WHINERY MARK DOUGLAS
ABENS MARY BARBARA
| WO1999040360A1 | 1999-08-12 |
| US5047965A | 1991-09-10 | |||
| DE4205010A1 | 1993-08-26 | |||
| EP0087766A1 | 1983-09-07 |
| 1. | A gas distribution system distributing gas to users comprising: a pipeline for supplying gas to a plurality of users; a pressure regulator coupled to said pipeline upstream of said users; and an adaptive pressure controller monitoring the gas pressure at a lowpressure point downstream of said users and controlling said pressure regulator to enable the gas pressure downstream of said users at said low pressure point to be greater than a predetermined low pressure value. |
| 2. | A gas distribution system according to claim 1, wherein said pressure regulator includes a set point, and said adaptive pressure controller changes said pressure regulator setpoint. |
| 3. | A gas distribution system according to claim 2, wherein said adaptive pressure controller includes a lowpressure point controller having a controller setpoint input, said lowpressure point controller monitoring said gas pressure at said low pressure point and comparing said monitored lowpressure point gas pressure with said controller setpoint input to provide a lowpressure point controller output for changing said pressure regulator setpoint. |
| 4. | A gas distribution system according to claim 3, including a plurality of said pipelines, each including a respective pressure regulator coupled to said lowpressure point controller, wherein said lowpressure point controller selectively changes the pressure regulator setpoint of each respective pressure regulator in response to the monitored lowpressure point of the gas pressure downstream of said users in said plurality of pipelines. |
| 5. | A gas distribution system according to claim 1, wherein said adaptive pressure controller includes an adaptive predictive controller responsive to the time of day and temperature and providing an adaptive predictive controller output for correspondingly controlling said pressure regulator. |
| 6. | A gas distribution system according to claim 5, wherein said pressure regulator includes a set point, and said adaptive predictive controller changes said pressure regulator setpoint. |
| 7. | A gas distribution system according to claim 3, wherein said adaptive pressure controller includes an adaptive predictive controller responsive to the time of day and temperature and providing an adaptive predictive controller output for correspondingly changing said pressure regulator setpoint. |
| 8. | A gas distribution system according to claim 7, including combining means for combining said lowpressure point controller output and said adaptive predictive controller output and coupled to said pressure regulator for changing said pressure regulator setpoint. |
| 9. | A gas distribution system according to claim 4, including a respective adaptive predictive controller coupled between each of said pressure regulators and said lowpressure point controller, each respective adaptive predictive controller responsive to the time of day and temperature and providing a respective adaptive predictive controller output for correspondingly controlling a respective pressure regulator setpoint. |
| 10. | A gas distribution system according to claim 9, including respective combining means for combining said lowpressure point controller output and a respective adaptive predictive controller output and coupled to a respective pressure regulator for corresponding controlling said pressure regulator setpoint. |
This invention relates to apparatus and a method for controlling pressure in a gas distribution system, and in particular to the adaptive predictive control of such pressure in response to the low-pressure point of the system.
The adaptive pressure control system of the present invention was developed to control and manage pressure in natural gas distribution systems. As shown in Figure 1, the adaptive pressure control system consists of two main components or subsystems: 1) The Pressure Control System.
2) The Pressure Management System.
The Pressure Control System is located entirely in the field and is responsible for controlling pressure on a localized basis. The Pressure Management System consists of a dedicated personal computer and software that communicates with the Pressure Control System. It provides a pressure information database as well as an operator interface to the field system.
Figure 2 shows the District Regulator Controller (DRC), and Figure 3 shows the Low-Pressure Point Controller (LPPC). The LPPC is located at the low-
pressure point (LPP) in the gas distribution system and is responsible for maintaining the low-pressure point at a user-defined setpoint. The LPPC can communicate with up to five DRCs, as shown in Figure 4.
