Multidimensional parallelepiped model a new type of non-probabilistic convex model for structural uncertainty analysis

Size: px
Start display at page:

Download "Multidimensional parallelepiped model a new type of non-probabilistic convex model for structural uncertainty analysis"

Transcription

1 INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING Int. J. Numer. Meth. Engng (015) Publshed onlne n Wley Onlne Lbrary (wleyonlnelbrary.com) Multdmensonal parallelepped model a new type of non-probablstc convex model for structural uncertanty analyss C. Jang*,, Q. F. Zhang, X. Han, J. Lu and D. A. Hu State Key Laboratory of Advanced Desgn and Manufacturng for Vehcle Body, College of Mechancal and Vehcle Engneerng, Hunan Unversty, Changsha Cty, Chna SUMMARY Non-probablstc convex models need to be provded only the changng boundary of parameters rather than ther exact probablty dstrbutons; thus, such models can be appled to uncertanty analyss of complex structures when expermental nformaton s lackng. The nterval and the ellpsodal models are the two most commonly used modelng methods n the feld of non-probablstc convex modelng. However, the former can only deal wth ndependent varables, whle the latter can only deal wth dependent varables. Ths paper presents a more general non-probablstc convex model, the multdmensonal parallelepped model. Ths model can nclude the ndependent and dependent uncertan varables n a unfed framework and can effectvely deal wth complex mult-source uncertanty problems n whch dependent varables and ndependent varables coexst. For any two parameters, the concepts of the correlaton angle and the correlaton coeffcent are defned. Through the margnal ntervals of all the parameters and also ther correlaton coeffcents, a multdmensonal parallelepped can easly be bult as the uncertanty doman for parameters. Through the ntroducton of affne coordnates, the parallelepped model n the orgnal parameter space s converted to an nterval model n the affne space, thus greatly facltatng subsequent structural uncertanty analyss. The parallelepped model s appled to structural uncertanty propagaton analyss, and the response nterval of the structure s obtaned n the case of uncertan ntal parameters. Fnally, the method descrbed n ths paper was appled to several numercal examples. Copyrght 015 John Wley & Sons, Ltd. Receved 5 May 01; Revsed 14 September 014; Accepted 19 December 014 KEY WORDS: multdmensonal parallelepped model; convex model; nterval analyss; ellpsodal model; non-probablstc uncertanty; parameter correlaton 1. INTRODUCTION Uncertanty wdely exsts n practcal engneerng problems, and t s commonly related to materal propertes, loads, boundary condtons, and other factors. The probablty model, n whch the dstrbutons of structural responses are obtaned based on statstcal technques [1 5], s most wdely used to quantfy uncertantes of these types. The probablty model has become the prncpal means for dealng wth uncertanty n engneerng and has been successfully appled to varous ndustral applcatons. In the processng of probablty models, a large number of samples are used to construct precse probablty dstrbutons for uncertan parameters; such samples, however, are not always avalable or are sometmes very costly to obtan for practcal reasons. Thus, n many cases, some assumptons regardng probablty dstrbutons have to be made when usng a probablty model. Nevertheless, research ndcates that even a small devaton of probablty dstrbutons from real values may result n an extremely large error n the uncertanty analyss [6]. In the early 1990s, Ben-Ham and Elshakoff [6 9] proposed a new knd of uncertanty analyss methodology based on a non-probablstc convex model. In ths method, t s assumed that the *Correspondence to: C. Jang, State Key Laboratory of Advanced Desgn and Manufacturng for Vehcle Body, College of Mechancal and Vehcle Engneerng, Hunan Unversty, Changsha Cty, Chna E-mal: jangc@hnu.edu.cn Copyrght 015 John Wley & Sons, Ltd.

2 C. JIANG ET AL. uncertanty of the parameters belongs to a convex set; thus, the uncertanty boundary can be obtaned based on a small number of samples nstead of an exact probablty dstrbuton. In addton, optmzaton methods are generally used n ths model to obtan the response nterval and thereby measure the degree of safety of structures. Because of ts weak dependence on sample number, the convex model approach s hghly sutable for uncertanty analyss of many complex engneerng problems. There are many excellent recent research fndngs n ths area. In [9], the probablty model and the convex model were compared. In other work, an uncertan trangle was adopted to descrbe the relatonshp of three uncertanty models - the probablty model, fuzzy sets, and the convex model [10]. A seres of numercal algorthms were developed to conduct uncertanty propagaton analyss for structural statc mechancs, egenvalues, and dynamcal problems [11 13]. A new nterval analyss technque was proposed to calculate the statc and dynamc responses of structures based on a frst-order Taylor nterval expanson [14]. An error estmaton method was proposed for nterval and subnterval analyss based on a second-order truncaton model [15]. A correlaton-analyzng technque based on a non-probablstc convex model was proposed as an effcent method for the constructon of multdmensonal ellpsodal convex models, and a covarance matrx was ntroduced to descrbe the correlaton among uncertan parameters [16]. The applcatons of convex model n engneerng mechancs also nclude non-lnear bucklng analyss of a column wth uncertan ntal mperfectons [17], stablty analyss of elastc bars on uncertan foundatons [18], boundary analyss of the structural responses of beams [19], uncertanty analyss n structural number determnaton n flexble pavement desgn [0], and so on. In recent years, the convex model has been appled to the relablty analyss of structures wth uncertanty, and some work n ths area has been publshed. By ntroducng the concept of the tradtonal frst-order relablty method nto problems wth convex models, a non-probablstc relablty ndex that represents a mnmal dstance n the standard convex space was defned [1, ]; an effcent soluton algorthm was further formulated for ths relablty ndex [3]. By placng the nonprobablstc relablty ndexes as constrants, several relablty-based optmzaton desgn methods were developed [4 6]. Based on the order relaton of ntervals, a nonlnear programmng method was proposed for desgn of structures wth nterval uncertanty [7, 8]. By ntegratng the probablty model and the convex model, the relablty analyss for a mxed uncertanty problem was also nvestgated [9 3]. Currently, two convex models, the nterval and the ellpsod, are prmarly used for analyss nvolvng non-probablstc convex models. In the nterval model, the fluctuatons of a sngle varable are descrbed through ts upper and lower boundares. The uncertanty doman s a multdmensonal box. For the ellpsod model, t s assumed that the parametrc uncertanty les wthn a multdmensonal ellpsod. The degree of uncertanty and the degree of correlaton of the varables are descrbed by the sze and shape of the ellpsod. In theory, the nterval model can deal only wth problems nvolvng ndependent varables, whle the ellpsod model can deal only wth correlated parametrc problems. However, many complex engneerng problems have mult-source uncertanty arsng from materal, load, geometrcal dmenson, and so on; modelng of these problems nvolves many uncertan parameters, some of whch are ndependent, whle others may have correlatons. Therefore, analyss based only on an nterval model or an ellpsodal model may result n the excessve expanson of the uncertanty doman, further resultng n a conservatve desgn. For ths reason, we need to develop a more general non-probablstc convex model that smultaneously takes nto account the ndependence and correlatons of the varables. Such a model would be expected to deal more effectvely wth the complex mult-source uncertanty problems, thus greatly expandng the scope of applcaton of the convex model method. Ths artcle proposes a novel non-probablstc convex model, the multdmensonal parallelepped model. A quanttatve descrpton of the degree of correlaton between any two varables was realzed by ntroducng a relevant angle. The proposed model can easly handle the uncertanty problem wth the coexstence of correlaton and ndependence and thus effectvely solve the complex ssues of mult-source uncertantes. The remander of ths artcle s organzed as follows: Secton ntroduces two types of conventonal convex models; Secton 3 proposes a new non-probablstc

3 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS convex model; Secton 4 gves an affne coordnate transformaton for the suggested convex model; Secton 5 apples the model to uncertanty propagaton analyss of structures; Secton 6 conducts the numercal analyss cases; and Secton 7 states the conclusons.. CONVENTIONAL NON-PROBABILISTIC CONVEX MODELS In theory, the buldng of convex models s not unque as long as the uncertanty doman belongs to a convex set. Because of the advantages of smplcty of expresson and ease of use, the nterval model and ellpsodal model have become the two most commonly used types of models n ths feld [33, 34]. In nterval model, the varaton range of an uncertan parameter s expressed through an nterval. Only the lower and upper bounds of the range are requred to be specfed, and there s no need to know the exact probablty dstrbuton of the parameter. Assume a bounded closed nterval I.R/. The upper and lower bounds of an uncertan varable X R are X U and X L, respectvely. Then, the nterval model can be expressed as or X I.R/ D¹XjX L 6 X 6 X U º (1) X X L ;X U () In the nterval model, the mdpont of the nterval s X C D X L C X ı U, and the nterval radus s X W D X U X Lı.Forn uncertan varables, the correspondng nterval model s h X X L ; X U ; X X L ;X R ; D 1; ; :::; n : (3) The uncertanty doman usng an nterval model s a multdmensonal box. As shown n Fgure 1, the uncertan doman of the two-dmensonal problem s a rectangle wth the margnal ntervals X1 I and X I of the two parameters. A margnal nterval represents the change range for a sngle varable. In an ellpsodal model, the uncertanty doman of the n-dmensonal varables X belongs to an ellpsod X X C T M X X C 6 1 (4) n whch 3 g 11 g 1 ::: g 1n g 1 g ::: g n M D : : : : 5 g n1 g n ::: g nn (5) Fgure 1. Interval model.

