Environmental Control

Togai InfraLogic, Inc. has completed a Phase I Small Business Innovative 
Research (SBIR) program for NASA on application of fuzzy control to manned 
spacecraft environments and building energy saving efforts.  A Phase II effort 
has been approved and will begin mid-1993.  Phase I was a feasibility 
activity.  Phase II provides implementation and product development 
opportunities  for commercial follow-on.  Major features of this program are:

o  Development supported by NASA SBIR program
o  Sensor-based comfort control
o  Simplified multiple zone temperature control and vent control
o  Energy saving using fuzzy decision system 

A principal product of this work is a comfort sensor based on the work of P.O. 
Fanger who defined ranges of variables such as dry air temperature, humidity, 
air speed, radiative temperature, activity level, exposure time, and clothing 
that characterize comfort.  Sensing dry air temperature, air speed, radiative 
temperature, and humidity is easily accomplished with existing sensors.  To 
augment these for establishing comfort, we are adding CO2 level,  infrared 
intensity, and infrared differential intensity sensors to support inferring 
activity levels, number of people present, and clothing levels from remote 
sensors using a fuzzy rule-based decision system.  Should expectations be 
realized, control of temperature and humidity in a zone will serve comfort as 
well as operational needs.  For example, exercise areas could be run dry and 
warm while astronauts are exercising and at nominal conditions at other times.
With a comfort sensor available, the only remaining question is the need for 
and the ability to regulate several zones.  With zone control, independent 
regulation of temperature and humidity becomes useful 

A second application of fuzzy control is reducing energy requirements by 
control of computers and lights.  The goals of reducing energy input and not 
interfering with computer and light users are conflicting.  Small savings are 
possible from a closer dynamical match of the heating/cooling system capacity 
to heat loads and running motors at optimal efficiency.  Except for the use of 
such techniques, energy savings are only realizable by decreasing power input 
into an area.  In this application, the fuzzy system is a decision support 
system for discerning times when activity in an area or use of a computer is 
such to allow turning off lights and at least portions of the computer system 
without interfering with work being performed.  The use of adaptive fuzzy 
controllers is likely to be useful in this area to permit any correction by a 
user to influence the decision making by the fuzzy controller.

Three demonstrations were developed during Phase I.  The first demonstration 
shows the use of fuzzy modeling to control a flow based on two level sensors.  
The idea is to simplify the data interface to the microprocessor for 
temperature measurements to enable use of multiple zones without undue cost.  
For heating, ventilation, and air conditioning (HVAC) applications, the level 
controlled is temperature and the sensor inputs are the conditions of two 
bimetallic switches that can be set to close at low and high settings.  The 
controller estimates the temperature and heat load and converges to those 
conditions using proportional-integral (PI) fuzzy control for a constant heat 
load.  A second demonstration  illustrates the ability of the controller to 
track a sinusoidally varying heat load that simulates the daily variation of 
solar heating.  The third demonstration is for a vent controller that directs 
conditioned air into a specific area.  Such a vent controller changes the air 
flow and thus the back pressure on a central distribution fan. In concert, a 
group of such controllers can provide a basis for fan speed control.

These systems are being integrated into a simulation of a test area in Building 
12 at Johnson Space Center called the Software Technology Laboratory.  The 
purpose of this simulation is to evaluate the behavior of the control system 
under varying building occupancy and heat loading due to seasonal variations. 
It will allow testing various control schemes as well as provide a baseline 
model for the current system.  The outside conditions, number and type of 
sensors, sensor set points, lights on, number of people, and number of 
operating computers can be set by a user of the simulation to conduct "what-if" 
studies.  These settings can be made on-line or under file control. 
