I
by R. F. Wall Monsanto Chemical Co.
INSTRUMENTATION A
W O R K B O O K
B.t.u. Recorders in Process Control B.t.u. measurement may make possible process control where other methods fail
M
ISUREMENT
of
the
rate
or
quantity of heat energy used or evolved from a process is infrequently used for process control, but may often be valuable. B.t.u. is an energy measurement, and directly useful in power plants or in refrigeration systems where energy is the quantity of interest. It is also useful in processes where rate of energy transfer is indicative of a desired measurement, often the rate of a reaction, and through integration the degree of completion of a reaction. It is probably most effective where methods of direct measurement of reactant or product composition are impractical, as where precipitate formation, extremely viscous materials, or formation of a slurry, etc., make sampling impossible. Occasionally materials being processed are mixtures of no definite composition, but the information necessary for control is related to the quantity of heat involved in the processing. B.t.u. measurements m a y be valuable in any application where the quantity of heat involved is significantly related to processing. These measurements can be simply made from an instrumentation standpoint with equipment that is conventional, reliable, and moderate in cost. Determination of B.t.u. involved in a process system is derived from measurement of the temperature difference of a stream flowing in and out of the system and the rate of flow of the stream, with a simple analog computer to take the product of temperature difference X flow X a constant (the constant including the heat capacity and density of the fluid and instrument constants). T h e product of these terms gives the B.t.u. rate, and can be integrated for total quantity of heat transferred over a period of time.
Temperature difference and flow are measured by conventional wellproved instruments, and the analog computation required is simple, as illustrated in the diagram of the Hays B.t.u. meter circuitry. Several companies manufacture complete B.t.u. meters, or a user can combine components if inclined toward gadgetry, though this probably cannot be justified economically. There appears no reason why information transmitted from already installed temperature and flow instrumentation could not be used in a B.t.u. computing unit. T h e accuracy of the B.t.u. indication will equal the summed errors of the individual instrumental measurements—i.e., temperature difference and flow—which will dominate as the precision of the analog computer can be made m u c h better. Flow measurements are not noted for great accuracy and may impose a limitation. Orifice-type flowmeters require that the square root of the signal be taken to provide an output linear with flow. Magnetic or turbine-type flowmeters are intrinsically linear with flow and permit more variation in flow rate. T h e accuracy of differential temperature measurements is generally considered to be good. M u c h of the error in the B.t.u. indication will be derived from the stream being measured, for the constant in the B.t.u. equation includes heat capacity and density, and these vary with temperature. Stream composition changes will usually cause variations in heat capacity and density, making a B.t.u. measurement impossible. T h e most successful process measurements of B.t.u. are probably those made on the water stream across a heat exchanger, as shown in Foxboro's I/EC
F E A T U R E
E C
application schematic. This eliminates the characteristics of the process stream and permits the measurement to be made on water, a material of accurately known characteristics. T h e best instrumental accuracy will probably be obtained with this type of system. Control accuracy is something else, as this will include the effect of random (unmeasured) heat transfer and the degree of correlation of B.t.u. measurement with the desired control parameter. Perhaps the factor most limiting control accuracy is the random or unmeasured quantity of heat transferred between the system under study and the environment. If the system upon which the measurement is made is large and poorly insulated and the temperature gradient to the atmosphere is considerable, the random heat transfer is relatively large, and if at the same time the quantity of heat absorbed or evolved in the processing to be controlled is small, a B.t.u. measurement will probably be so inaccurate as to be meaningless. If the heat evolved or absorbed in the processing is large and random heat transfer is small, accuracy may be high and very satisfactory process control feasible. If in addition the conditions of operation are held constant, reproducibility and precision may be significantly better than accuracy and excellent: control obtained. A B.t.u. measurement may be useful for process control if the stream upon which the measurement is m a d e is essentially constant in heat capacity and density; if the rate or quantity of heat measured is large compared to the random heat transfer of the system; and if the measurement of heat energy is significantly correlated with a rate or degree of completion of a process under consideration. T h e measurement is directly applicable in the sale of energy in heating or refrigeration. T h e indirect application of B.t.u. measurement to process control is
W
ORKBOOK FEATURES
71 A
I/EC
INSTRUMENTATION
«
Λ Workbook
Feature
INFLUENCE POTENTIOMETERS
VOLTAGE SUPPLY TRANSFORMER
FLOW
BTU AMPLIFIER BTU PEN DRIVE ~ MOTOR BTU CALIBRATION
CAM Simplified to its essentials, this schematic of the Btu M e t e r electrical circuitry shows the interdependence of f l o w a n d t e m p e r a t u r e difference. If either o f these values is reduced to z e r o , t o t a l Btu signal is also reduced to zero since a w h o l e n u m b e r times zero yields zero.
of interest here, and has apparently received little attention, for little has been published on specific appli cations and the B.t.u. measurement seems to have been little used. One of the most appropriate areas of application would appear to be rate or degree of polymerization, as polymerization processes most com monly involve materials that are viscous, emulsions, or slurries, and the problems involved in continuous process analyses on this type of sample are prohibitive. Satisfac torily operating sample systems for handling this type of sample on a continuous basis are difficult to impossible. As far as is known, B.t.u. measure FOXBORO TYPE 2 8 ELECTRICAL FLOW TRANSMITTER
ments have been applied only to liquid streams, although in theory there is no reason why they could not be made on a gas stream, with pressure held constant or included in the analog computation. The heat capacity of a gas will be considerably less than that of the equivalent volume of liquid, and random heat transfer may be con siderably more significant as a cause of error. Sample techniques for gas streams are successful, and con ventional analysis instrumentation can usually provide much more specific control information, so that B.t.u. measurements on gas streams, though theoretically feasible, seem of relatively little value. FOXBORO DYNALOG BTU RECORDER
FOUR WIRE CABLE
HEAJ^ EXCHANGER
72 A
CHART
The B.t.u. measurement is con siderably less specific than process analysis instrumentation, but may improve process control where the more selective analytical control is impractical. The measurement is based on conventional and proved instrumentation, and reliable and trouble-free instrument operation may be expected. The accuracy of control obtained may be rather low, for there are many causes of error. Random heat transfer be tween the system measured and the environment is rather indeterminate and may be very significant, and the correlation of B.t.u. with the desired process information may not be completely satisfactory. However, even a rough indication can be far better than blind operation, and in favorable cases results mav be cood.
THREE WIRE BULB CABLES H 5 V - 6 0
