Technical Field
This invention relates generally to a control method and apparatus for controlling the rotational speed of a hot gas expander, using a model based approach. More particularly, it relates to a method which anticipates the speed increase attributed to a sudden rejection of load.
Background Art
The operation of recovering hot gases for use by expanders not only adds supplementary drivers to the process, it also results in cost savings. Present-day high cost of energy mandates that "waste" energy (often discarded) be salvaged and used to drive electric generators, compressors, and pumps.
Expander speed is inherently variable, and even a "constant speed" application such as electric power generation must be brought up to speed and shutdown again. Therefore, control of the rotational speed of an expander is important during at least some of its operation.
When at least one of the loads driven by an expander is a synchronous or induction generator, there is the possibility of a very sudden rejection of load. A similar scenario is possible during the surge of a process compressor driven by an expander. Under these load shedding conditions, it is important to maintain a fairly constant speed to avoid overspeed, or for the sake of the process (in the case of one of the loads being a compressor, a generator, or a pump). For the above reasons, there is a need to easily and accurately control the rotational speed of expanders during certain critical control conditions.
Disclosure of the Invention
The purpose of this invention is to provide a control method for hot gas expanders that anticipates the shedding of a known power. This method employs the use of dimensional analysis applied to the characteristic maps that describe the expander (reduced power versus reduced flow rate or pressure ratio) together with a feedforward control action.
When the load (or a portion of it) on an expander is shed suddenly--following a breaker trip or compressor surge--the event may be sensed before the ensuing speed increase is detected by the speed control loop. In this situation, power applied by the expander to the shaft must be reduced. A feedforward approach can be employed to anticipate this requirement for control action before the feedback system indicates the need.
Open-loop control can be applied, but an intelligent application of this type of control should be made regarding how much to reduce the power. The intent is to balance the power applied to the shaft with the power absorbed on the shaft, to maintain a constant speed, as described in the equation ##EQU1## where:
I=moment of inertia
N=rotational speed
P.sup.+ =power added by the driver(s) on the shaft
P.sup.- =power absorbed by the load(s) on the shaft
It can be seen from the equation that speed will not change with time if the power applied to the shaft is exactly absorbed by the load(s). However, this equation also describes the rotational dynamics when these powers are not balanced. As mentioned, the intent is to balance the power applied to the shaft with the power absorbed--making the right-hand side of the equation close to zero, thereby keeping the speed under tight control.
Following through with the open-loop approach, this invention warns the speed control function of load loss and of impending speed increase. Next, it estimates the magnitude of the load lost. With a generator, a measure of the actual power produced the instant before breaker opening will suffice. For a compressor in surge, a computation of the polytropic power divided by the appropriate efficiency is sufficient.
The feedforward control action comprises three scenarios:
(1) Expander with Flow Measurement--Initially, reduced flow is calculated and reduced power is subsequently determined (using characteristic maps). From this power value, shaft power is calculated knowing the inlet conditions of pressure and temperature. The value of power rejected is then subtracted from the shaft power (the power being absorbed on the shaft). Next, the value of reduced power corresponding to this required power, is calculated. The value of reduced flow is now determined, representing the reduced flow rate needed to keep the expander and its loads from increasing speed. The required flow rate can be correlated to a valve position using the valve's characteristics.
(2) Expander with Downstream Pressure Measurement--The value of pressure ratio across the expander is required, and reduced power is then determined. The value of power can be calculated as in the preceding scenario, and the power which was shed subtracted from it. The process continues as above, resulting in the reduced flow rate required to maintain speed, and a valve position set point for the open-loop step.
(3) Expander Valve as a Function of Load Power--A mapping between the expander inlet valve position and the power of the driven load is constructed. When load rejection (partial or full) occurs, the required valve position is determined based on this mapping.
Once these three feedforward scenarios are initiated, closed-loop control will be applied to overcome any remaining errors.
Brief Description of the Drawings
FIG. 1 shows a power recovery system comprising a hot gas expander and instrumentation (measuring devices).
FIG. 2 shows a typical characteristic map for a hot gas expander depicting shaft power versus mass flow rate.
FIG. 3 shows a hot gas expander map (reduced power versus reduced flow rate squared) obtained through dimensional analysis.
FIG. 4 shows a hot gas expander map (reduced power versus pressure ratio) obtained through dimensional analysis.
FIG. 5 shows a flow diagram for calculating a set point to adjust a power control using R.sub.c.
FIG. 6 shows a flow diagram for calculating a set point to adjust a power control element, using q.sub.r.
FIG. 7 shows a flow diagram for calculating a set point to adjust a power control element, using R.sub.t.
FIG. 8 shows a flow diagram for calculating a set point to adjust a power control element, using P and P.sub.shed.
FIG. 9 shows a flow diagram for calculating a set point to adjust a power control element, using R.sub.c and P.sub.shed.
FIG. 10 shows a flow diagram for calculating a set point to adjust a power control element, using q.sub.r and P.sub.shed.
FIG. 11 shows a flow diagram for calculating a set point to adjust a power control element, using R.sub.t and P.sub.shed.
FIG. 12 shows a flow diagram for adjusting a power control element, using a predetermined set point value, C.sub.sp.
FIG. 13 shows a flow diagram for decrementing a power control element set point, by a value based on P.sub.shed.
FIG. 14 shows a flow diagram for calculating a set point to adjust a power control element, using .DELTA.T and m.
FIG. 15 shows a flow diagram for calculating a set point to adjust a power control element, using h.sub.1 and h.sub.2 as functions of T.sub.1 and T.sub.2, respectively.
