MXPA00007666A - Adaptive system for predictive control of district pressure regulators - Google Patents

Abstract

A method and apparatus for controlling a fluid distribution regulator disposed in a fluid distribution system in which an intelligne controller (15) is operatively connected to at least one fluid district regulator (12) for controlling the district regulator. At least one intelligent low-pressure monitor (30) is connected to the distribution system at a distance from the fluid district regulator and monitors ambient temperature (31) and fluid pressure (34) in the distribution system. A computer is operatively connected to the intelligent controller and the intelligent monitor and provides status checks and manual overrides of the intelligent controller and the intelligent monitor. An adaptive algorithm proximate the intelligent controller generates a prediction of fluid demand an a corresponding district regulator fluid outlet pressure setting based upon feedback from the intelligent monitor.

Description

ADAPTABLE SYSTEM FOR THE PREDICTIVE CONTROL OF DISTRICT PRESSURE REGULATORS BACKGROUND OF THE INVENTION This invention relates to a system for controlling the outlet pressures of district regulators in a fluid distribution system proportional to the fluid demand. More particularly, this invention relates to an adaptive system that varies the outlet pressures of district regulators in natural gas distribution systems in proportion to the gas demand. The system predicts natural gas demand and raises or lowers outlet pressures in advance of demand to maintain a relatively constant pressure at points in the distribution system away from district regulators. The prediction is formulated automatically based on the temperature, time of day and operation observed in the past of the distribution system. The invention includes a distributed intelligence system that adaptably converges to an accurate prediction of natural gas loads. The prediction function is carried out by means of intelligent controllers located in the district regulators. Smart low pressure monitors are located at remote points of the gas distribution system. A Centralized computer is used to communicate with distributed controllers and to provide a manual disconnection, but is not necessary in routine operation. DESCRIPTION OF THE PREVIOUS TECHNIQUE Currently, outlet pressures for district regulators in fluid distribution systems such as natural gas distribution systems are manually set seasonally. This represents a very crude prediction of gas demand. The outlet pressure is set at the highest expected value necessary for the station. The operator selects the value based on historical data for the particular distribution system. Because the demand can fluctuate greatly during the season, the outlet pressure is greater than optimal most of the time. Several techniques have been employed to achieve a real-time correlation between fluid demand, such as natural gas, and output pressures from district regulators. The feedback approach uses pressure sensors at pressure points in the fluid distribution system in constant communication with controllers located in the district regulators. For example, the United States patent 3,878,376 presents a control system of pressure by solenoid valve operated by computer that includes a computer, which converts the instructions of the computer into electrical energy to operate solenoid valves which, with associated components, pressurize or depressurize a closed volume, and a pressure measuring system which provides feedback to the computer. Although the feedback approach provides control based on demand in real time, it requires the presence of a constant communication channel and, therefore, with the costs related to it. An alternative approach which avoids the need for a constant communications channel is the purely predictive approach which involves the placement of a preprogrammed load profile in a controller in the district controller. The load profile contains the gas demand versus the time of day and temperature. For example, U.S. Patent 4,200,911 discloses a method and apparatus for optimal water distribution in which actual water consumption at selected nodes of a network from which water is supplied to consumers is measured to detect a standard pattern of water demand in each selected area. Predicted patterns of demand for each of the nodes are determined by comparing the characteristics or attributes of pumps and valves installed in the pipe network that are then controlled based on predicted demand patterns. See also the United States patent 4, 562,552 and U.S. Patent 4,569,012, which present predictive approaches. See also U.S. Patent 5,047,965 which discloses a gas pressure regulating valve controlled by a microprocessor having a pilot valve controlled by a spring-guided diaphragm with which the adjustment is carried out automatically providing increases of pressure of the spring side of the diaphragm by means of an electrically adjustable regulator valve under the control of the microprocessor. The historical pressure data according to the day of the week, time of day and ambient temperature are stored in the microprocessor. The temperature sensor supplies the microprocessor with a signal indicative of the ambient temperature, and this is correlated with the stored historical data to determine the setting of the main throttle valve. The problem with these predictive approaches is the requirement that the profile be compiled by an experienced operator and that it be updated by varying the load on the system in the long term. Failure to periodically update the profile Periodic and often can significantly reduce the predictive approach and, of course, the requirement of a human operator to update the profile significantly increases the cost of this approach. SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a control system for adjusting the outlet pressure of a district regulator of a fluid distribution system to meet the real-time demand for a fluid distributed by the fluid distribution system which uses a predictive approach and automates the creation of a load profile. It is another object of this invention to provide a control system for adjusting the outlet pressure of a district regulator of a fluid distribution system which, in addition to automating the creation of a load profile, also provides a mechanism for updating the load profile. It is another object of this invention to provide a control system for adjusting the outlet pressure of a district regulator of a fluid distribution system to meet a real-time demand of a fluid distributed by the distribution system of fluid which eliminates