Créer une présentation
Télécharger la présentation

Télécharger la présentation
## ECE 530 – Analysis Techniques for Large-Scale Electrical Systems

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -

**ECE 530 – Analysis Techniques for Large-Scale Electrical**Systems Lecture 8: Advanced Power Flow Topics; Gaussian Elimination Prof. Hao Zhu Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign haozhu@illinois.edu 9/18/2014**Announcements**• HW 2 is due today • HW 3 is due Oct 2, Thur**Switched Shunts and SVCs**• Switched capacitors andsometimes reactors are widely used at both the transmissionand distribution levels tosupply or (for reactors)absorb discrete amounts of reactive power • Static Var compensators (SVCs) are also used to supplycontinuously varying amounts of reactive power • In the power flow SVCs are sometimes represented as PV buses with zero real power**Switched Shunt Control**• The status of switched shunts can be handled in an outer loop algorithm, similar to what is done for LTCs and phase shifters • Because they are discrete they need to regulate a value to a voltage range • Switches shunts often have multiple levels that need to be simulated • Switched shunt control also interacts with the LTC and PV control • The power flow modeling needs to take into account the control time delays associated with the various devices**NR Initialization**• A textbook starting solution for the NR is to use “flat start” values in which all the angles are set to the slack bus angle, all the PQ bus voltages are set to 1.0, and all the PV bus voltages are set to their PV setpoint values • This approach usually works for small systems • It seldom works for large systems. The usual approach for solving large systems is to start with an existing solution, modify it, then hope it converges • If not make the modification smaller • More robust methods are possible, but convergence is certainly not guaranteed!!**Power Flow with Optimal Multiplier**• Consider an n bus power system where S is the vector of the constant real and reactive power load minus generation at all buses except the slack, x is the vector of the bus voltages in rectangular coordinates: Vi = ei + jfi, and f is the function of the power balance constraints**Power Flow with Optimal Multiplier**• With a standard NR approach we would get • If we are close enough to the solution the iteration converges quickly, but if the system is heavily loaded it can diverge • Key idea: once NR method has selected a direction, we can analytically determine the distance to move in that direction to minimize the norm of the mismatch**Power Flow with Optimal Multiplier**• Optimal multiplier approach modifies the iteration as • Scalar m is chosen to minimize the norm of the mismatch F in direction x • Paper by Iwamoto et al “A Load Flow Calculation Method for Ill-Conditioned Power Systems” from 1981 shows mcan be computed analytically with little additional calculation when rectangular voltages are used m**Power Flow with Optimal Multiplier**• Determination of minvolves solving a cubic equation, which gives either three real solutions, or one real and two imaginary solutions • 1989 PICA paper by Iba(“A Method for Finding a Pair of Multiple Load Flow Solutions in Bulk Power Systems”) showed that NR tends to converge alongline joining the high and a low voltage solution**Solving Linear Systems**• We are concerned with the solution of linear systems Ax = b where A is an n by n matrix with elements aij, and x and b are n-vectors with elements xi and bi respectively • We are particularly interested when n is relatively large and A is sparse • A matrix is sparse if a large percentage of its elements have zero values**Sparse Matrices**• Sparse matrices arise in many areas, and can have domain specific structures • Symmetric matrices • Structurally symmetric matrices • Tridiagnonal matrices • Banded matrices • A good (and free) book on sparse matrices is available at www-users.cs.umn.edu/~saad/IterMethBook_2ndEd.pdf • ECE 530 is focused on problems that arise in the electric power domain; it is not a general sparse matrix course**Gaussian Elimination**• The best known and most widely used method for solving linear systems of algebraic equations is attributed to Gauss • Gaussian elimination avoids having to explicitly determine the inverse of A • Gaussian elimination leverages the fact that scaling a linear equation does not change its solution, nor does adding on linear equation to another**Gaussian Elimination, cont.**• Gaussian elimination is the elementary procedure in which we use the first equation to eliminate the first variable from the last n-1 equations, then we use the new second equation to eliminate the second variable from the last n-2 equations, and so on • After performing n-1 such eliminations we end up with a triangular system which is easily solved in a backward direction**Example 1**• We need to solve for x in the system • The three elimination steps are given on the next slides; for simplicity, we have appended the r.h.s. vector to the matrix • First step is set the diagonal element of row 1 to 1 (i.e., normalize it)**Example 1, Step 1**• Eliminate x1 by subtracting row 1from all the rows below it**Example 1, Step 2**• Eliminate x2 by subtracting row 2from all the rows below it**Example 1, Step 3**• Elimination of x3 from row 3and 4**Example 1 Solution**• Then, we solve for x by “going backwards”, i.e., using back substitution:**Triangular Decomposition**• In this example, we have: • triangularized the original matrix by Gaussian elimination using column elimination • then, we used backsubstitution to solve the triangularizedsystem • The following slides present a general scheme for triangular decomposition by Gaussian elimination • The assumption is that A is a nonsingular matrix (hence its inverse exists) • Gaussian elimination also requires the diagonals to be nonzero; this can be achieved through ordering • If b is zero then we have a trivial solution x = 0**Triangular Decomposition**• We form the matrix Aa using A and b with and show the steps of the triangularization scheme**Triangular Decomposition, Step 1**• Step 1: normalize the first equation**Triangular