Technical Field
This invention relates generally to a method and apparatus for antisurge control of multistage machines such as refrigeration compressors. More particularly, the invention relates to a method that more accurately calculates the distance of the operating point to its surge line.
Background Art
Ethylene and propylene refrigeration compressors are commonly used in refineries for gas separation operations. Control and protection of these compressors is not only among the most complex of control applications, but a challenge fraught with inaccuracies. However, it is also an application where the economics of good control is large. Although energy savings can be profitable, an important objective is to maximize production while minimizing process upsets, surge, and shutdowns.
Most refrigeration systems are multistage machines with sidestreams entering and/or exiting between the stages. Since flow measurement devices are not available between stages, and the gas temperature entering most stages is unknown, it is impossible to calculate reduced flow reliably.
The method commonly used today requires the assumption be made that temperatures into the separate stages do not change considerably. When temperatures do change, the safety margin must be increased to protect the compressor under worst-case conditions.
Disclosure of the Invention
The purpose of the invention is to provide a method for more accurately calculating the distance of the operating point to its surge line for use in antisurge control.
For the most part, refrigeration systems are multistage units with sidestreams entering and/or exiting between the stages. In this type configuration, calculating a reliable reduced flow directly is impossible because a flow measurement device is not available at the inlet of all stages. It is also impossible to infer reduced flow from the available flow measurements since the gas temperature entering most stages is unknown. Therefore, a method used today assumes that the temperatures into the separate stage do not change significantly, but when they do change, the compressors must be protected under worst-case conditions by increasing the safety margin.
The patent incorporates a method of antisurge control, whereby reduced flow squared (q.sup.2) alone, is replaced by reduced flow times equivalent speed squared (q.sup.2 N.sub.e.sup.2), which allows accurate calculation of the operating point to the surge line; and inlet temperatures into the separate stages are not necessary.
Brief Description of the Drawings
FIG. 1 shows a representative multistage compression system.
FIG. 2 shows a compressor map comprising surge lines.
FIG. 3 shows a block diagram of a control scheme for a multistage compression system, inputting to a PID controller.
Best Mode for Carrying Out the Invention
Although the disclosure for the best mode refers specifically to refrigeration compressors, it should be understood that the invention has broader application to multistage compressors in other processes.
To minimize process upsets, surge, and shutdowns of a refrigeration system, precise and fast control application is paramount. Most such systems comprise multistage machines, wherein flow measurements are not available between stages, and the gas temperature entering most stages is unknown. It is, therefore, difficult to calculate an accurate reduced flow with which to calculate the distance to surge, and to effectively deal with compressor transients.
FIG. 1 depicts a representative multistage compression system 101 consisting of a low-pressure stage (LP) 102, high-pressure stage (HP) 103, a sidestream for flow admission or extraction (B) 104, and designated locations (1, 2, 3, and 4). Measurement devices are included for rotational speed (N) and for flow (.DELTA.p.sub.o), pressure (p), and temperature (T) at B and location 1.
The proposed method suggests that the reduced flow (q.sup.2 =.DELTA.p.sub.o /p) alone be replaced by the squared product of reduced flow and equivalent speed (q.sup.2 N.sub.e.sup.2). It employs a phantom differential pressure signal in the suction of each stage (subscripts denote locations shown in FIG. 1 ), which would be equal to ##EQU1## where: .DELTA.p.sub.o =pressure drop across flow measurement device (orifice plate or venturi)
A=orifice or venturi constant (converts m.sup.2 /.rho. to a pressure difference)
1=suction of first stage (subscript)
3=suction of second stage (subscript)
m=mass flow rate
.rho.=density
B=sidestream (subscript)
Mass flow rates can also be represented as functions of differential pressure measurements: ##EQU2## A present-day approach is to calculate .DELTA.p.sub.o,3 as ##EQU3## where: C.sub.1 =constant
C.sub.B =constant
f()=converts differential pressure signal from the suction of the low-pressure stage to the discharge of that stage
R.sub.c =pressure ratio
LP=low-pressure stage (subscript)
Similarity between Eqs. (2) and (3) can be seen when expanding out the densities in Eq. (2) and noting that p.sub.B =p.sub.3 : ##EQU4## where: Z=compressibility
R=gas constant
T=temperature
If compressibilities do not change considerably, the quotient ##EQU5## where:
and, assuming that the ratios ##EQU6## are approximately constant, then ##EQU7## where:
which is seen to be Eq. (3) with f.sub.1 (R.sub.cLP)=R.sub.cLP.sup..sigma.-1, C.sub.1 =.sqroot..LAMBDA..sub.1 , and C.sub.B =.sqroot..LAMBDA..sub.B where it has been assumed that p.sub.2 =p.sub.3.