As depicted in Figure 5, in previous gas control systems, the district regulator maintains a constant pressure at the station outlet, and control is implemented independent of the pressure at the low- pressure point (LPP). The setpoint at the LPP is based on system requirements during peak load and, therefore, in order to maintain the LPP above a minimum safe pressure, the district regulator local setpoint (LSP) must be set much higher than is actually needed most of the time during a typical day. Some systems allow the operator to remotely change the district regulator setpoint, but even this type of control is not based on the pressure at the LPP.
In contrast with such previously known systems, Figure 6 shows a closed loop with an"ideal"pressure controller closing the loop. The setpoint at the LPP is achieved by constantly varying the setpoint of the district regulator, which is the opposite of what happens in previously known control systems.
Figure 7 shows how the adaptive pressure control system 20 of the present invention provides "ideal"control of the LPP, using closed-loop setpoint control. The adaptive pressure control system 20 includes an Adaptive Predictive Controller (APC) 22 and a
Low-Pressure Point Controller (LPPC) 24. The district regulator (DR) 26 still has a local mechanical setpoint (LSP) that is set to some minimum setting in case of electronic controller failure. The remote setpoint (RSP) is set from the Pressure Management System and is used during override and in setting initial conditions. The pressure setpoint (SP) is set at the LPPC to the desired minimum pressure. A deadband is also set by the operator around the setpoint. If the pressure wanders outside the deadband, then the adaptive predictive controller 22 algorithm drives the process back to the setpoint.
The purpose of the adaptive pressure control system 20 is to control the pressure at the LPP 28 to a user-defined setpoint. The LPPC calculates an average of the pressure over a minute and then checks the average against the user-defined setpoint. et = PVipp,-Spipp, (1) where: et = Error at time t, PVlppc = Process Variable for LPPC = Average Pressure, and SPlppc = Setpoint for LPPC.
If et is less than the error deadband, ed, then the pressure at the LPP is acceptable and nothing further will happen. However, if et is greater than ed, then the LPPC will calculate a change to the Inner Loop setpoint for each DRC. Because each DRC can have different gains
and affect the system differently, the change in Inner Loop setpoint calculated by the LPPC is specific to each DRC. To calculate the change in Inner Loop setpoint, the following calculation is used: ApgAiAAA where: IL KP = Proportional gain, F, = Scale Factor, t-1#t et = Error at time t, et1 = Error at time t-1, KI = Integral gain (repeats/minute), <BR> <BR> <BR> <BR> <BR> #<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> D<BR> <BR> <BR> <BR> <BR> <BR> <BR> t-1#t Rt = Rate of change of error at time t = (et-et1)/E Rt1 = Rate of change of error at time t-1.
After the LPPC has calculated the change in Inner Loop setpoint(AA that DRC.
If the system is in Adaptive mode (see Figure 8), then the DRC will change its Inner Loop setpoint every minute based upon the following calculation: SPIL = SPILold + A IL m m
where: SPIL = New Inner Loop setpoint, SPiLoid = Current Inner Loop setpoint, IL received from LPPC, TODm = Time-of-Day value for current minute of day, and A. = Adaptive value for current minute of day.
The Time-of-Day value for the current minute of the day (TODm) is calculated based upon the received change in Inner Loop setpoint(AA minute of the day. A Time-of-Day profile (see Figure 9) stores 1440 values that represent each minute of the day and how each minute of the day affects the LPP in the system. When the Time-of-Day profile is being built, the current minute of the day and the 15 minutes leading up to the current minute have their Time-of-Day values changed depending upon the mode of the system. To build the Time-of-Day profile, the following calculation is used if the system is in Adaptive mode: TODX = TODXO1d + (K*AAA where: TODX = New Time-of-Day value for current minute of day + previous 15 minutes, TODXOld = Existing Time-of-Day value for current minute of day + previous 15 minutes, KTOD = Time-of-Day profile gain (1% - 0. 01), and IL
received from LPP.
However, if the DRC is against the minimum or maximum Inner Loop setpoint or output while in Adaptive mode, the Time-of-Day profile is decayed to move the setpoint or output away from its limits. To decay the Time-of-Day profile, the following calculation is used: TODX = TOD, *DoD(5) where: TOD, = New Time-of-Day value for current minute of day + previous 15 minutes, TODxOld = Existing Time-of-Day value for current minute of day + previous 15 minutes, and DTOD = Time-of-Day profile decay (99% = 0.99).
