Modeling and Control Studies of an Integrated Biological Wastewater Treatment System 2005-01-2963
From 1997 to 2001, the third author worked with a team of engineers at JSC to develop the requirements and basic design for the Bioregenerative Planetary Life Support Systems Test Complex, or BIO-Plex. Under the Advanced Integration Matrix (AIM) Project, this earlier effort is extended to an investigation of methods and approaches for Advanced Systems Integration and Control. The intent is to understand and validate the use of software as an integrating function for complex heterogeneous systems, particularly for Advanced Life Support (ALS), in the context of support of mission operations. Preliminary investigations undertaken in the summer of 2004 indicate that integration of controls for the type of dynamic, non-linear, closed-loop biological systems under investigation for ALS systems require a different systems engineering approach than that required for traditional avionics systems. Under AIM, the authors are currently involved in the development of a testbed to elicit the controls integration questions and criteria that will enable evaluation and comparison of controls technologies for ALS systems. This paper will discuss the current status of these efforts.
In traditional avionics engineering schedules, the software requirements specification follows the hardware system design review. In contrast, process control synthesis requires that the software and hardware be designed co-currently. A particularly important difference is that the hardware must be designed for controllability of the process; this forces control design to precede hardware requirements. Because software is the embodiment of the control design, this is a reversal of traditional spacecraft design. The AIM testbed is being developed to explore the ramifications of this engineering approach on requirements definition, modeling, design, testing, and integration. The focus is directed to closed-loop life support where the impacts are most obvious. Closed-loop life support, whether biological or physico-chemical, has greater similarities to process control systems than to flight avionics systems.
There are several steps that must be followed to system engineer such a process. Several models must be developed. A controls-relevant model, based on fundamental kinetics of the process must be developed to determine open-loop stability of the process in the absence of controls. That model has to capture the basic dependant and independent variables of the process, which are determined by the major phenomena of the process, but also by the choice of the objectives to be controlled. Observability of the control variables, and finally controllability of the process has to be validated by the model. Until these steps are completed, there are no assurances that a control system can be realized or that systems integration is feasible.
These traits are design dependent and sensitivity to various design parameters is process dependent. The accuracy of these models must be confirmed by experiment, as does their sensitivity. At this stage we can begin to review the suitability of various control architectures for system integration.
Currently, the process of building the testbeds to validate and refine our models is in progress. Once that is done, the investigation of the sensitivity of system integration requirements to operations and design parameters will begin.