance scheduling incorporating uncertain forced outage rate and load representation, Electric Power Systems Research 72 (1) (2004) 73 – 84. [71] Y.-Y. Hong, P.-H. Chen, Genetic-based underfrequency load shedding in a stand- alone power system considering fuzzy loads, Power Delivery, IEEE Transactions on 27 (1) (2012) 87 –95. [72] Z. Xu, M. Liu, G. Yang, N. Li, Application of interval analysis and evidence theory to fault location, Electric Power Applications, IET 3 (1) (2009) 77 –84. [73] Y. Zhu, Y. Li, G. Huang, Planning municipal-scale energy systems under functional interval uncertainties, Renewable Energy 39 (1) (2012) 71 – 84. [74] Y. Wang, Q. Xia, C. Kang, Unit commitment with volatile node injections by using interval optimization, IEEE Transactions on Power Systems, 26 (3) (2011) 1705 – [75] A. Street, F. Oliveira, J. Arroyo, Contingency-constrained unit commitment with n - k security criterion: A robust optimization approach, IEEE Transactions on Power Systems, 26 (3) (2011) 1581 –1590. [76] R. Singh, B. Pal, R. Vinter, Measurement placement in distribution system state estimation, IEEE Transactions on Power Systems, 24 (2) (2009) 668 –675. [77] E. Caro, J.-M. Morales, A. Conejo, R. Minguez, Calculation of measurement correla- tions using point estimate, IEEE Transactions on Power Delivery, 25 (4) (2010) 2095 –2103. [78] K.-R. Shih, S.-J. Huang, Application of a robust algorithm for dynamic state esti- mation of a power system, IEEE Transactions on Power Systems, 17 (1) (2002) 141 –147. 21 [79] A. Saric, R. Ciric, Integrated fuzzy state estimation and load flow analysis in distri- bution networks, IEEE Transactions on Power Delivery, 18 (2) (2003) 571 – 578. [80] Y. Wang, W. Li, P. Zhang, B. Wang, J. Lu, Reliability analysis of phasor mea- surement unit considering data uncertainty, IEEE Transactions on Power Systems, PP (99) (2012) 1 –8. [81] C. Rakpenthai, S. Uatrongjit, S. Premrudeeprechacharn, State estimation of power system considering network parameter uncertainty based on parametric interval lin- ear systems,, IEEE Transactions on Power Systems, 27 (1) (2012) 305 –313. [82] M. Perninge, F. Lindskog, L. Soder, Importance sampling of injected powers for electric power system security analysis, IEEE Transactions on Power Systems, 27 (1) (2012) 3 –11. [83] P. Duenas, J. Reneses, J. Barquin, Dealing with multi-factor uncertainty in electricity markets by combining monte carlo simulation with spatial interpolation techniques, Generation, Transmission Distribution, IET 5 (3) (2011) 323 –331. [84] A. Saric, A. Stankovic, An application of interval analysis and optimization to electric energy markets, IEEE Transactions on Power Systems, 21 (2) (2006) 515 – 523. [85] A. Conejo, J. Morales, L. Baringo, Real-time demand response model, IEEE Trans- actions on Smart Grid 1 (3) (2010) 236 –242. [86] M.-P. Cheong, D. Berleant, G. Sheble, Information gap decision theory as a tool for strategic bidding in competitive electricity markets, in: Probabilistic Methods Applied to Power Systems, 2004 International Conference on, 2004, pp. 421 –426. [87] A. Parisio, C. D. Vecchio, A. Vaccaro, A robust optimization approach to energy hub management, International Journal of Electrical Power & Energy Systems 42 (1) (2012) 98 – 104. [88] K. Zare, M. Moghaddam, M. Sheikh-El-Eslami, Risk-based electricity procurement for large consumers, IEEE Transactions on Power Systems, 26 (4) (2011) 1826 –1835. 