4 C. JIANG ET AL. Fgure. Ellpsodal model. M s a symmetrc postve defnte matrx and s called the characterstc matrx of the ellpsodal model. It determnes the sze and orentaton of the ellpsod. For the two-dmensonal problem shown n Fgure, the uncertanty doman s an ellpse, and X1 I and X I represent the margnal ntervals of the two parameters. The ellpsodal model could deal wth the correlaton of the varables. For a dscusson on how to defne ts correlaton and precsely construct a multdmensonal ellpsodal model, one can refer to our recent work [16]. Although the nterval model and the ellpsodal model can both effectvely solve many uncertanty problems, they possess some nadequaces. In theory, the nterval model can deal wth the problem wth only ndependent varables, whle the ellpsodal model can process problems wth only dependent varables. However, many practcal engneerng problems have mult-source uncertantes related to materal, load, geometrcal dmenson, and other factors and thus have many uncertan parameters. Some parameters are ndependent, whle others mght be dependent to each other. If ether the nterval model or the ellpsodal model s used separately to conduct the analyss, t may result n the excessve expanson of the uncertanty doman, leadng to a conservatve desgn. For ths purpose, we need to develop a more general non-probablstc convex model that could take nto account both the ndependence and correlaton of the varables. Usng ths model, the mportant mult-source uncertanty problems wll be handled more easly, thus effectvely expandng the scope of applcaton of the convex model method. 3. NON-PROBABILISTIC CONVEX MODEL BASED ON A MULTIDIMENSIONAL PARALLELEPIPED In ths secton, a more general convex model, namely multdmensonal parallelepped model that could contan the ndependent and dependent uncertan varables n a unfed framework, wll be presented. As long as nformaton on the varables margnal ntervals and the correlaton between any two varables s gven, the uncertanty doman can then be constructed. In geometry, the uncertan doman that s bult n our model s a multdmensonal parallelepped, whch s stll a convex set. For better understandng, examples of the constructon of the gven convex model n two dmensons, three dmensons, and n dmensons are gven Two-dmensonal problem As shown n Fgure 3, for a two-dmensonal problem, our multdmensonal parallelepped model wll degenerate nto a parallelogram. In ths model, a quadrlateral sde s set to be parallel to the abscssa axs. In theory, the parallelogram s requred to be obtaned from expermental samples of the parameters. X1 I andx I are the respectve ranges of the two varables, namely ther margnal ntervals, and X1 W and X W are the nterval rad of the uncertan varables X 1 and X, respectvely. We can fnd that under the fxed X1 W and X W, the angle 1 actually could depct the correcton

5 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Fgure 3. Two-dmensonal parallelogram model: (a) Postve correlaton and (b) negatve correlaton. Fgure 4. Three-dmensonal parallelepped model. of the two varables, and hence, t s defned as correlaton angle n ths paper. 1 s a value n the followng 1 arctan X W X W 1 ; arctan X W X W 1 When 1 < 90 ı, X 1 and X are postvely correlated, as shown n Fgure 3(a). When 1 D 90 ı, X 1 and X are ndependent of each other, and the parallelogram model degenerates to the nterval model. When 1 >90 ı, the varables X 1 and X are n negatve correlaton, as shown n Fgure 3(b). 3.. Multdmensonal problem For a three-dmensonal problem, the uncertanty doman s a parallelepped, as shown n Fgure 4. Smlarly, one surface of the parallelepped s set to be parallel to the X-Y coordnate plane. The margnal ntervalsx1 I ;XI,andX 3 I represent the uncertanty range for each varable, whle the correlaton angles 1, 13,and 3 descrbe the dependence of any two varables. For example, when only X 1 and X have correlaton among the three varables, the uncertanty doman wll be a parallelepped, as shown n Fgure 5. For a general n-dmensonal problem, the uncertanty doman wll be a multdmensonal parallelepped. X I ; D 1; ; :::; n represent the margnal ntervals of the varables. The degree of correlaton of any two varables X and X j can be depcted by ther correlaton angle j. In addton, for convenence of descrpton, we defne the correlaton coeffcent j between the two varables as follows j D (6) X W j X W tan j (7)

6 C. JIANG ET AL. Fgure 5. Parallelepped model wth correlaton only between X 1 and X. Fgure 6. Parallelepped models under some specal correlaton cases: (a) j D 1, (b) j D 0, and(c) j D 1. n whch j arctan X W j X W ; arctan X W j X W and the range of the correlaton coeffcent j s Œ 1; 1. As shown n Fgure 6, when j D arctan X j W, the correlaton coeffcent X W j D 1, and X and X j are totally lnearly postvely correlated. When j D 90 ı, the correlaton coeffcent j D 0, and the two parameters are ndependent. When j D arctan X j W, the correlaton X W coeffcent j D 1, and the two parameters are totally lnearly negatvely correlated. From the aforementoned analyss, t can be seen that n the stuaton n whch the correlaton coeffcents of all the varables are 0, the multdmensonal parallelepped model degenerates nto the tradtonal nterval model; that s, the nterval model s actually a specal case of the parallelepped model Constructon of the uncertanty doman From the aforementoned dscusson, t can be seen that n our model, as long as the margnal ntervals of all the varables and the correlaton angles or correlaton coeffcents between any two

7 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS varables are known, a multdmensonal parallelepped that represents the uncertanty doman of the parameters can be bult. In practcal engneerng problems, the changng ranges of the varables, that s, ther margnal ntervals, are often not hard to obtan based on the experence or exstng knowledge of the problems concerned. For example, we know that gven a partcular machnng accuracy, the dmenson of a product wll le wthn the nterval of a fxed tolerance. In addton, n many cases, nformaton on the correlaton between varables can also be obtaned based on our experence. For example, n a gven structure, we know that the parameters of materals, loads, and geometrc dmensons arse from dfferent uncertanty sources and are usually not correlated wth one another. Stochastc process analyss of an ocean wave s heght and wnd speed usually shows strong correlaton between these two parameters and thus should be gven a relatvely large correlaton coeffcent. However, f we lack experence wth a partcular problem, the buldng of the uncertanty doman wll be based entrely on the samples avalable. In the followng, we provde an entrely sample-based method to create a multdmensonal parallelepped convex model for a general uncertanty problem. Assume a structure wth n-dmensonal uncertan varables X, D 1; ; :::; n, andtherearem test samples X.r/ ;rd1; ; :::; m. Then, the process of buldng the convex model s as follows: Step 1: Take any two uncertan varables X and X j, wth j. Step : Extract the values of X and X j from the sample X.r/ to obtan a two-dmensonal sample set X.r/ ;X.r/ j ;rd1; ; :::; m. Step 3: In the X -X j two-dmensonal varable space, establsh a parallelogram envelopng all the samples X.r/ ;X.r/ j ;rd1; ; :::; m, thus obtanng the margnal ntervals X I and Xj I of the varables, the correlaton angle j, and correlaton coeffcent j. Step 4: Repeat the aforementoned steps for any two uncertan varables and obtan the margnal ntervals and correlaton coeffcents of all the uncertan varables. Step 5: Based on the obtaned margnal ntervals and correlaton coeffcents of all the varables, buld the multdmensonal parallelepped convex model. The bult multdmensonal parallelepped model should satsfy: (1) ts changng range along each axs s equal to the correspondng margnal nterval and () the angle of each two edges of the parallelepped model s equal to correspondng correlaton angle. For a multdmensonal problem, partcularly n cases of hgh-dmensonal parameters, t s relatvely dffcult to buld a parallelepped model drectly through the samples. However, n the aforementoned method, ths dffculty can be effectvely elmnated through decomposng the complex n-dmensonal problem nto n.nc1/ smple two-dmensonal problems. Just usng the margnal ntervals and correlaton angles that can be easly obtaned through the samples, a multdmensonal uncertanty doman can be bult, whch geometrcally s a parallelepped. In addton, t should be noted that n many practcal cases, the margnal ntervals of some parameters can be drectly known based on the engneers experence or the exstng knowledge of the problem concerned. Thus, only ther correlaton angles need to be obtaned usng the samples by means of the aforementoned procedure. 4. AFFINE COORDINATE TRANSFORMATION FOR THE CONVEX MODEL In the aforementoned analyss, we constructed a multdmensonal parallelepped model for the descrpton of the uncertanty of parameters. However, the doman formed by a multdmensonal parallelepped can hardly be expressed mathematcally by explct functons such as the boundares of the nterval model and the quadratc functon of the ellpsodal model. Ths results n dffculty n performng subsequent structural uncertanty analyses, such as uncertanty propagaton analyss and relablty analyss. In ths secton, we ntroduce the use of affne coordnates [35, 36], whch provdes an effectve soluton to the problem. In the affne coordnates of our formulaton, the angle between the axes wll no longer be constant (equal to 90 ı ); rather, t wll be equal to the correlaton angle between the correspondng varables. Through affne transformaton, the parallelepped model

