Scribd
Upload
184 views2 pages

Influence Coefficient Matrix Guide

The document discusses the relationship between the input and response of a linear system using matrices. It introduces an influence coefficient matrix [C] that can model an unknown system matrix [Q] by recording the system's responses to individual inputs. The [C] matrix is constructed by applying separate unit inputs to obtain individual response vectors, forming the columns of [C]. Algebraically, this shows [C] is equivalent to the discretized system matrix [Q], even though [Q] is unknown. The influence coefficient technique allows obtaining responses from the system as a "black box" without explicitly knowing the system matrix.

Uploaded by

someonefromparadise6855
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
184 views2 pages

Influence Coefficient Matrix Guide

The document discusses the relationship between the input and response of a linear system using matrices. It introduces an influence coefficient matrix [C] that can model an unknown system matrix [Q] by recording the system's responses to individual inputs. The [C] matrix is constructed by applying separate unit inputs to obtain individual response vectors, forming the columns of [C]. Algebraically, this shows [C] is equivalent to the discretized system matrix [Q], even though [Q] is unknown. The influence coefficient technique allows obtaining responses from the system as a "black box" without explicitly knowing the system matrix.

Uploaded by

someonefromparadise6855
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd

F - 1 [Q] {x} {r} (F.1) Figure F.1 An arbitrary system subjected to certain loads and its response.

Appendix F: INFLUENCE COEFFICIENT MATRIX &onsider an arbitrary system as depicted in Fig. F.1. Let the vector {x} represe nt a set of input applied simultaneously to the system and the vector {r} represent the r esponse of the system. The response {r} can also be considered as the difference between the fi nal state of the system and the initial state in which no input is applied. The relationship betw een the input and the response for a linear system is shown in Eq. (F.1). The matrix [Q] contains all the properties and characteristics of the system, i. e. it defines the system. For a complex system, the matrix [Q] is unknown. Relating this to pure displacement shape control, displacement values {r} are pr escribed and the applied voltages {x} are to be determined. The relationship between thes e two electromechanical state variables is represented by the analytically unknown [Q]. The concept is t o construct an Influence Coefficient matrix [C] to model the real system represent ed by the unknown [Q]. When the system is linear, [C] is constructed by recording the responses due to each independent actuation. For example when there is three actuation degrees of free dom {x1, x2, x3}, then obtain 3 separate responses, {q1}, {q2}, {q3} from 3 different unit input a t the corresponding F - 2 [Q] 1 0 0 {q1} ; [Q] 0 1 0 {q2} ; [Q] 0 0 1 {q3} ; (F.2) [C] {q1} {q2} {q3} (F.3) [C] [Q] 1 0 0 [Q] 0

1 0 [Q] 0 0 1 [Q] 1 0 0 0 1 0 0 0 1 [Q] [I] [Q] (F.4) states as shown in Eq. (F.2) Then use the {q} vectors to form columns of the [C] matrix which is the influenc e coefficient matrix as shown in Eq. (F.3) Note that when Eq. (F.2) is inserted into Eq. (F.3) then [C] is equivalent to th e discretized system matrix [Q] as shown in Eq. (F.4). Algebraically, this proof seems trivial, but note that the Q-matrix is unknown, and the Cmatrix is obtained by sampling the responses at the 3 states. So the main mechanism in the IC technique is the ability to obtain responses from the system due to given inputs by using [Q] as a black box operator (e.g. the FE process) although the system [Q] itself is not explicitly known. Several items to note: 1. The method presented above can be applied to a system with n states in genera l although it was illustrated as a 3-state system. 2. The [C] and [Q] matrices are in general non-square matrices. As shown in Fig. F.1 there are 3 degrees of actuation freedom and 7 degrees of response freedom. The latter dep ends on how many points of interest the analyst wishes to observe. 3. Following from the last point, it is obvious that in general, [C] is non-inve rtible. Thus the solution Eq.(F.1) cannot be obtained by inverting [C] or the system matrix [Q] e ven if it is known.

You might also like