Matthew C. Ruschmann

Probing the NASA generic transport aircraft in real-time for health monitoring

This brief paper describes the development of a pulse compression probing algorithm for real-time monitoring of nonlinear dynamic systems. The probing outputs of a system being monitored are estimates of the sequences of Markov parameters at the current operating point. By imposing signal to noise ratios at both the controlled outputs and the probing outputs, the probing inputs are designed to be non-intrusive and with the intention that the resulting probing outputs are unaffected by other applied signals to the system. The probing algorithm is implemented on a Virtex-5 field-programmable gate-array evaluation platform to expedite real-time processing, and is applied to monitoring the health of the NASA Generic Transport Aircraft Model (GTM). Samples of probing output residuals are evaluated using the Hotellings T2-test for change detection when significant noise is present in the system. This brief paper focuses its discussion on the method of rapid collection of independent samples to reduce detection delay.

Data-Availability-Constrained Placement of PMUs and Communication Links in a Power System

This paper presents a solution to placing a minimum number of phasor measurement units (PMUs) and communication links in a power system so that the steady-state availability of synchrophasor data at each bus meets a prescribed level. For this purpose, a Markov model suitable for the evaluation of synchrophasor availability is built. The model contains a set of binary decision variables representing actions or inactions to maintain PMUs or to utilize communication links. The decision variables are solved to minimize an expected cost for maintaining a PMU network and utilizing data links for synchrophasor inference, subject to a set of synchrophasor availability requirements at the buses. The new aspects of development in this paper concern first the selection of decision variables so that their solutions immediately imply whether a PMU should be placed at a bus and whether a communication link should be established between two buses and, second, the solution to PMU network placement in a large-scale electric network. In addition, permanent and random intermittent PMU outages are modeled. A five-bus power system is used to demonstrate the problem formulation, and placement results are presented. A method to treat a large power system as interconnected smaller power systems with common boundary buses is proposed to address the curse of dimensionality encountered. Each smaller power system formulates an independent optimal placement problem with constraints on matching boundary placement. The steps involved are illustrated through a 14-bus system. The complexity and optimality of a divided problem are compared to that of the original problem.

Redundancy architecture design of PMU networks for availability of synchrophasors

Wide-spread introduction of phasor measurement units into the power grid increases the opportunity to utilize the supplied synchrophasor estimates for potentially significant improvement of monitoring and control performance. On the other hand, a growing grid reliance on synchrophasor estimates may accentuate the grids vulnerability whenever such estimates become unavailable or faulty. This papers studies the redundancy architecture of a PMU network for which required synchrophasor availability at each location is specified. A Markov chain containing unspecified decision variables is established to allow the formal computation of the steady-state availability of synchrophasors. The decision variables are selected to minimize the cost of requiring communication links among the PMUs and the restoration of failed PMUs while maintaining the prescribed synchrophasor availability. The optimized decision variables determines the structure of the Markov chain, which in turn dictates the redundancy architecture of the PMU network. The redundancy architecture design method is illustrated through a 3-bus power system.

Probing the NASA Generic Transport Aircraft in Real-Time for Health Monitoring

This paper investigates the effectiveness of a pseudo-random binary sequence as a probing signal for real-time monitoring of a closed-loop nonlinear GTM aircraft (NASA Generic Transport Model). Two forms of probing outputs based on two probing algorithms are studied: a small signal pulse response matrix at trim using pulse-compression, and a linearized state-space parameter set using regression. A new structure of the pulse-compression method is introduced to simultaneously monitor the small signal pulse sequences for all input-output channels of the aircraft, and the probing algorithm is implemented on a Virtex-5 FPGA evaluation platform for speed. The paper also exploits the role of the probing signal in regression to identify a local state-space model for GTM as the first step in our effort toward real-time faulty model identification to support flight envelope assessment. The pseudo random sequence probing compares favorably to white noise and doublet excitations. Probing outputs in both forms are shown to be highly sensitive to the occurrence of actuator faults in the plant.

