GB1313861A - System and method for optimizing processor or equipment profit - Google Patents

Abstract

1313861 Automatic retroactive control of density INDUSTRIAL NUCLEONICS CORP 15 May 1970 [2 June 1969] 23754/70 Heading G3R The average value of an output parameter of a manufacturing processor e.g. a cigarette making or paper making machine, is controlled in a manner so that it is as close as possible to a predetermined limit value to achieve optimum economic profit, by producing a signal indicative of the spread of values of the output parameter (e.g. density of cigarette), and controlling the average value accordingly so that for larger spreads the percentage output for which the parameter has a value outside the limit value is greater than for smaller spreads, see Fig.2. Thus the average value is adjusted toward said limit value (lowest cigarette weight acceptable to consumer) to produce a larger percentage defective output so long as profit is thereby increased, and is adjusted away from the limit value if the profit decreases due to too high a percentage of defective output. A measure of the spread of values or 'standard deviation' is obtained by finding the percentage of cigarettes having values greater or less than the lower limit value LL below which a cigarette is considered defective. In Fig.1, tobacco is fed by a conveyer belt 14 to a rod former 15 in quantities controlled in accordance with the outlet of a feeder 13 and the vertical position of a knife 16 which is varied by a controller 17 having tacho-feedback. A cutter 20 produces predetermined lengths. A tachogenerator 21 produces one pulse for each length of rod of desired cigarette length. Rod density is detected by a radiation type gauge 23 which produces a D.C. voltage inversely proportional to density. A gauge network 24 derives therefrom a positive or negative analogue output voltage proportional to the direction and magnitude of deviation from standard weight/unit length set on a unit 26. This output is fed to a reject classifying network 28 which energizes a solenoid 46 operating a kicker 45 to reject below limit LL cigarettes the tobacco from which is reclaimed and returned to the feeder. The output of 24 is also fed to a summer 29 where it is combined with either a manually set deviation from standard weight signal at contact 32 from a source 34 or signal at contact 33 from an automatic control network 35, which signals at 32 and 33 act to change the target average value of standard weight to maximize profit. The output of summer 29 is accordingly fed to the controller 17 to control the knife 16. To determine the percentage defects, a counter 71 counts the output of tachogenerator 21 and a timing generator 72 produces three out of phase pulses M 1 , M 2 , M 3 the leading edge of M 1 and trailing edge of M 3 coinciding with the leading and trailing edges respectively of each one thousandth pulse from the tacho. A counter 73 stores the 'defect' signals from a trigger 43, the M 1 pulses terminating the count of counter 73 so that the count therein is the defects/1000 cigarettes. A digital/analogue converter 74 produces an output voltage significant of the percentage defects and maintains this output until the next M 1 pulse arrives. The D/A output is fed to a Safety Limit comparator 76 and a network 75 which automatically performs an initial computation to produce a change in average weight for maximum profit and operates an alarm if percentage defective is outside a high or low limit thereof. The target (average value) deviation from the lower limit LL calculated for the immediately preceding 1000 cigarettes is compared in a network 86 with that for a previous 1000 cigarettes. This network 86 is responsive to the current target deviation signal from summer 83 and includes a feedback network 88, 89, 91 responsive to the previously computed target optimum i.e. the output of a sample and hold memory 87 significant of percentage deviation from standard weight to achieve maximum profit for preceding 1000 cigarettes. This output is subtracted from a lower limit LL signal in subtractor 91. The resulting output of 91 is compared in subtractor 92 with the current percentage deviation signal from subtractor 83. The output of network 86 is derived from the sample and hold type analogue memory 87 having an input responsive to each M 3 pulse so that the signal stored therein is updated once every 1000 cigarettes. The output of memory 87 is significant of per cent of standard weight and this is converted by a multiplier network 84 into a voltage significant of the deviation of desired target weight from standard weight and this is applied to the selector switch contact 33 as described above. A meter 95 monitors this voltage and it is stated that alternatively the standard weight deviation input to summer 29 may be controlled manually in response to meter observations.