曝光台 注意防骗
网曝天猫店富美金盛家居专营店坑蒙拐骗欺诈消费者
not repeated since the setup information will never change. Following setup,
the program enters the initialization loop by set.ting the interrupt timer
(fig. 6). This loop does not have a real-time cycle constraint (it has no time
dependency) but it repeats every 50 msec. Here it executes INLET (fig. I0)
which calculates the ambient conditions based on the altitude and Mach number.
Then it goes to EMODEL (fig. II) which determines the base points and matrix
elements by evaluating polynomials whose coefficients are functions of the
ambient conditions. The scheduled values of engine variables are calculated
in the subroutlnes RPFAND (fig. 12) and RPLIMD (fig. 13) which are called from
EMODEL. The operating point is requested using MINDS. The operating point is
automatically initialized to sea-level static, standard day conditions, 83 °
PLA. Base point values at this condition are also stored as the initial control
values for that operating point in the open loop mode. Any extra time In
this loop is used by MINDS to accept inputs from the user. He can change altitude
and Mach number and the next time through the loop everything is recalculated
for the new conditions. Since everything in the initialization loop is
calculated directly, the loop need only be executed once after a change is made
for the values to be correct. The user can also set the switch to go from the
initialization loop to the run loop while in MINDS. Figure 14 shows the program
flow as the initialization/run switch, RLOOP, is set and reset. The
update interval is short enough to essentially guarantee that the loop will be
executed at least once after the conditions are changed to obtain the correct
values before the switch can be set. The program is ready to be used interactively
once the MINDS prompt (>) appears.
Setting the appropriate software switch puts the program into the realtime
mode. The run loop consists of the dynamic routines. This loop has an
update interval of 12 msec and during that tlme the control input routine,
actuator routine, the system evolution routine (numerical integration), and
the output signal routine all execute. The first section receives the control
signals from the CIM unit and converts the scaled integers to real numbers.
ACTUAT (fig. 15) takes the real commanded values and evolves the actuator
models to their value at the current time step. This output is used by EVOLVE
(fig. 16) to integrate the differential equations describing the engine itself.
Over tlme, the numerical integration will bring the simulation from its previous
steady-state point up to the new steady-state point with a linear, nonrealistic
transient. The new steady-state point is, however, accurate and
realistic. After EVOLVE executes, the engine outputs, actuator feedbacks, PLA,
and the ambient conditions are converted to scaled integers and sent via D/A
converters to the CIM unit. The I/0 sections are part of the multiplexer
interrupt service routine section of the executlve (fig. 17). Any spare time
is used by the messagegeneration routine or MINDS. The messagegeneration
routine takes prlority over MINDSif it needs to execute but it is only used
to print out error messages. A more In-depth description of the simulator's
operation is given in appendix A.
Many of the routines listed above call their own subroutines which do
table lookups or some other type of calculation. The relationships are shown
in figure 18. A complete list of the routines with a description of each
appears in appendix B.
Exception Handling
There are three noncatastrophic exceptions which, if they occur, will
cause incorrect operation of the simulator. They are: (1) floating point,
<2) divide by zero, and (3) update failure. The first two produce an interrupt
and are handled by interrupt service routines (figs. 19 and 20). Update failures
are detected by a flag check routine within the timer interrupt service
routine (fig. 21). When the timer signaling the start of the real-time run
loop interrupts the simulation, the service routine checks the update failure
flag. Since the flag is reset near the end of the multiplexer interrupt serv-
Ice routine (fig. 17), a reset update failure flag indicates no problem. However,
if the flag is still set at the start of the timer service routine, it
means that the cycle was unable to finish the previous time through the loop
and an_update failure is declared.
One of the results of these interrupt service routines is to give the user
an indication that the error took place by printing a message to the user terminal.
This printing is done in the time remaining at the end of the real-time
loop. Printing a message is a slow process and may take several cycles of the
中国航空网 www.aero.cn
航空翻译 www.aviation.cn
本文链接地址：
A_real-time_simulator_of_a_turbofan_engine(5)