8 C. JIANG ET AL. n the orgnal parameter space can be converted to the nterval model n the affne space, thus greatly facltatng subsequent structural uncertanty analyss and calculaton Two-dmensonal problem As shown n Fgure 7, the margnal ntervals x 1 x1 I D x1 L;xU 1 and x x I D x L;xU and the correlaton± angle 1 of the parallelogram model are known. A new Cartesan coordnate system O 0 I e 1 ; e s defned by makng ts orgn and the center of the parallelogram concde. And the affne coordnate system ¹O 00 I e 1 ; e º s establshed accordng to the parallelogram model. As shown n Fgure 8, the center of the parallelogram s taken as the orgn; axs e 1 s parallel to the bottom edge of the parallelogram, that s, n the same drecton as axs e 1 n the Cartesan coordnate system, whle axs e s parallel to another sde of the parallelogram. Obvously, through the affne coordnates ¹O 00 I e 1 ; e º, the correlaton between two varables can be transformed nto the ncluded angle between the two axes n the affne coordnates. Wth respect to the Cartesan coordnate system ¹O 0 I e 1 ; e º and the affne coordnate system ¹O 00 I e 1 ; e º, accordng to the prncples of the affne coordnate system [35], t can be assumed that e 1 D a 11 e 1 C a 1e e D a 1e 1 C a (8) e where a 11 ;a 1 ;a 1 and a are weght coeffcents. From the condton that the parallelogram bottom edge s parallel to the axs e 1, we can conclude that a 1 D 0. Consderng the condton that e 1?e, Equaton (8) can be rewrtten n matrx form as e 1 e D e 1 e A (9) Fgure 7. The mdpont shft n the coordnate system. Fgure 8. The establshment of an affne coordnate system.

9 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Fgure 9. Rad of the margnal ntervals n dfferent coordnate systems: (a) In orgnal coordnates and (b) n affne coordnates. In whch A s the coordnate transformaton matrx A T 1 0 D cos 1 sn 1 : (10) Combnng the coordnate translaton, we obtan the mappng relaton between the ntal parametrc space and the fnal affne space T 00 T x1 x D A x 1 x00 C x1 RCxL 1 x R CxL T (11) As shown n Fgure 9, after transformaton of the affne coordnate system, the rad of the margnal ntervals n the orgnal Cartesan coordnate system and the ones n the affne coordnates wll change and have the followng relaton x W 00 1 x W 00 T D A 1 x1 W x W T (1) n whch x1 W and xw are the margnal ntervals of the uncertan parameters n the orgnal space, and x W 00 1 and x W 00 are the ones n the affne space. 4.. Multdmensonal problem In n-dmensonal parallelepped model, as for the two-dmensonal case, the correlaton angle between varables can be transformed nto the angle between the correspondng axes n affne coordnates. In a smlar way as the two-dmensonal case, we can obtan the weght coeffcents of the coordnate transformaton, whch are the elements of matrx A T 8 0.j > / ˆ< a j D s ˆ: jp 1 cos k a m a jm md1 1 ajj.j < / jp 1 ld1 a l.j D / For smplcty of expresson, the aforementoned equaton uses k to represent angle j,nwhch.n j /.j 1/ the subscrpt k D C. j/. Combnng the coordnate translaton, the mappng relaton of the ntal parameter space and the fnal affne space for an n-dmensonal problem can be expressed as T 00 T T x1 x ::: x n D A x 1 x00 ::: x00 n C x1 RCxL 1 x RCxL ::: xr n CxL n (14) (13)

10 C. JIANG ET AL. Accordngly, the relatonshp between the rad of the parameter ntervals before and after the affne transformaton can be represented as follows x W 00 1 x W 00 ::: x W 00 T n D A 1 x1 W x W ::: xn W T (15) 5. APPLICATION TO STRUCTURAL UNCERTAINTY PROPAGATION ANALYSIS In structural analyss, uncertantes n parameters such as materal propertes, loads, boundares, and other factors wll lead to uncertantes n structural responses such as dsplacement, stress, and stran. Ths represents an uncertanty propagaton problem. Through the analyss of uncertanty propagaton, t s possble to determne the quanttatve nfluence of the degree of ntal uncertanty on the structural response. Thus, relablty desgn can be conducted. In the convex model method, the parameter uncertanty belongs to a bounded set. Therefore, the structural response usually belongs to an nterval. In the convex model, performng an accurate calculaton of ths response nterval s the man task of uncertanty propagaton analyss. Assume that the uncertan parameters for a structure are X ; D 1; ; :::; n. The structural response functon s then g.x/ ; X X I ; (16) In Equaton (16), s the uncertanty doman, and n ths paper, t comprses a multdmensonal parallelepped. The sze and shape of the uncertanty doman are decded by the margnal ntervals X I and the correlaton angles of the parameters. Frst, through the affne transformaton Equaton (14), the parallelepped convex model s transformed to the affne coordnates. Accordngly, the response functon G n the affne space can be generated g.x/ D g T X 00 D G X 00 ; X (17) where T s the transformaton functon determned by Equaton (14). 00 represents the uncertanty doman n the affne space. As shown n Fgure 9, n the affne space, the uncertanty doman becomes an nterval model n whch the center pont s the orgn of the affne coordnate system. The parameters nterval rad are X W 00 ; D 1; ; :::; n. Therefore, the exstng structural uncertanty propagaton methods such as [11 15, 37], whch are based on the conventonal nterval analyss, can be drectly used. Consderng that the level of uncertanty of parameters s generally small, n order to decrease the amount of calculaton, especally for some complex engneerng problems, the response functon can be expanded to the frst-order Taylor seres G.X 00 / D G.X C 00 / C C 00 / 00 X 00 X C 00 n whch the deployment pont X C 00 s the affne coordnate followng equaton 00 The value of X 00 X C C 00 / 00 nx j C j can be explctly obtaned from Equaton (14). belongs to an nterval below h X 00 X C 00 X W 00 ;X W 00 (18) can be calculated by the ; (19) ;D 1; ; :::; n (0)

11 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Thus, applyng the nterval extenson [37] to Equaton (18), an explct expresson can be obtaned for the response nterval; the lower and upper bounds wll be 8 ˆ< G L.X / D G.X C 00 P / n 00 / ˇX W D1 ˆ: G R.X / D G.X C 00 P / C n 00 / (1) ˇ ˇX W 00 The rad of the ntervals n the affne space are X W 00 ;D 1; ; :::; n, whch can be obtaned from Equaton (15) NUMERICAL EXAMPLES AND DISCUSSIONS In ths secton, three numercal examples are analyzed. In the frst problem, subgrade settlement predcton problem was ntroduced to descrbe the process of modelng of multdmensonal parallelepped model. In the second problem of the explct functon, a comparson between the multdmensonal parallelepped model proposed n ths paper, the conventonal nterval model, and the ellpsodal model was conducted. In the thrd and fourth examples, parallelepped models were used n the uncertanty modelng and propagaton analyss of statcs and dynamcs problems, respectvely. Addtonally, the thrd example orgnates from a real engneerng problem of the roll steerng characterstcs The model constructon of engneerng problems Analyss for a subgrade settlement problem. Subgrade settlement s an mportant ndcator of road safety and a frequent cause of the road traffc accdents. Presently, some methods have been developed to mprove the accuracy and relablty of the subgrade settlement predcton. However, most of them are based on the sngle-pont montorng mode and thus could only study the local deformaton, whle correlaton of the settlements at dfferent ponts could not be obtaned. Subgrade settlement s a complex systematc process, n whch the deformaton of a montorng pont wll be effected by other montorng ponts on the same subgrade cross-secton. Thus, to mprove the predcton accuracy, t s often necessary to use the mult-pont montorng mode and then nvestgate the correlaton of the deformaton amounts from dfferent montorng ponts. Based on the reference [38], a practcal two-way four-lane hghway s nvestgated here. As shown n Fgure 10, three montorng ponts A, B,andC were arranged n the left half secton of the road, and three dsplacement sensors were nstalled to test ther responses. Ten groups of measured data for the montorng ponts were gven n Table I [38], n whch a measurng perod of 15 days was adopted. In ths problem, we expect to analyze the correlatons between the settlements of the three montorng ponts and furthermore gve a quanttatve descrpton for ther whole uncertanty doman. There are only 10 groups of data; thus, the tradtonal probablty approach seems not Fgure 10. Arrangement of the montorng ponts (cm) [38].