To accurately calculate distance between the operating point of a compressor and its surge line, it is prudent to use parameters for which that distance is invariant to changes in suction conditions like temperature. Reduced flow (q) is such a parameter: ##EQU8##
Pressure ratio (R.sub.c =p.sub.d /p.sub.s) is another such parameter in which p.sub.d is absolute pressure at discharge and p.sub.s is absolute pressure in suction. Still another is equivalent speed (N.sub.e) with N=rotational speed, and defined as ##EQU9##
Any combination--linear or nonlinear--of invariant coordinates is also invariant and can be used for control. To eliminate the dependency on temperature values not available, Eqs. (4) and (5) can be combined as follows: ##EQU10## in which all variables are measured or can be computed easily for these multistage applications. Mass flow, in particular, must be calculated using flow measurements available and the temperatures and pressures, at the locations of the flow measurement devices. In other words, working from one end of the compressor (e.g., inlet end), mass flows can be calculated entering and exiting the compressor case, so simple subtraction or addition is all that is required to determine mass flow passing through a given stage. It should be noted that the same is true if the flow measurement signal is not available at location 1 in FIG. 1, but is, instead, available at the discharge of the high-pressure stage 103 (location 4). In this case, the mass flow rate through the low-pressure stage 102 can be calculated as the difference between the flow exiting the high-pressure stage and flow entering at the sidestream (B) 104.
It is ideal to control each stage using two invariant coordinates-one can be used as the independent variable in a function to characterize the other on the surge limit line, to construct an antisurge parameter: ##EQU11## Because R.sub.c is known for each stage of compression, it is a sound choice for y. Using reduced flow (.DELTA.p.sub.o /p) alone, for the high-pressure stages, does not appear to be a good choice because of the lack of flow and temperature measurements between stages. However, the product introduced in Eq. (6) is a good choice and can be represented as ##EQU12## Equation 8 reveals a parameter which can be used to define the distance to surge. This distance will be invariant to suction conditions. In other words, q.sub.3.sup.2 N.sub.e.sup.2 can be used in place of x in Eq. (7); moreover, all variables appearing in this equation are known. In terms of the pressures and temperatures that are known, ##EQU13## where it is assumed that the compressibilities (Z) are fairly constant. For control purposes, the square root of Eq. (10) could be taken to simplify the implementation. The antisurge parameter [Eq. (7)] is then, ##EQU14## where N is the rotational speed; m is mass flow rate through the stage of interest (calculated using known mass flows in the compression system); and f.sub.1 (R.sub.c) returns the value of q.sub.3.sup.2 N.sub.e.sup.2 at surge, so S.sub.s goes to unity on the surge limit line-the square root of q.sub.3.sup.2 N.sub.e.sup.2 is also valid for defining the distance to surge, i.e., S.sub.s =f.sub.2 (R.sub.c)/q.sub.3 N.sub.e. The implication or Eq. (11) is that it can be used for any stage in the refrigeration system which lacks either a flow measurement device or a temperature transmitter.
The control algorithm which employs the antisurge parameter in Eq. (11) uses the fact that S.sub.s will be equal to one on the surge limit line, and it is less than unity to the safe (right) side of the surge limit line. A safety margin (b) is added to S.sub.s to construct the surge control line (S=S.sub.s +b) as depicted in FIG. 2. The definition for the distance between the operating point and the surge control line can then be described simply as
This distance, or error, describes a parameter that is positive in the safe region (to the right of the surge control line), and zero on the surge control line. It can then be passed to a PID control algorithm which calculates the required output for the recycle valve associated with that particular compressor stage. In particular, this output would be calculated as follows: ##EQU15## where: G=proportional gain
K.sub.r =integral reset rate
T.sub.d =derivative time constant
A block diagram of the .delta. calculation [Eq. (12)] is shown in FIG. 3, wherein output information from a low-pressure compressor stage and from a sidestream is computed to define .delta.. In the figure, transmitter data (T.sub.1, p.sub.1, and .DELTA.p.sub.o,1) 20 are acted on by algebraic operations to produce a value proportional to mass flow ##EQU16## 21. Concurrently, data from the sidestream transmitters (T.sub.B, p.sub.B, and .DELTA.p.sub.o,B) 22 produce a similar value ##EQU17## 23.
A summing module 24 computes the two prior values 21, 23 as a mass flow rate (m.sub.3) which, in turn, is multiplied by rotational speed (N) generating a product (m.sub.3 N) 25, and is then divided by absolute pressure (p.sub.B) to specify a parameter (q N.sub.e) 26. Following this, pressure ratio (R.sub.c) 27 is characterized [f.sub.1 (R.sub.c)] 28 and divided by q N.sub.e to generate an antisurge parameter (S.sub.s) 29, from which summing operations determine the distance equation value (.delta.) 30 for input to a PID controller 31.
It should be noted that the present invention can be applied to compressors of any number of stages as long as the required instrumentation is available. Also, any number of stages can be lumped together as a controlled compression unit and treated the same way as described above for individual stages.
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.