GB1013048A - Improvements in or relating to monitoring systems for fault detection - Google Patents

Abstract

1,013,048. Automatic controL INTERNATIONAL SYSTEMS CONTROL Ltd. April 28, 1964 [May 6, 1963], No. 17786/63. Heading G3R [Also in Division G1] Monitoring system for detecting faults in apparatus incorporating plural parts each constituting a single translating device having at lease one input and an output, the system including a comparator incorporating a stored program digital computer in whose program is stored information defining a required functional relationship between the output and input conditions of each such device; and means for supplying to the comparator signals representing the input and output conditions of each device; the comparator being arranged to compare intermittently for each device a first signal derived by operation of the computer with a second signal, such first signal representing the required performance of the device and the second signal correspondingly representing the actual performance thereof; with means provided to operate an alarm or perform other functions in response to the occurrence of a difference of predetermined degree between the desired and actual performance of a device as represented by the relevant first and second signals. Fig. 1 shows a control apparatus e.g. for boilers comprising plural equipments 1 consisting of measuring devices 2 responsive to several parameters e.g. feedwater flow, blower fan speed, water level; producting electric signals compared with the desired value signals in regulators 3 to operate control devices 4 for the appropriate parameters over two-way switches 5. Fig. 2 shows a typical regulator providing output K 1 E+ #00 Edt + K 3 dE/dt where E is the difference between the applied signals, t is time, and K 1 , K 2 , K 3 are constants; wherein network 6 applies an output representing the difference between the measured value and desired value signals to parallel amplifying, integrating, and differentiating networks whose outputs are added at 10 to operate control device 4 over switch 5. An electronic digital computer 11 has a program of information defining the desired relationship between input and output signals of each network of the apparatus. Analog input scanning unit measures simultaneously the input and output signals of each network in torn and converts them to digital form so that computer 11 calculates the desired relation between the input and output signals of the network from program information and generates a signal representing the desired output signal for the measured input signal of the corresponding network when correctly operating. This signal is then compared in the computor with a signal representing the measured output signal of the network, and if the ratio is withinpredetermined limits the next succeeding network is similarly tested. If the ratio is outside such limits the appropriate network is considered faulty, and alarm 13 is operated to indicate failure of the network. Also the appropriate switch 5 is operated to disconnect the equipment 1 containing the failed network from its associated control device, and computer 11 is switched to produce from the input signal of the disconnected equipment 1 and its programmed information a signal corresponding to the output of the faulty network when operating correctly, which is fed over switch 5 to the appropriate control device 4 after conversion to analogue form at 14 so that the computer assumes the function of the failed equipment 1. The computor may be programmed to make a quick check of each network in turn e.g. by fast scanning the input and output signals and checking that rates of change thereof fall within preset limits to determine possible failure, and to make a full test only when possible failure is indicated. Alternatively the computer may generate from a measured actual output of a network the desired input signal for correct operation, which is then compared with the actural input signal; or a signal representing a desired relation between input and output signals for correct operation, which is then compared with a computed signal representing the actual relationship there-between. The computor may assume only the function of the faulty element instead of that of the entire equipment comprising it, and fault detection may be effected only by measurement of the input and output of the entire equipment.