FIG. 16 shows a flow diagram for calculating a set point to adjust a power control element, using h.sub.1 and h.sub.2 as functions of T.sub.1,p.sub.1 and T.sub.2,p.sub.2, respectively.
FIG. 17 shows an equation for calculating mass flow rate as a function of .DELTA.p.sub.o, T, and p.
FIG. 18 shows an equation for calculating a value proportional to mass flow rate.
FIG. 19 shows a flow diagram for the calculation of mass flow rate, using R.sub.c and q.sub.r.
Best Mode for Carrying Out the Invention
To prevent overspeed of a hot gas expander and its resultant load (e.g., electric generator, compressor, or pump), the expander's rotational speed must be accurately controlled. Overspeed occurs when the expander's load (or a portion of it) is shed suddenly, resulting from a breaker trip or compressor surge. The balanced relationship of the power added (P.sup.+) by the driver(s) on the shaft, and the power absorbed (P.sup.-) by the load(s) on the shaft, is shown as ##EQU2##
Since P.sup.- drops off upon load rejection, the only way to avoid increasing the rotational speed (N) is to decrease P.sup.+ to keep the right-hand side of the equation near zero; thus, maintaining a fairly constant speed for the sake of avoiding overspeed, or for the sake of the process should the load be a compressor, a generator, or a pump. Normal feedback to the speed controller may be too slow to be effective; however, this load shedding event may be sensed by a feedforward system before a significant speed increase is detected by the speed control loop. This invention uses such a feedforward approach to anticipate and avoid a speed increase during load rejection conditions.
FIG. 1 depicts a power recovery system consisting of a hot gas expander 101, a load 102, a control valve 103, an orifice plate 104, and instrumentation which includes a flow rate transmitter 105, an inlet pressure transmitter 106, an inlet temperature transmitter 107, an outlet temperature transmitter 108, and an outlet pressure transmitter 109. Each of the signals (104) through (109) is received into processor (110), which calculates an estimate of the expander power.
FIG. 2 shows a typical characteristic map for a hot gas expander, depicting shaft power versus mass flow rate and consisting of four parameters: mass flow rate, power, inlet pressure, and inlet temperature. These expander characteristics are typically a weak function of rotational speed. When the influence of rotational speed is significant, a third dimension can be added, being Equivalent Speed: ##EQU3##
By using dimensionless parameters obtained through dimensional analysis (similitude), the data will collapse into a single curve (shown in FIGS. 3 and 4), and the useful parameters are ##EQU4## where:
m=mass flow rate
z=compressibility
R=gas constant (R.sub.u /MW)
R.sub.u =universal gas constant
MW=molecular weight
T.sub.1 =inlet temperature
p.sub.1 =inlet pressure
p.sub.2 =outlet pressure
K=constant
.DELTA.p.sub.o,1 =inlet flow-measurement signal
P=power
N=rotational speed
Reduced Flow, Pressure Ratio, Reduced Power, and Equivalent Speed are nearly invariant to changes in inlet conditions. Note that these parameters are not truly dimensionless; however, since the geometry of the expander is fixed, dimensions of length can be eliminated by a constant characteristic length. Constants can be dropped and ignored for the purpose of this invention.
Because pressure is also changing in FIG. 2, a map (FIG. 4) can be constructed showing the relationship between the pressure ratio across the expander and the reduced power associated with it.
The implementation of this invention enlists two approaches: expander with flow measurement, and expander with downstream pressure measurement. With a flow measurement, the approach is to calculate the value of reduced flow, q.sub.r, then use FIG. 3 to determine reduced power, P.sub.r. From this value, shaft power is calculated knowing the inlet conditions, p.sub.1 and T.sub.1. The value of power rejected is then subtracted from this power which is being absorbed on the shaft; thus, the power to be provided. The value of the reduced power, corresponding to this required power, is calculated; and the value of reduced flow at which such reduced power is realized, is extracted from FIG. 3. This is the reduced flow rate needed to keep the expander and its load(s) from increasing speed. The required flow rate can be correlated to a valve position using the valve's characteristics.
The mode of an expander with a downstream pressure measurement, requires two maps. The value of pressure ratio across the expander is required, and FIG. 4 can be used to determine the reduced power at this point. The value of power can be calculated and the power which was shed subtracted from it. Switching back to FIG. 3, the process continues as above, resulting in the reduced flow rate required to maintain speed, and a valve position set point for the feedforward response,
Using the valve manufacturers equations (or the ISA standard equation), it is seen that reduced flow is all that is needed to correlate the flow with C.sub.v (or valve position). The ISA equation is shown below in terms of Reduced Flow ##EQU5## where:
.DELTA.p.sub.v =valve flow-measurement signal
p.sub.v,1 =valve inlet pressure
K.sub.1 =constant
K.sub.2 =constant
C.sub.v =valve flow coefficient
Therefore, the required mass flow rate need not be known, and the new value of C.sub.v can be calculated as a function of the reduced flows (assuming that .DELTA.p.sub.v /p.sub.v,1 is constant): ##EQU6##
As an alternative to measuring the inlet temperature, T.sub.1, and the discharge pressure, p.sub.2, determining one or both of these can be accomplished by estimating them. Examples of when they can be estimated instead of measured are:
when the inlet temperature does not change greatly during operation; and
when the expander discharges to atmospheric (or nearly atmospheric) pressure, so p.sub.2 is approximately constant.
FIGS. 5-16 and 19 exhibit flow diagrams for calculating a set point to adjust a power control element, whereas FIGS. 17 and 18 show equations for calculating mass flow rate.
Obviously many modifications and variations of the present invention are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described.