12 C. JIANG ET AL. Table I. Measured data of the montorng ponts [38]. Settlement(mm) No. Days Pont A Pont B Pont C Table II. Results from the two-dmensonal analyss n A-B parameter space. Montorng ponts Lower bound (mm) Upper bound (mm) Correlaton angle( ı ) A B Table III. Results from the two-dmensonal analyss n A-C parameter space. Montorng ponts Lower bound (mm) Upper bound (mm) Correlaton angle( ı ) A C Table IV. Results from the two-dmensonal analyss n B-C parameter space. Montorng ponts Lower bound (mm) Upper bound (mm) Correlaton angle( ı ) B C applcable here. The proposed parallelepped convex model s then used to deal wth these expermental data. We frst created three parallelograms fttng the samples n the three two-dmensonal parameter spaces, and the correspondng analyss results were gven n Tables II IV. Syntheszng these results, the margnal ntervals and correlaton angles of the dsplacements at the three montorng ponts can be obtaned, as A Œ11:35 mm; 30:78 mm, B Œ8:69 mm;4:75 mm, C Œ1:03 mm;7:16 mm, AB D 4:39 ı, BC D 41:51 ı,and AC D 45:53 ı, respectvely. Here, we also use A, B, and C to represent the dsplacements of the three montorng ponts. The three parallelograms fttng the measured data n the two-dmensonal spaces are llustrated n Fgure 11(a) (c). It can be observed that all the parallelograms seem thn and long, whch means that for these problems, the settlements of these three montorng ponts are sgnfcantly correlated wth each other. Furthermore, the correlatons of the three settlements seem at a smlar level. Based on all the margnal ntervals and correlaton angles, the whole uncertanty doman of the subgrade settlements at the three montorng ponts can be created as shown n Fgure 11(d), whch geometrcally s a parallelepped. It can be found that the parallelepped has a good fttng to the 10 samples. Also, through the parallelepped, the correlatons between each par of settlements can be well descrbed. Based on ths uncertanty doman, some mportant structural uncertanty analyss such as uncertanty propagaton and relablty desgn can then be carred out for the subgrade settlement problem.

13 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS B / mm (a) A / mm C / mm (c) C / mm (b) B / mm A / mm 6 4 C / mm (d) B / mm A / mm 10 Fgure 11. The multdmensonal parallelepped model for the subgrade settlement predcton problem Analyss for a ple load test problem. In ple foundaton desgn, research studes have ndcated that model uncertanty s an mportant factor. We consder a ple load test problem, n whch the ple load test reports were collected from varous plng companes n South Afrca [39]. In order to evaluate the measured or nterpreted ultmate capactes of all the test ples n the database, a combnaton of Chn s extrapolaton procedure and Davsson s falure crteron was employed.

14 C. JIANG ET AL. And a and b are the hyperbolc curve-fttng parameters for the load-settlement curve, and they are not ndependent. A dataset for D-C (drven ples n cohesve sols) wth 57 data (two among the orgnal 59 samples are abnormal ponts, and hence, they were removed), and a dataset for B-NC (bored ples n noncohesve sols) wth 31 data (two among the orgnal 33 samples are abnormal ponts, and hence, they were removed) as shown n Table V and VI, are used for llustraton [39]. In ths problem, we expect to analyze the correlaton between the hyperbolc curve-fttng parameters and furthermore gve a quanttatve descrpton for ther whole uncertanty doman. The proposed parallelepped convex model s then used to deal wth these expermental data, and we can create parallelogram (for a two-dmensonal problem) fttng the samples n the two-dmensonal parameter spaces. For D-C, the margnal ntervals and correlaton angle of the varables a and b can be obtaned: a Œ 0:8; 7:78, b Œ0:61; 0:95, ab D 176:6 ı. And for B-NC, the margnal ntervals and correlaton angle of the varables a and b can be obtaned: a Œ 1:44; 11:58, b Œ0:44; 0:99, ab D 176:4 ı. The correspondng analyss results are gven n Table VII and VIII, and the uncertanty domans are llustrated n Fgure 1(a) and (b) (a parallelogram doman, respectvely). When adoptng the nterval model, the correlaton of the varables s not taken nto account. The uncertanty domans are constructed wth the margnal ntervals a Œ0:57; 6:93, b Œ0:61; 0:95 for D-C and a Œ0:06; 10:93, b Œ0:44; 0:99 for B-NC, respectvely, whch are also llustrated n Fgure 1(a) and (b) (a rectangular doman, respectvely). Fgure 1 shows that the uncertanty domans from nterval model are much larger than those from parallelepped model. Wth the much more compact uncertanty doman of the parameters obtaned accordng to the parallelepped model, the subsequent uncertanty analyss for the ple load test problem thus wll be more precse and useful. Table V. Hyperbolc curve-fttng parameters a and b for ples (D-C) [37]. No. a b No. a b

15 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Table VI. Hyperbolc curve-fttng parameters a and b for ples (B-NC) [37]. No. a b No. a b Table VII. The margnal ntervals and correlaton angle of a and b for ples (D-C). Parameters Lower bound Upper bound Correlaton angle( ı ) a b Table VIII. The margnal ntervals and correlaton angle of a and b for ples (B-NC). Parameters Lower bound Upper bound Correlaton angle( ı ) a b a (a) D-C b b a (b) B-NC Fgure 1. The multdmensonal parallelepped model and nterval model for the Hyperbolc curve-fttng parameters: (a) D-C and (b) B-NC. 6.. An analytcal functon problem Consder the followng response functon g.x/ D X 1 X 3X 1 C X X 3 C 4X C 5X 3 X 1 X 3 C 10 ()

16 C. JIANG ET AL. Table IX. The frst sample case data. No. x 1 x x 3 No. x 1 x x In ths functon, the margnal ntervals of the three uncertan parameters are known as X1 I D Œ 5; 5 ; X I D Œ 4; 4 and X 3 I D Œ 8; 8. It s supposed there are 54 samples of the uncertan parameters, as shown n Table IX. In the followng, uncertanty modelng and propagaton analyss wll be conducted usng the nterval, ellpsod, and parallelepped models, and ther results wll be compared Interval model. When adoptng the nterval model, the correlaton of the varables s not taken nto account. The uncertanty doman s a cubod composed of the margnal ntervals X1 I ;XI, and X3 I. In uncertanty propagaton analyss, the response nterval can be approxmately obtaned drectly n the orgnal parameter space based on a frst-order Taylor expanson 8 ˆ< g L.X/ D g.x C P / n / ˇˇˇX D1 ˆ: g R.X/ D g.x C P / C n (3) / ˇˇˇX 6... Ellpsodal model. In ths model, all samples are frst ftted usng the mnmum volume method [40]. The ellpsodal model s then obtaned as follows X 1 3 T 0 1 C 0:0049 0:0540 0:0419 0:0008 X 0:0540 A 6 4 0:0419 0:0550 0: X 1 1 C 0:0049 X 0:0540 A 6 1 (4) X 3 0:0076 0:0008 0:0010 0:0075 X 3 0:0076 Thus, the changng bounds of the response functon over the ellpsod uncertanty doman can be calculated n the orgnal parameter space. It can be obtaned by the followng equaton [37] D1

17 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS 8 q ˆ< g L.X/ D g.x C / rg T M -1 rg q ˆ: g R.X/ D g.x C / C rg T M -1 rg (5) where M s the characterstc matrx of the aforementoned ellpsod. rg C 1 T s the gradent vector of the response functon wth respect to the uncertan C Multdmensonal parallelepped model. Usng the method descrbed n Secton 3.3, all sample ponts can be ftted. In ths way, the parallelepped model for the uncertanty doman, n whch the correlaton angles of the parameters are 1 D 60 ı, 13 D 90 ı,and 3 D 90 ı, s obtaned. Ths ndcates that only X 1 and X have correlaton; thus, ths problem has both dependent parameters and ndependent parameters. The coordnate transformaton matrx A s 1 cos 60 ı cos 90 ı A D 4 0 sn 60 ı cos 90ı cos 60 ı cos 90 ı sn 60 ı D p (6) From Equaton (18), the mappng relaton between the orgnal parameter space and the affne space s X X X 00 X C 1 X 00 1 X A B D X 00 C 6 p A D B 0 X 00 p C B A 3 X 00 C A : (7) X 3 X X 00 3 X 00 3 Thus, the expresson of the response functon n the affne space can be explctly obtaned for ths problem as p p G.X 00 / D 3X C X 00 1 X X1 X C 4 X 00 C p 3 3 p! X 00 3 C 1 X 00 X C5X3 C10 (8) The margnal ntervals of the parameters n the affne space are X C p 4 ;5 4 p ;X 00 p 8 ; p ;X 00 3 Œ 8; 8 : (9) 3 Accordng to Equaton (1), the uncertanty nterval of the response functon can eventually be solved Comparson of the results. In ths secton, the geometrc shape and sze of the uncertanty doman constructed from three convex models wll be compared. For smplcty, the uncertanty doman s projected onto the two-dmensonal parameter spaces, as shown n Fgure 13. It can be seen from the fgure that, n the X 1 X parameter space, the ellpsodal and parallelepped models both reflect the correlaton of the parameters. However, n the spaces of X 1 X 3 and X X 3,the ellpsodal model was not able to deal wth the problem of ndependent parameters, thus enlargng the uncertanty doman to a great extent, whereas the parallelepped model was able to envelope all the samples well. In the spaces of X 1 X 3 and X X 3, due to the fact that the parameters do not exhbt correlaton wth each other, the nterval and parallelepped models have the same uncertanty domans. In the X 1 X space, because the nterval model cannot handle the correlaton, t sgnfcantly expands the uncertanty doman, whle the parallelepped model fts all the samples very well. Fgure 14 shows three-dmensonal graphcs depctng the uncertanty domans from the three models. It can be seen that the uncertanty doman obtaned from the parallelepped model s sgnfcantly smaller than that obtaned by ether of the other two methods.

