This Small Business Innovation Research Phase I project is to develop and evaluate a promising new integration algorithm for solving stiff problems in real time. Such problems occur frequently in Aerospace, Chemical, Nuclear, Electric Power, and Automotive applications. In such applications, a common design technique is to combine a computer model of part of the system under evaluation (usually consisting of ordinary differential equations) with actual system hardware. Such "hardware in the loop" applications require real-time operation; the simulation of one second of system operation may take no more than one second of computer time. Some models contain very fast subsystems (e.g. a "tight" controller in a mechanical system or a small parasitic impedance in an electrical system). Such large eigenvalues (high frequencies, short time constants) require a very short computational step for stability. If the computer can't solve the entire set of system equations in this short time, real-time operation is impossible. The proposed integration algorithm, called the LIMP (Linearized IMPlicit) method, allows much larger steps to be taken for a given system, enabling real-time solution of many previously "intractable" problems. Potential commercial applications include the design of aircraft, missiles, chemical reactors, power generators, and automobiles.
