WELL-CONDITIONED OBSERVER DESIGN FOR OBSERVER-BASED MONITORING SYSTEMS

Observer-based monitoring systems for machine diagnostics and control are receiving increased attention. These observer techniques can estim ate process and machine variables from inexpensive, easy to install re mote sensors based on state-space models of the machine structure betw een the machine variables of interest and the location of the remote s ensors. Unfortunately, these observers can be ill-conditioned and this leads to poor performance. The authors have previously shown that obs erver performance can be represented by a single performance index, th e condition number of the eigensystem of the state observer matrix and that there exists an upper bound for the index in non-normal matrices and the bound can be determined by the structure and eigenvalues of t he observer matrix. In this paper a design methodology for synthesizin g well-conditioned observers is proposed based on the upper bound of t he performance index. The methodology is based on the fact that a smal l upper bound guarantees small values of the performance index. A well -conditioned matrix form is defined and a block-by block design strate gy to produce a well-conditioned observer matrix is presented. A compl ete design procedure for well-conditioned deterministic state observer s is given for the single-output case. The design strategy is illustra ted with an example that shows that the proposed well-conditioned obse rver performs much better than an observer designed with traditional p ole placement techniques.