18 C. JIANG ET AL. Fgure 13. Projectons of the uncertanty domans n two-dmensonal parameter spaces: (a) X 1 X plane, (b) X 1 X 3 plane, and (c) X X 3 plane. Fgure 14. Three-dmensonal uncertanty domans from three convex models.

19 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Through the foregong analyss, we found that, for problems wth ether correlaton or ndependence, the nterval and ellpsodal model are both lkely to construct a too-large uncertanty doman, resultng n an ultra-conservatve desgn. The parallelepped model, however, can handle both dependent varables and ndependent varables and thus can construct a more compact uncertanty doman, resultng n more reasonable uncertanty analyss results. The uncertanty propagaton analyss results based on the aforementoned convex models are shown n Table X. The results show that, among these three methods, the response nterval obtaned by the parallelepped model s the narrowest and s contaned n the results of both the nterval model and the ellpsodal model. Ths occurs because the parallelepped model constructs a more compact doman of uncertanty, so that the calculaton results are not as conservatve as those produced by the nterval and the ellpsodal models. In the aforementoned example, the margnal ntervals of the parameters do not change. We also consdered three other knds of sample dstrbutons, as shown n Table XI XIII; the three convex models were used separately for analyzng these samples. When uncertanty modelng s conducted usng the multdmensonal parallelepped model, 13 and 3 are both 90 ı n all three sample cases, whle the correlaton angle 1 takes on values 5:5 ı, 75 ı,and90 ı, respectvely. The correspondng correlaton coeffcents 1 are 0.614, 0.14, and 0.000, respectvely. Fgure 15 shows the geometry of the uncertanty doman under dfferent correlaton angles. It can be seen that dfferent correlaton angles 1 drectly result n uncertanty domans of dfferent shapes and szes. Wth a smaller 1, the parallelepped s more flat, and the volume s also smaller. As 1 gradually approaches 90 ı,the shape of the uncertanty doman gradually becomes a cube, and ts volume gradually ncreases. The results of uncertanty propagaton analyses for the aforementoned cases are shown n Table XIV and Fgure 16. From the calculaton results of the ellpsodal and parallelepped models, t can be Table X. Uncertanty propagaton analyss results for the frst sample case. Model Response Interval model Ellpsodal model Parallelepped model Response nterval [ 61.00, 81.00] [ 51.5, 7.06] [ 47.14, 67.14] Radus of nterval Table XI. The second sample case data. No. x 1 x x 3 No. x 1 x x

20 C. JIANG ET AL. Table XII. The thrd sample case data. No. x 1 x x 3 No. x 1 x x Table XIII. The fourth sample case data. No. x 1 x x 3 No. x 1 x x

21 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Fgure 15. Uncertanty domans under dfferent correlatons of the parameters: (a) 1 D 40 ı,(b) 1 D 5:5 ı,(c) 1 D 67:5 ı,(d) 1 D 75 ı,(e) 1 D 80 ı, and (f) 1 D 90 ı. Table XIV. Uncertanty propagaton analyss results for dfferent sample cases. Case Model Response nterval of case Response nterval of case 3 Response nterval of case 4 Interval model [ 61.00,81.00] [ 61.00,81.00] [ 61.00,81.00] Ellpsodal model [ 50.85,70.07] [ 63.7,83.30] [ 69.55,90.47] Parallelepped [ 4.58,6.58] [ 54.57,74.57] [ 61.00,81.00] Model 1 D 5:5 ı 1 D 75 ı 1 D 90 ı seen that the correlaton of parameters X 1 and X has a sgnfcant effect on the varaton of response functon n ths example. Wth the correlaton decreasng (.e., 1 becomng larger), the response ntervals calculated by both the ellpsodal and parallelepped model show a progressvely larger trend. However, because the nterval model cannot descrbe the correlaton of parameters, only a constant analyss results can be generated n all three cases. In the fourth case, all the parameters are not correlated, and the parallelepped model results are consstent wth the nterval model results. However, because of the nablty of the ellpsodal model to deal wth ndependent varables, the model gves a wder range of the response. In addton, t should be noted that the parallelepped model gves the narrowest range of the response functon n all cases, and ts result are ncluded by the nterval model and ellpsodal model results. Ths further shows that the uncertanty doman of the parallelepped model s more compact than the ones of the other two models and that t generally could provde more reasonable and effectve uncertanty analyss results.

22 C. JIANG ET AL Fgure 16. The bounds of the response functon from three convex models wth dfferent parametrc correlatons. Fgure 17. A 10-bar alumnum truss A 10-bar truss structure A well-known 10-bar alumnum truss structure [41], shown n Fgure 17, s used as an example. The length of the vertcal and horzontal bars s L D 360 n. The cross-sectonal area of the bar s A D 10 n. The elastc Modulus of the materal s E D 10 4 ks. A vertcal load F 1 s appled on node 4. A vertcal force F and a horzontal force F 3 are appled on node. Loads F 1, F,andF 3 are all uncertan parameters. A multdmensonal parallelepped model s used to buld the uncertanty doman. Suppose the margnal ntervals are known to be F1 I D Œ90; 110 kps, F I D Œ90; 110 kps, and F3 I D Œ190; 10 kps, respectvely. All the three correlaton angles of the parameters have the same value. In ths example, the response nterval of the vertcal dsplacement ı of node wll be analyzed. Standard structural mechancs analyss gves [41], ı D " 6X D1 N 0 N A C p X10 D7 N 0N # L A E n whch N ;D 1; ; :::; 10 are the axal forces n the bar, as follows N 1 D F p p p N 8;N D N 10;N 3 D F 1 F C F 3 N 8 (30)

23 A NON-PROBABILISTIC CONVEX MODEL FOR STRUCTURAL UNCERTAINTY ANALYSIS Table XV. Response nterval of the vertcal dsplacement ı for cases wth dfferent parametrc correlatons. Correlaton angle Response nterval 50 ı 55 ı 60 ı 65 ı 70 ı 75 ı 80 ı 85 ı 90 ı Lower bound (n) Upper bound (n) Radus of nterval (n)

24 C. JIANG ET AL. N 4 D F C F 3 p p p p N 10;N 5 D F N 8 N 10;N 6 D N 10 N 7 D p.f 1 C F / C N 8 ;N 8 D a b 1 a 1 b a 11 a a 1 a 1 N 9 D p F C N 10 ;N 10 D a 11b a 1 b 1 a 11 a a 1 a 1 a 11 D 1 C 1 C 1 C p C p! A 1 A 3 A 5 A 7 a D A 8 1 A C 1 C 1 A 4 A5 C 1 p C p A6 A 9 A 10 L E ;a 1 D a 1 D! L E L A 5 E (31) b 1 D F A 1 F 1 C F F 3 A 3 F A 5 b D p.f3 F / A 4 p! F 4F A 5 A 7 p.f1 C F / A 7 L E ;! p L E N 0 s the value of N when F 1 D F 3 D 0 and F D 1. Durng the structural uncertanty propagaton analyss, a set of dfferent parametrc correlatons are taken nto account. The computatonal results by usng the present parallelepped model are gven n Table XV and Fgure 18. The results show that for the 10 bar truss, the correlaton between the three load parameters also has a sgnfcant nfluence on the vertcal dsplacement ı of node. Under dfferent correlaton angles, the response nterval of ı shows a relatvely large fluctuaton. In addton, when the correlaton angle ncreases, the response radus of ı frst exhbts a downtrend and then a rsng trend. In all nne stuatons, ı has a mnmum response radus equal to ı W D n when D 60 ı Fgure 18. The relatonshp between the response nterval of vertcal dsplacement ı and the parametrc correlaton angle.

