DocumentCode :
985070
Title :
A multiple DSP environment for on-line signature analysis of plant instrumentation
Author :
Hassapis, G.D.
Author_Institution :
Dept. of Electr. Eng., Aristotelian Univ., Thessaloniki
Volume :
39
Issue :
1
fYear :
1990
fDate :
2/1/1990 12:00:00 AM
Firstpage :
126
Lastpage :
128
Abstract :
It is shown that a typical multiprocessor architecture with four digital signal processors (DSPs) can be used to identify more than one faulty sensor within the time limits imposed by the response time of the process control loops in which the sensors belong. This is achieved through a heuristic algorithm, running on a host computer and assigning to the four processors in a suboptimal way the algorithms used to detect the faulty sensors. Testing of the heuristic algorithm has been made possible by simulating on a computer the operation of the multiprocessing system. If this system is properly programmed, it can detect faulty sensors in a large number of process control loops before the next update of the output of each control loop occurs. Thus the respective controllers can avoid using the data supplied by the faulty sensor and instead can use a known output signal value which will drive the process to a safe state. The program execution time requirements can be met if the individual program routines are properly allocated to the four processors. A heuristic algorithm has been developed to carry out this allocation
Keywords :
automatic test equipment; automatic testing; computer architecture; computerised signal processing; multiprocessing systems; optimal control; process computer control; computerised signal processing; digital signal processors; faulty sensor; heuristic algorithm; multiple DSP environment; multiprocessing system; multiprocessor architecture; online signature analysis; plant instrumentation; process control loops; response time; suboptimal; Computer architecture; Delay; Digital signal processing; Digital signal processors; Fault detection; Fault diagnosis; Heuristic algorithms; Process control; Signal processing algorithms; System testing;
fLanguage :
English
Journal_Title :
Instrumentation and Measurement, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9456
Type :
jour
DOI :
10.1109/19.50430
Filename :
50430
Link To Document :
بازگشت