The adaptive value for the current minute of the day (Am) is calculated based upon the change in temperature and the coefficient of temperature (see Figure 9). It will only be calculated if the Inner Loop setpoint and the temperature have changed over the last minute. To calculate the Adaptive value for the current minute of the day, the following calculation is used: k = CT * (TMPm-TMPmold) (6) where: A, = Adaptive value for current minute of day, CT = Current Coefficient of Temperature,
TMPm = Temperature at current minute of day, and TMPmold = Temperature from previous minute of day.
After calculating Am, the adaptive error is calculated to ensure that Am is not too large. The following calculation is used to discover the adaptive error: (SPIlm-SPILmold-Am(7)Ea= where: Ea = Adaptive error, SPILm = Inner Loop setpoint at current minute of day, SPxrLmoid = Inner Loop setpoint from previous minute of day, and An, = Adaptive value for current minute of day.
If Ea is too large or too small, then A is set to zero (0) and the coefficient of temperature is not recalculated. But, if Ea is within range, then Au is found to be acceptable and Ea is used to recalculate the coefficient of temperature if the minimum or maximum Inner Loop setpoint or output has not been reached. To recalculate the coefficient of temperature, the following calculation is used: CT = CTold + (Ka * Ea * (TMP-TMP)) (8) where:
CT = New Coefficient of Temperature, CTold = Current Coefficient of Temperature, Ka = Adaptive gain, Ea = Adaptive error, TMPm = Temperature at current minute of day, and TMP,, ld = Temperature from previous minute of day.
If the system is in Outer Loop mode (Figure 10), then the DRC will change its Inner Loop setpoint every minute based upon the following calculation: SPIL = #IL+ where: SPI, = New Inner Loop setpoint, SPILold = Current Inner Loop setpoint, and IL received from LPP.
Notice that the Time-of-Day value and the Adaptive value for the current minute of the day are not used to change the Inner Loop setpoint when the system is in Outer Loop mode. They are, however, calculated to allow for the Time-of-Day profile and the coefficient of temperature to be built and calculated respectively. The Adaptive value (A,), adaptive error (Ea), and the coefficient of temperature (CT) are calculated the same way in Outer Loop mode as in Adaptive mode. The Time-of- Day profile, however, is calculated differently. To build the Time-of-Day profile in Outer Loop mode, the
following calculation is used: TODX = (TODXO1d * (1-1/BP)) + (AAA where: TODx = New Time-of-Day value for current minute of day + previous 15 minutes, TODXOld = Current Time-of-Day value for current minute of day + previous 15 minutes, BP = Back Propagation into profile in minutes of day = 15, and IL From LPP.
After the DRC calculates the new Inner Loop setpoint (SPIL), it is checked against an Inner Loop setpoint minimum and maximum to ensure that the district regulator is not getting out of control. To control the District Regulator, the Inner Loop (Figure 11) of the DRC uses the following algorithm to create the PID control functionality: If et is greater than ed, then: Ot-1=KPKS(#It#D##Ot- If et is less than or equal to ed, then: Ot-1=0.0Ot-
where: Ot - Ot-1 = Change in output to be added to the actual output, Ot = Output at time t (0-o), Ot-1 = output at time t-1 (W), Kp = Proportional gain, KI = Integral gain (repeats/minute), KD = Derivative gain (minute), Rt = Rate of change of error at time t = (et-etl Rt1 = Rate of change of error at time t-1, et = Error at time t = PVDR-SPDRI et1 = Error at time t-1, ed = Error deadband, <BR> <BR> <BR> <BR> A<BR> <BR> <BR> <BR> <BR> <BR> s PVDR = Process Variable for DR = Outlet Pressure, and SPDR = Setpoint for DR.
The foregoing detailed description has been given fcr clearness of understanding only, and no unnecessary limitations should be understood therefrom, as modifications will be obvious to those skilled in the art.