Indeterminate pin-jointed frames (trusses)

Indeterminate pin-jointed frames (trusses) Indetermnate pn-jonted frames (trusses) Calculaton of member forces usng force method I. Statcal determnacy. The degree of freedom of any truss can be derved as: w= k d a =, where k s the number of all

More information

Chapter 13: Multiple Regression

Chapter 13: Multiple Regression Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to

More information

NUMERICAL DIFFERENTIATION

NUMERICAL DIFFERENTIATION NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the

More information

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity Week3, Chapter 4 Moton n Two Dmensons Lecture Quz A partcle confned to moton along the x axs moves wth constant acceleraton from x =.0 m to x = 8.0 m durng a 1-s tme nterval. The velocty of the partcle

More information

COMPOSITE BEAM WITH WEAK SHEAR CONNECTION SUBJECTED TO THERMAL LOAD

COMPOSITE BEAM WITH WEAK SHEAR CONNECTION SUBJECTED TO THERMAL LOAD COMPOSITE BEAM WITH WEAK SHEAR CONNECTION SUBJECTED TO THERMAL LOAD Ákos Jósef Lengyel, István Ecsed Assstant Lecturer, Professor of Mechancs, Insttute of Appled Mechancs, Unversty of Mskolc, Mskolc-Egyetemváros,

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

One-sided finite-difference approximations suitable for use with Richardson extrapolation

One-sided finite-difference approximations suitable for use with Richardson extrapolation Journal of Computatonal Physcs 219 (2006) 13 20 Short note One-sded fnte-dfference approxmatons sutable for use wth Rchardson extrapolaton Kumar Rahul, S.N. Bhattacharyya * Department of Mechancal Engneerng,

More information

DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM

DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM Ganj, Z. Z., et al.: Determnaton of Temperature Dstrbuton for S111 DETERMINATION OF TEMPERATURE DISTRIBUTION FOR ANNULAR FINS WITH TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY BY HPM by Davood Domr GANJI

More information

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

More information

2016 Wiley. Study Session 2: Ethical and Professional Standards Application

2016 Wiley. Study Session 2: Ethical and Professional Standards Application 6 Wley Study Sesson : Ethcal and Professonal Standards Applcaton LESSON : CORRECTION ANALYSIS Readng 9: Correlaton and Regresson LOS 9a: Calculate and nterpret a sample covarance and a sample correlaton

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng

More information

Chapter 9: Statistical Inference and the Relationship between Two Variables

Chapter 9: Statistical Inference and the Relationship between Two Variables Chapter 9: Statstcal Inference and the Relatonshp between Two Varables Key Words The Regresson Model The Sample Regresson Equaton The Pearson Correlaton Coeffcent Learnng Outcomes After studyng ths chapter,

More information

Linear Approximation with Regularization and Moving Least Squares

Linear Approximation with Regularization and Moving Least Squares Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...

More information

FUZZY FINITE ELEMENT METHOD

FUZZY FINITE ELEMENT METHOD FUZZY FINITE ELEMENT METHOD RELIABILITY TRUCTURE ANALYI UING PROBABILITY 3.. Maxmum Normal tress Internal force s the shear force, V has a magntude equal to the load P and bendng moment, M. Bendng moments

More information

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and

More information

Finite Element Modelling of truss/cable structures

Finite Element Modelling of truss/cable structures Pet Schreurs Endhoven Unversty of echnology Department of Mechancal Engneerng Materals echnology November 3, 214 Fnte Element Modellng of truss/cable structures 1 Fnte Element Analyss of prestressed structures

More information

Errors in Nobel Prize for Physics (7) Improper Schrodinger Equation and Dirac Equation

Errors in Nobel Prize for Physics (7) Improper Schrodinger Equation and Dirac Equation Errors n Nobel Prze for Physcs (7) Improper Schrodnger Equaton and Drac Equaton u Yuhua (CNOOC Research Insttute, E-mal:fuyh945@sna.com) Abstract: One of the reasons for 933 Nobel Prze for physcs s for

More information

Global Sensitivity. Tuesday 20 th February, 2018

Global Sensitivity. Tuesday 20 th February, 2018 Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values

More information

Report on Image warping

Report on Image warping Report on Image warpng Xuan Ne, Dec. 20, 2004 Ths document summarzed the algorthms of our mage warpng soluton for further study, and there s a detaled descrpton about the mplementaton of these algorthms.

More information

4DVAR, according to the name, is a four-dimensional variational method.

4DVAR, according to the name, is a four-dimensional variational method. 4D-Varatonal Data Assmlaton (4D-Var) 4DVAR, accordng to the name, s a four-dmensonal varatonal method. 4D-Var s actually a drect generalzaton of 3D-Var to handle observatons that are dstrbuted n tme. The

More information

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method Appled Mathematcal Scences, Vol. 7, 0, no. 47, 07-0 HIARI Ltd, www.m-hkar.com Comparson of the Populaton Varance Estmators of -Parameter Exponental Dstrbuton Based on Multple Crtera Decson Makng Method

More information

Second Order Analysis

Second Order Analysis Second Order Analyss In the prevous classes we looked at a method that determnes the load correspondng to a state of bfurcaton equlbrum of a perfect frame by egenvalye analyss The system was assumed to

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

A Robust Method for Calculating the Correlation Coefficient

A Robust Method for Calculating the Correlation Coefficient A Robust Method for Calculatng the Correlaton Coeffcent E.B. Nven and C. V. Deutsch Relatonshps between prmary and secondary data are frequently quantfed usng the correlaton coeffcent; however, the tradtonal

More information

EVALUATION OF THE VISCO-ELASTIC PROPERTIES IN ASPHALT RUBBER AND CONVENTIONAL MIXES

EVALUATION OF THE VISCO-ELASTIC PROPERTIES IN ASPHALT RUBBER AND CONVENTIONAL MIXES EVALUATION OF THE VISCO-ELASTIC PROPERTIES IN ASPHALT RUBBER AND CONVENTIONAL MIXES Manuel J. C. Mnhoto Polytechnc Insttute of Bragança, Bragança, Portugal E-mal: mnhoto@pb.pt Paulo A. A. Perera and Jorge

More information

A new Approach for Solving Linear Ordinary Differential Equations

A new Approach for Solving Linear Ordinary Differential Equations , ISSN 974-57X (Onlne), ISSN 974-5718 (Prnt), Vol. ; Issue No. 1; Year 14, Copyrght 13-14 by CESER PUBLICATIONS A new Approach for Solvng Lnear Ordnary Dfferental Equatons Fawz Abdelwahd Department of

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

Fundamental loop-current method using virtual voltage sources technique for special cases

Fundamental loop-current method using virtual voltage sources technique for special cases Fundamental loop-current method usng vrtual voltage sources technque for specal cases George E. Chatzaraks, 1 Marna D. Tortorel 1 and Anastasos D. Tzolas 1 Electrcal and Electroncs Engneerng Departments,

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS

NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS IJRRAS 8 (3 September 011 www.arpapress.com/volumes/vol8issue3/ijrras_8_3_08.pdf NON-CENTRAL 7-POINT FORMULA IN THE METHOD OF LINES FOR PARABOLIC AND BURGERS' EQUATIONS H.O. Bakodah Dept. of Mathematc

More information

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.

More information

Power law and dimension of the maximum value for belief distribution with the max Deng entropy

Power law and dimension of the maximum value for belief distribution with the max Deng entropy Power law and dmenson of the maxmum value for belef dstrbuton wth the max Deng entropy Bngy Kang a, a College of Informaton Engneerng, Northwest A&F Unversty, Yanglng, Shaanx, 712100, Chna. Abstract Deng

More information

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan Wnter 2008 CS567 Stochastc Lnear/Integer Programmng Guest Lecturer: Xu, Huan Class 2: More Modelng Examples 1 Capacty Expanson Capacty expanson models optmal choces of the tmng and levels of nvestments

More information

ALGORITHM FOR THE CALCULATION OF THE TWO VARIABLES CUBIC SPLINE FUNCTION

ALGORITHM FOR THE CALCULATION OF THE TWO VARIABLES CUBIC SPLINE FUNCTION ANALELE ŞTIINŢIFICE ALE UNIVERSITĂŢII AL.I. CUZA DIN IAŞI (S.N.) MATEMATICĂ, Tomul LIX, 013, f.1 DOI: 10.478/v10157-01-00-y ALGORITHM FOR THE CALCULATION OF THE TWO VARIABLES CUBIC SPLINE FUNCTION BY ION

More information

This column is a continuation of our previous column

This column is a continuation of our previous column Comparson of Goodness of Ft Statstcs for Lnear Regresson, Part II The authors contnue ther dscusson of the correlaton coeffcent n developng a calbraton for quanttatve analyss. Jerome Workman Jr. and Howard

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

Module 3: Element Properties Lecture 1: Natural Coordinates

Module 3: Element Properties Lecture 1: Natural Coordinates Module 3: Element Propertes Lecture : Natural Coordnates Natural coordnate system s bascally a local coordnate system whch allows the specfcaton of a pont wthn the element by a set of dmensonless numbers

More information

Uncertainty as the Overlap of Alternate Conditional Distributions

Uncertainty as the Overlap of Alternate Conditional Distributions Uncertanty as the Overlap of Alternate Condtonal Dstrbutons Olena Babak and Clayton V. Deutsch Centre for Computatonal Geostatstcs Department of Cvl & Envronmental Engneerng Unversty of Alberta An mportant