Toward a Highly Available Modern Grid

The concept of fault-coverage and how it affects the availability of a dynamic grid is explained through a two-area power system represented by an aggregated swing model. Fault-coverage is intended to serve as a criterion for decisions in redundancy management to benefit system availability upon occurrence of a disturbance due to loss of equipment. The criterion allows the incorporation of formal measures of uncertainties associated with real-time fault diagnosis, as well as formal control performance measures. Also investigated is the effect of the availability of a modern grid’s supporting structure on the availability of the grid, with focus on a network of measurement units. The focus stems from the recognition of a greater need for real-time diagnosis and control in a modern power grid. A redundancy architecture design problem is formulated based on a Markov model of a measurement network, and a solution is presented that minimizes the number of phasor measurement unit (PMU) restorations and the usage of communication links to a PMU while maintaining a prescribed data availability at any PMU. A 3-bus/3-PMU network is used as an example to explain the formulation and the solution of the redundancy architecture design problem.

Fault-tolerant control of a distributed database system

Optimal state information-based control policy for a distributed database system subject to server failures is considered. Fault-tolerance is made possible by the partitioned architecture of the system and data redundancy therein. Control actions include restoration of lost data sets in a single server using redundant data sets in the remaining servers, routing of queries to intact servers, or overhaul of the entire systemfor renewal. Control policies are determined by solving Markov decision problems with cost criteria that penalize system unavailability and slow query response. Steady-state system availability and expected query response time of the controlled database are evaluated with the Markov model of the database. Robustness is addressed by introducing additional states into the database model to account for control action delays and decision errors. A robust control policy is solved for the Markov decision problem described by the augmented state model.

Fault-tolerant control of power grids for security and availability

This paper presents a framework for fault-tolerant control of a modernizing electric power grid. The motivation for the new development hinges on the potential benefits of utilizing fast control devices and high sample rate measurement devices to enhance the grids resilience beyond that offered by the traditional protection devices. A set of quantifiable security indices based on a fault-coverage concept is used for selection of control actions and the times to exert them. Viewed as a discrete state stochastic process, the grids long-run availability is shown to be an increasing function of the security indices, whereas the security indices are decreasing functions of the level of uncertainty in the grid state information, the the level of imprecision of its dynamic model. An example of an aggregated two-area power system is used to demonstrate the principle and the computation involved in achieving a minimum risk control. It also illustrates how uncertainties cause delay in control actions and reduction in the critical clearance time in order to maintain a prescribed level of security.

Pulse-compression probing of small signal characteristics for nonlinear systems in dynamic operating conditions

This paper is concerned with real-time monitoring of a nonlinear system using pulse-compression probing, which evaluates the pulse response produced by compressing a small elongated random probing sequence around an operating point. The focus is to study the extent and the condition under which the pulse-compression principle is upheld when the system undergoes transients between equilibria. The paper uses an incremental, linear time-varying representation of the nonlinear system not only to establish the validity of the pulse-compression principle under the uniform bounded-input bounded-output condition, but also to assist in determining the parameters of probing signals. Using the incremental representation, probing outputs are accurate regardless of knowledge about the trim condition. The results from pulse-compression probing in the longitudinal axis dynamics of the NASA Generic Transport Aircraft Model (GTM) are reported. The probing algorithm proves effective during changing operating conditions caused by a thrust command and by stuck faults in the outer elevators.

Online estimation of linear state-space parameters of the Generic Transport Model for a set of structural damage scenarios

This paper reports the results of online state-space parameter estimation for a nonlinear Generic Transport Model during changing operating conditions and a set of six damage scenarios including lost control surfaces. The motivation is to provide rapid and accurate model information for use in online assessment of the flight envelope. Providing this information to the pilot and control systems is critical for aircraft safety. The parameters are estimated based on a linearized aircraft model using the modified sequential least squares algorithm. An excitation signal is proposed to reduce linear dependencies among the data. Common tuning mechanisms are discussed to improve parameter estimation during upset conditions.

Optimal control of a database with reduced & full state models

A Markov decision problem is formulated and solved for a redundant database system. The objective function is expressed in terms of the expenditure of average time required for the system to recover from server failures or to serve queries. Control actions include data restoration and system overhaul, and are taken based on the state of servers and the service demands. The steady-state availability of the database and the expected query response time are quantified under optimal control policies derived from several sets of cost criteria for the case of reduced state model that disregards the service demand information, as well as the full state model.