Decomposition, Step 2**• Step 2: a) eliminate x1 from row 2: • Step 2: b) normalize the second equation**Triangular Decomposition, Step 2**and we end up at the end of step 2 with**Triangular Decomposition, Step 3**• Step 3: a) eliminate x1 and x2 from row 3: • Step 3: b) normalize the third equation: =**Triangular Decomposition, Step 3**and we have the system at the end of step 3**Triangular Decomposition, Step k**• In general, we have for step k:a)eliminate from row k:b) normalize the kth equation:**Triangular Decomposition:Upper Triangular Matrix**• and proceed in this manner until we obtain the upper triangular matrix (the nth derived system):**Triangular Decomposition**• Note that in the scheme presented, unlike in the example, we triangularly decompose the system by eliminating row – wise rather than column –wise • In successive rows we eliminate (reduce to 0) each element to the left of the diagonal rather than those below the diagonal • Row-wise operations will work better for power system sparse matrices**Solving for X**• To compute x we perform back substitution**Upper Triangular Matrix**• The triangular decomposition scheme applied to the matrix A results in the upper triangular matrix U with the elements • The following theorem is important in the development of the sparse computational scheme**LU Decomposition Theorem**• Any nonsingular matrix Ahas the following factorization: where U could be the upper triangular matrix previously developed (with 1’s on its diagonals) and L is a lower triangular matrix defined by**LU Decomposition Application**• As a result of this theorem we can rewrite • Can also be set so U has non unity diagonals • Once A has been factored, we can solve for x by first solving for y, a process known as forward substitution, then solving for x in a process known as back substitution • In the previous example we can think of L as a record of the forward operations preformed on b.**LDU Decomposition**• In the previous case we required that the diagonals of U be unity, while there was no such restriction on the diagonals of L • An alternative decomposition is where D is a diagonal matrix, and the lower triangular matrix is modified to require unity for the diagonals**Symmetric Matrix Factorization**• The LDU formulation is quite useful for the case of a symmetric matrix • Hence only the upper triangular elements and the diagonal elements need to be stored, reducing storage by almost a factor of 2 (30)**Symmetric Matrix Factorization**• There are also some computational benefits from factoring symmetric matrices. However, since symmetric matrices are not common in power applications, we will not consider them in-depth • However, topologically symmetric sparse matrices are quite common, so those will be our main focus**Pivoting**• An immediate problem that can occur with Gaussian elimination is the issue of zeros on the diagonal; for example • This problem can be solved by a process known as “pivoting,” which involves the interchange of either both rows and columns (full pivoting) or just the rows (partial pivoting) • Partial pivoting is much easier to implement, and actually can be shown to work quite well**Pivoting, cont.**• In the previous example the (partial) pivot would just be to interchange the two rowsobviously we need to keep track of the interchanged rows! • Partial pivoting can be helpful in improving numerical stability even when the diagonals are not zero • When factoring row k interchange rows so the new diagonal is the largest element in column k for rows j >= k**LU Algorithm Without PivotingProcessing by row**• We will use the more common approach of having ones on the diagonals of L. Also in the common, diagonally dominant power system problems pivoting is not neededBelow algorithm is in row form (useful with sparsity!) For i:= 2 to n Do Begin // This is the row being processed For j := 1 to i-1 Do Begin // Rows subtracted from row i A[i,j] = A[i,j]/A[j,j] // This is the scaling For k := j+1 to n Do Begin // Go through each column in i A[i,k] = A[i,k] - A[i,j]*A[j,k] End; End; End;**LU Example**• Starting matrix • First row is unchanged; start with i=2 • Result with i=2, j=1; done with row 2**LU Example, cont.**• Result with i=3, j=1; • Result with i=3, j=2; done with row 3; done!**LU Example, cont.**• Original matrix is used to hold L and U With this approachthe original A matrixhas been replacedby the factored values!**Forward Substitution**Forward substitution solves with values in bbeing over written (replaced by the y values)For i := 2 to n Do Begin // This is the row being processed For j := 1 to i-1 Do Begin b[i] = b[i] - A[i,j]*b[j] // This is just using the L matrix End; End;**Backward Substitution**• Backward substitution solves (with values of y contained in the b vector as a result of the forward substitution) For i := n to 1 Do Begin // This is the row being processed For j := i+1 to nDo Begin b[i] = b[i] - A[i,j]*b[j] // This is just using the Umatrix End; b[i] = b[i]/A[i,i] // The A[i,i] values are ≠ 0 if it is nonsingular End**Computational Complexity**• Computational complexity indicates how the number of numerical operations scales with the size of the problem • Computational complexity is expressed using the “Big O” notation; assume a problem of size n • Adding the number of elements in a vector is O(n) • Adding two n by n full matrices is O(n2) • Multiplying two n by n full matrices is O(n3) • Inverting an n by n full matrix, or doing Gaussian elimination is O(n3) • Solving the traveling salesman problem by brute-force search is O(n!)**Computational Complexity**• Knowing the computational complexity of a problem can help to determine whether it can be solved (at least using a particular method) • Scaling factors do not affect the computation complexity • an algorithm that takes n3/2 operations has the same computational complexity of one the takes n3/10 operations (though obviously the second one is faster!) • With O(n3) factoring a full matrix becomes computationally intractable quickly! • A 100 by 100 matrix takes a million operations (give or take) • A 1000 by 1000 matrix takes a billion operations • A 10,000 by 10,000 matrix takes a trillion operations!