More information

Uncertainty and auto-correlation in. Measurement

Uncertainty and auto-correlation in. Measurement Uncertanty and auto-correlaton n arxv:1707.03276v2 [physcs.data-an] 30 Dec 2017 Measurement Markus Schebl Federal Offce of Metrology and Surveyng (BEV), 1160 Venna, Austra E-mal: markus.schebl@bev.gv.at

More information

NUMERICAL RESULTS QUALITY IN DEPENDENCE ON ABAQUS PLANE STRESS ELEMENTS TYPE IN BIG DISPLACEMENTS COMPRESSION TEST

NUMERICAL RESULTS QUALITY IN DEPENDENCE ON ABAQUS PLANE STRESS ELEMENTS TYPE IN BIG DISPLACEMENTS COMPRESSION TEST Appled Computer Scence, vol. 13, no. 4, pp. 56 64 do: 10.23743/acs-2017-29 Submtted: 2017-10-30 Revsed: 2017-11-15 Accepted: 2017-12-06 Abaqus Fnte Elements, Plane Stress, Orthotropc Materal Bartosz KAWECKI

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

Formulas for the Determinant

Formulas for the Determinant page 224 224 CHAPTER 3 Determnants e t te t e 2t 38 A = e t 2te t e 2t e t te t 2e 2t 39 If 123 A = 345, 456 compute the matrx product A adj(a) What can you conclude about det(a)? For Problems 40 43, use

More information

Composite Hypotheses testing

Composite Hypotheses testing Composte ypotheses testng In many hypothess testng problems there are many possble dstrbutons that can occur under each of the hypotheses. The output of the source s a set of parameters (ponts n a parameter

More information

ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM

ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM ELASTIC WAVE PROPAGATION IN A CONTINUOUS MEDIUM An elastc wave s a deformaton of the body that travels throughout the body n all drectons. We can examne the deformaton over a perod of tme by fxng our look

More information

Inductance Calculation for Conductors of Arbitrary Shape

Inductance Calculation for Conductors of Arbitrary Shape CRYO/02/028 Aprl 5, 2002 Inductance Calculaton for Conductors of Arbtrary Shape L. Bottura Dstrbuton: Internal Summary In ths note we descrbe a method for the numercal calculaton of nductances among conductors

More information

Comparative Studies of Law of Conservation of Energy. and Law Clusters of Conservation of Generalized Energy

Comparative Studies of Law of Conservation of Energy. and Law Clusters of Conservation of Generalized Energy Comparatve Studes of Law of Conservaton of Energy and Law Clusters of Conservaton of Generalzed Energy No.3 of Comparatve Physcs Seres Papers Fu Yuhua (CNOOC Research Insttute, E-mal:fuyh1945@sna.com)

More information

Chapter - 2. Distribution System Power Flow Analysis

Chapter - 2. Distribution System Power Flow Analysis Chapter - 2 Dstrbuton System Power Flow Analyss CHAPTER - 2 Radal Dstrbuton System Load Flow 2.1 Introducton Load flow s an mportant tool [66] for analyzng electrcal power system network performance. Load

More information

An Algorithm to Solve the Inverse Kinematics Problem of a Robotic Manipulator Based on Rotation Vectors

An Algorithm to Solve the Inverse Kinematics Problem of a Robotic Manipulator Based on Rotation Vectors An Algorthm to Solve the Inverse Knematcs Problem of a Robotc Manpulator Based on Rotaton Vectors Mohamad Z. Al-az*, Mazn Z. Othman**, and Baker B. Al-Bahr* *AL-Nahran Unversty, Computer Eng. Dep., Baghdad,

More information

The Two-scale Finite Element Errors Analysis for One Class of Thermoelastic Problem in Periodic Composites

The Two-scale Finite Element Errors Analysis for One Class of Thermoelastic Problem in Periodic Composites 7 Asa-Pacfc Engneerng Technology Conference (APETC 7) ISBN: 978--6595-443- The Two-scale Fnte Element Errors Analyss for One Class of Thermoelastc Problem n Perodc Compostes Xaoun Deng Mngxang Deng ABSTRACT

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

SIMPLE LINEAR REGRESSION

SIMPLE LINEAR REGRESSION Smple Lnear Regresson and Correlaton Introducton Prevousl, our attenton has been focused on one varable whch we desgnated b x. Frequentl, t s desrable to learn somethng about the relatonshp between two

More information

A Hybrid Variational Iteration Method for Blasius Equation

A Hybrid Variational Iteration Method for Blasius Equation Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 1 (June 2015), pp. 223-229 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) A Hybrd Varatonal Iteraton Method

More information

Uncertainty in measurements of power and energy on power networks

Uncertainty in measurements of power and energy on power networks Uncertanty n measurements of power and energy on power networks E. Manov, N. Kolev Department of Measurement and Instrumentaton, Techncal Unversty Sofa, bul. Klment Ohrdsk No8, bl., 000 Sofa, Bulgara Tel./fax:

More information

Assessment of Site Amplification Effect from Input Energy Spectra of Strong Ground Motion

Assessment of Site Amplification Effect from Input Energy Spectra of Strong Ground Motion Assessment of Ste Amplfcaton Effect from Input Energy Spectra of Strong Ground Moton M.S. Gong & L.L Xe Key Laboratory of Earthquake Engneerng and Engneerng Vbraton,Insttute of Engneerng Mechancs, CEA,

More information

A new construction of 3-separable matrices via an improved decoding of Macula s construction

A new construction of 3-separable matrices via an improved decoding of Macula s construction Dscrete Optmzaton 5 008 700 704 Contents lsts avalable at ScenceDrect Dscrete Optmzaton journal homepage: wwwelsevercom/locate/dsopt A new constructon of 3-separable matrces va an mproved decodng of Macula

More information

829. An adaptive method for inertia force identification in cantilever under moving mass

829. An adaptive method for inertia force identification in cantilever under moving mass 89. An adaptve method for nerta force dentfcaton n cantlever under movng mass Qang Chen 1, Mnzhuo Wang, Hao Yan 3, Haonan Ye 4, Guola Yang 5 1,, 3, 4 Department of Control and System Engneerng, Nanng Unversty,

More information

DUE: WEDS FEB 21ST 2018

DUE: WEDS FEB 21ST 2018 HOMEWORK # 1: FINITE DIFFERENCES IN ONE DIMENSION DUE: WEDS FEB 21ST 2018 1. Theory Beam bendng s a classcal engneerng analyss. The tradtonal soluton technque makes smplfyng assumptons such as a constant

More information

Grover s Algorithm + Quantum Zeno Effect + Vaidman

Grover s Algorithm + Quantum Zeno Effect + Vaidman Grover s Algorthm + Quantum Zeno Effect + Vadman CS 294-2 Bomb 10/12/04 Fall 2004 Lecture 11 Grover s algorthm Recall that Grover s algorthm for searchng over a space of sze wors as follows: consder the

More information

PHYS 450 Spring semester Lecture 02: Dealing with Experimental Uncertainties. Ron Reifenberger Birck Nanotechnology Center Purdue University

PHYS 450 Spring semester Lecture 02: Dealing with Experimental Uncertainties. Ron Reifenberger Birck Nanotechnology Center Purdue University PHYS 45 Sprng semester 7 Lecture : Dealng wth Expermental Uncertantes Ron Refenberger Brck anotechnology Center Purdue Unversty Lecture Introductory Comments Expermental errors (really expermental uncertantes)

More information

1 Derivation of Point-to-Plane Minimization

1 Derivation of Point-to-Plane Minimization 1 Dervaton of Pont-to-Plane Mnmzaton Consder the Chen-Medon (pont-to-plane) framework for ICP. Assume we have a collecton of ponts (p, q ) wth normals n. We want to determne the optmal rotaton and translaton

More information

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6 Department of Quanttatve Methods & Informaton Systems Tme Seres and Ther Components QMIS 30 Chapter 6 Fall 00 Dr. Mohammad Zanal These sldes were modfed from ther orgnal source for educatonal purpose only.