22 List of Figure Captions: • Figure 1. General classification of uncertain parameters in energy system studies • Figure 2. Uncertainty modeling tools • Figure 3. Uncertainty modeling trends: past, present and future • Figure 4. Simple two-bus illustrative network • Figure 5. Concept of Z-number 23 Table 1: Summaries of uncertainty modeling applications Probabilistic Possibilistic Hybrid Interval RO IDGT Applications MC PEM Scenario DG units [26, 27] [28] [29, 30] [31] [17, 18] PHEV [29, 32] [33] [33] [34] Available transfer capability (ATC) [35] [36] [37] Renewable energy (operation and planning) [38, 39] [40] [41, 42, 43, 44, 45] [46] [47] Load flow/Optimal power flow [48] [49] [50, 51, 52] [53] Reliability evaluation [54, 55] [56] [57, 55] [58] [59, 60] Distribution operation and planning [27] [61] [62] [63] Transmission/generation planning and operation/control [64] [65, 66, 13, 67] [68, 69, 70, 71] [69] [72, 73, 74] [75] State estimation [76] [77] [78] [79, 80] [79] [81] Electricity market [82, 83] [68] [84] [85, 86, 87] [88, 86] Unexplored research directions —— 24 Table 2: Summaries of uncertainty modeling attributes Input repre- Method Output attributes Advantages Disadvantages sentation Computationally ex- pensive, needs a large Statistics like expecta- Probabilistic PDF Easy to implement amount of historic tion, variance, etc. data, approximate result Converting linguistic Complex implementa- Possibilistic MF MF knowledge to numer- tion ical values Membership function Dealing with both Computationally ex- Hybrid MF & PDF with probabilistic pa- uncertainty types si- pensive rameters multaneously Decision variables Forecasted Useful for severe un- IGDT satisfying the require- Too conservative values certainties ments Robust Op- Controlled conserva- Useful when just an Difficult to use in non- Intervals timization tiveness interval is available linear models The correlations among intervals Interval Useful when just an Intervals Bounds of the outputs are neglected this Analysis interval is available would make it too conservative Table 3: Examples for Z-numbers Parameter A B Demand value High Very sure Wind speed Weibul PDF Normally Voltage magnitude Uniform distribution in [0.951.05] In most cases 25 Table 4: Describing the load values as Z-numbers A B Load Not sure L = (A , B ) 1 1 Low Almost certain L = (A , B ) 1 2 Quit sure L = (A , B ) 1 3 Not sure L = (A , B ) 2 1 Medium Almost certain L = (A , B ) 2 2 Quit sure L = (A , B ) 2 3 Not sure L = (A , B ) 3 1 High Almost certain L = (A , B ) 3 2 Quit sure L = (A , B ) 3 3 26 Uncertain Parameters !"#$%& yz{|}~ ’()*+,-./01<=>?@ABCDEF 23456789:;GHIJKLMNOP ¡¢£⁄¥ƒ §¤'“«‹›fifl–†‡·¶• klmnQRSTUVWXYZ[ ‚„”»…‰¿`´ˆ˜ ƪŁØŒºæı opqrstuvwx \]^_‘abcdefghij¯˘˙ łøœß ¨˚¸˝˛ˇ— Figure 1: General classification of uncertain parameters in energy system studies "#$%&’( RSTUVW ‰¿`´ˆ˜¯˘˙¨ ¢£⁄¥ƒ§¤'“«‹›fiXYZ[\]^_‘abcde IGDT ˚¸˝˛ˇ— fghijklmnopqr =>?@ABCD )*+,-./012stuvwxyz{|} ¡EFGHIfl–†‡·¶•‚„”»… 3456789:;<~ JKLMNOPQ Figure 2: Uncertainty modeling tools ıłø ƪŁØŒ ºæ Intervals Real Value Numbers Figure 3: Uncertainty modeling trends: past, present and future 28 V <δ Xj E <0 P+Qj Figure 4: Simple two-bus illustrative network efghij <=>?@ABopqrstuv klmn Ad Ab Ac 9;8 œß !"#$% xyz{|} CDEFGHIJ Q[ N B Bw BK Figure 5: Concept of Z-number 29 1 0.8 0.6 0.4 0.2 a a a min L U max Predicted Value of A Universe of discourse Membership Degree

Journal

Electrical Engineering and Systems SciencearXiv (Cornell University)

Published: Nov 25, 2019

There are no references for this article.