More information

Affine transformations and convexity

Affine transformations and convexity Affne transformatons and convexty The purpose of ths document s to prove some basc propertes of affne transformatons nvolvng convex sets. Here are a few onlne references for background nformaton: http://math.ucr.edu/

More information

Chapter 2 A Class of Robust Solution for Linear Bilevel Programming

Chapter 2 A Class of Robust Solution for Linear Bilevel Programming Chapter 2 A Class of Robust Soluton for Lnear Blevel Programmng Bo Lu, Bo L and Yan L Abstract Under the way of the centralzed decson-makng, the lnear b-level programmng (BLP) whose coeffcents are supposed

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Experment-I MODULE VII LECTURE - 3 ANALYSIS OF COVARIANCE Dr Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Any scentfc experment s performed

More information

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models Computaton of Hgher Order Moments from Two Multnomal Overdsperson Lkelhood Models BY J. T. NEWCOMER, N. K. NEERCHAL Department of Mathematcs and Statstcs, Unversty of Maryland, Baltmore County, Baltmore,

More information

On a direct solver for linear least squares problems

On a direct solver for linear least squares problems ISSN 2066-6594 Ann. Acad. Rom. Sc. Ser. Math. Appl. Vol. 8, No. 2/2016 On a drect solver for lnear least squares problems Constantn Popa Abstract The Null Space (NS) algorthm s a drect solver for lnear

More information

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH Turbulence classfcaton of load data by the frequency and severty of wnd gusts Introducton Oscar Moñux, DEWI GmbH Kevn Blebler, DEWI GmbH Durng the wnd turbne developng process, one of the most mportant

More information

Study on Non-Linear Dynamic Characteristic of Vehicle. Suspension Rubber Component

Study on Non-Linear Dynamic Characteristic of Vehicle. Suspension Rubber Component Study on Non-Lnear Dynamc Characterstc of Vehcle Suspenson Rubber Component Zhan Wenzhang Ln Y Sh GuobaoJln Unversty of TechnologyChangchun, Chna Wang Lgong (MDI, Chna [Abstract] The dynamc characterstc

More information

MMA and GCMMA two methods for nonlinear optimization

MMA and GCMMA two methods for nonlinear optimization MMA and GCMMA two methods for nonlnear optmzaton Krster Svanberg Optmzaton and Systems Theory, KTH, Stockholm, Sweden. krlle@math.kth.se Ths note descrbes the algorthms used n the author s 2007 mplementatons

More information

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k)

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k) ISSN 1749-3889 (prnt), 1749-3897 (onlne) Internatonal Journal of Nonlnear Scence Vol.17(2014) No.2,pp.188-192 Modfed Block Jacob-Davdson Method for Solvng Large Sparse Egenproblems Hongy Mao, College of

More information

Professor Terje Haukaas University of British Columbia, Vancouver The Q4 Element

Professor Terje Haukaas University of British Columbia, Vancouver  The Q4 Element Professor Terje Haukaas Unversty of Brtsh Columba, ancouver www.nrsk.ubc.ca The Q Element Ths document consders fnte elements that carry load only n ther plane. These elements are sometmes referred to

More information

OFF-AXIS MECHANICAL PROPERTIES OF FRP COMPOSITES

OFF-AXIS MECHANICAL PROPERTIES OF FRP COMPOSITES ICAMS 204 5 th Internatonal Conference on Advanced Materals and Systems OFF-AXIS MECHANICAL PROPERTIES OF FRP COMPOSITES VLAD LUPĂŞTEANU, NICOLAE ŢĂRANU, RALUCA HOHAN, PAUL CIOBANU Gh. Asach Techncal Unversty

More information

Comparison of Regression Lines

Comparison of Regression Lines STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence

More information

Module 9. Lecture 6. Duality in Assignment Problems

Module 9. Lecture 6. Duality in Assignment Problems Module 9 1 Lecture 6 Dualty n Assgnment Problems In ths lecture we attempt to answer few other mportant questons posed n earler lecture for (AP) and see how some of them can be explaned through the concept

More information

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests Smulated of the Cramér-von Mses Goodness-of-Ft Tests Steele, M., Chaselng, J. and 3 Hurst, C. School of Mathematcal and Physcal Scences, James Cook Unversty, Australan School of Envronmental Studes, Grffth

More information

RELIABILITY ASSESSMENT

RELIABILITY ASSESSMENT CHAPTER Rsk Analyss n Engneerng and Economcs RELIABILITY ASSESSMENT A. J. Clark School of Engneerng Department of Cvl and Envronmental Engneerng 4a CHAPMAN HALL/CRC Rsk Analyss for Engneerng Department

More information

Laboratory 1c: Method of Least Squares

Laboratory 1c: Method of Least Squares Lab 1c, Least Squares Laboratory 1c: Method of Least Squares Introducton Consder the graph of expermental data n Fgure 1. In ths experment x s the ndependent varable and y the dependent varable. Clearly

More information

Section 8.3 Polar Form of Complex Numbers

Section 8.3 Polar Form of Complex Numbers 80 Chapter 8 Secton 8 Polar Form of Complex Numbers From prevous classes, you may have encountered magnary numbers the square roots of negatve numbers and, more generally, complex numbers whch are the

More information

Lab 2e Thermal System Response and Effective Heat Transfer Coefficient

Lab 2e Thermal System Response and Effective Heat Transfer Coefficient 58:080 Expermental Engneerng 1 OBJECTIVE Lab 2e Thermal System Response and Effectve Heat Transfer Coeffcent Warnng: though the experment has educatonal objectves (to learn about bolng heat transfer, etc.),

More information

Canonical transformations

Canonical transformations Canoncal transformatons November 23, 2014 Recall that we have defned a symplectc transformaton to be any lnear transformaton M A B leavng the symplectc form nvarant, Ω AB M A CM B DΩ CD Coordnate transformatons,

More information

Chapter 8 Indicator Variables

Chapter 8 Indicator Variables Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n

More information

An identification algorithm of model kinetic parameters of the interfacial layer growth in fiber composites

An identification algorithm of model kinetic parameters of the interfacial layer growth in fiber composites IOP Conference Seres: Materals Scence and Engneerng PAPER OPE ACCESS An dentfcaton algorthm of model knetc parameters of the nterfacal layer growth n fber compostes o cte ths artcle: V Zubov et al 216

More information

CHAPTER 14 GENERAL PERTURBATION THEORY

CHAPTER 14 GENERAL PERTURBATION THEORY CHAPTER 4 GENERAL PERTURBATION THEORY 4 Introducton A partcle n orbt around a pont mass or a sphercally symmetrc mass dstrbuton s movng n a gravtatonal potental of the form GM / r In ths potental t moves

More information

arxiv:cs.cv/ Jun 2000

arxiv:cs.cv/ Jun 2000 Correlaton over Decomposed Sgnals: A Non-Lnear Approach to Fast and Effectve Sequences Comparson Lucano da Fontoura Costa arxv:cs.cv/0006040 28 Jun 2000 Cybernetc Vson Research Group IFSC Unversty of São

More information

Negative Binomial Regression

Negative Binomial Regression STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...

More information

χ x B E (c) Figure 2.1.1: (a) a material particle in a body, (b) a place in space, (c) a configuration of the body

χ x B E (c) Figure 2.1.1: (a) a material particle in a body, (b) a place in space, (c) a configuration of the body Secton.. Moton.. The Materal Body and Moton hyscal materals n the real world are modeled usng an abstract mathematcal entty called a body. Ths body conssts of an nfnte number of materal partcles. Shown

More information

COMPLEX NUMBERS AND QUADRATIC EQUATIONS

COMPLEX NUMBERS AND QUADRATIC EQUATIONS COMPLEX NUMBERS AND QUADRATIC EQUATIONS INTRODUCTION We know that x 0 for all x R e the square of a real number (whether postve, negatve or ero) s non-negatve Hence the equatons x, x, x + 7 0 etc are not

More information

A Fast Computer Aided Design Method for Filters

A Fast Computer Aided Design Method for Filters 2017 Asa-Pacfc Engneerng and Technology Conference (APETC 2017) ISBN: 978-1-60595-443-1 A Fast Computer Aded Desgn Method for Flters Gang L ABSTRACT *Ths paper presents a fast computer aded desgn method

More information

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

More information

Polynomial Regression Models

Polynomial Regression Models LINEAR REGRESSION ANALYSIS MODULE XII Lecture - 6 Polynomal Regresson Models Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Test of sgnfcance To test the sgnfcance

More information

LECTURE 9 CANONICAL CORRELATION ANALYSIS

LECTURE 9 CANONICAL CORRELATION ANALYSIS LECURE 9 CANONICAL CORRELAION ANALYSIS Introducton he concept of canoncal correlaton arses when we want to quantfy the assocatons between two sets of varables. For example, suppose that the frst set of

More information

Linear Feature Engineering 11

Linear Feature Engineering 11 Lnear Feature Engneerng 11 2 Least-Squares 2.1 Smple least-squares Consder the followng dataset. We have a bunch of nputs x and correspondng outputs y. The partcular values n ths dataset are x y 0.23 0.19

More information

Time-Varying Systems and Computations Lecture 6

Time-Varying Systems and Computations Lecture 6 Tme-Varyng Systems and Computatons Lecture 6 Klaus Depold 14. Januar 2014 The Kalman Flter The Kalman estmaton flter attempts to estmate the actual state of an unknown dscrete dynamcal system, gven nosy

More information

/ n ) are compared. The logic is: if the two

/ n ) are compared. The logic is: if the two STAT C141, Sprng 2005 Lecture 13 Two sample tests One sample tests: examples of goodness of ft tests, where we are testng whether our data supports predctons. Two sample tests: called as tests of ndependence

More information

The equation of motion of a dynamical system is given by a set of differential equations. That is (1)

The equation of motion of a dynamical system is given by a set of differential equations. That is (1) Dynamcal Systems Many engneerng and natural systems are dynamcal systems. For example a pendulum s a dynamcal system. State l The state of the dynamcal system specfes t condtons. For a pendulum n the absence

More information