the need for an open communication channel between the district regulator and the remote portions of the fluid distribution system. These and other objects of this invention are achieved by a fluid distribution system comprising a plurality of pipes, at least one district fluid regulator and control means for controlling the at least one district fluid regulator, one controller means intelligent is operatively connected to at least one district fluid regulator for controlling the district regulator, at least one intelligent monitor means for monitoring the ambient temperature and the fluid pressure in the distribution pipes operatively connected to at least one of the distribution pipes at a distance from the district fluid regulator, a computer means for providing state checks and over manual control of the intelligent controller means and the intelligent monitor means operatively connected to the intelligent controller means and the smart monitor means, and a selfadd algorithm suitable, said self-adaptive algorithm generates a prediction of the fluid demand and a fluid outlet pressure in the district regulator. According to a particularly preferred embodiment of this invention, the algorithm autoadaptable comprises a self-adapting finite pulse response filter (AFIR). More particularly, this invention provides a self-adaptive / predictive control system for output pressure adjustment of at least one district regulator of a fluid distribution system to meet a real-time demand for a fluid distributed by the distribution system of the fluid distribution system. fluid, at least one intelligent monitor means in said fluid distribution system at a distance from said district controller, a supervisory computing means for providing status verification and over manual control of the intelligent controller means and the monitor means intelligent, and a self-adaptive algorithm placed close to the means of intelligent controller, whereby the self-adaptive algorithm generates a prediction of fluid demand and an adjustment of output pressure of the corresponding district regulator. The intelligent controller means for controlling the district controller comprises controller communication means for communicating with the intelligent monitor means, pressure adjusting means of the controller to adjust the output pressure of the district controller, sensor input sensor means for detect a fluid inlet pressure from the district regulator, the room's ambient temperature sensor to detect an ambient temperature of a controller near the district regulator, and a clock with an hour of the day placed next to the district regulator. The at least one intelligent monitor means comprises a monitor communication means for communicating with the communication controller means, pressure sensor means monitor for detecting the fluid pressure in the fluid distribution system next to the intelligent monitor means and an ambient temperature sensor means for detecting an ambient temperature monitor close to the next monitor medium. This invitation also includes a method for controlling a fluid distribution regulator positioned in a fluid distribution system comprising the steps of collecting actual fluid pressure data and ambient temperature data from the fluid distribution system at a distance from the regulator of the fluid. fluid distribution, communicating the actual data of the fluid pressure and the ambient temperature data to an intelligent controller operatively connected to the fluid distribution regulator, processing the fluid pressure data and the ambient temperature data Actuals using a self-adaptive algorithm in the intelligent controller along with a local time of the controller. The ambient trature of the controller and the output pressure data of the fluid distribution regulator, resulting in the generation of a predicted fluid distribution regulator outlet pressure, and comparing the actual fluid pressure with the regulator outlet pressure of predicted fluid distribution. The processing of the fluid pressure data and the ambient trature data to generate the output pressure of the predicted fluid regulator, and comparing the outlet pressure of the predicted fluid distribution regulator with the actual fluid pressure data is iteratively repeated until a difference between the actual fluid pressure and the outlet pressure of the fluid distribution regulator is sufficiently small. At this point, the control of the outlet pressure of the fluid distribution regulator to the autoadaptable algorithm. At this point, communication between the district regulatory station and the fluid distribution system monitor is no longer required, and therefore, as long as the flow conditions within the distribution system remain within a predetermined bandwidth , the communication link or the channel can be deactivated. BRIEF DESCRIPTION OF THE DRAWINGS These and other objects and features of this invention will be better understood with the following detailed description taken in conjunction with the drawings in which: Figure 1 is a schematic diagram showing a portion of the fluid distribution system comprising a station district controller and a remote low pressure monitor according to one embodiment of this invention. DESCRIPTION OF THE PREFERRED MODALITIES Figure 1 is a schematic diagram of a portion of a natural gas distribution system comprising a district buffer station 11 and a remote low pressure monitor 30. Although this invention will be described in the context of a natural gas distribution system, it will be apparent to those skilled in the art that the concepts are equally applicable to other fluid distribution systems, including water distribution systems, and, accordingly, should not infer the intention to limit the invention to natural gas fluid distribution systems from this description. In order to carry out the adjustment of the pressures of output of the district controller 12 of the district control station 11, the district controller 12 is supplied with an intelligent controller 15. To carry out the adjustment function, means for converting an electrical signal into a suitable pneumatic signal for controlling the District regulator is supplied. This is typically carried out by turning a set screw of the pilot regulator 13 with an electric motor 14, although other means are available. In communication with the intelligent controller 15 are the inlet pressure sensor of the district regulator 16 which monitors the inlet pressure to the district regulator 12 to verify that there is sufficient driving pressure for proper operation and an outlet pressure sensor of the District regulator 17 which monitors the outlet pressure of district regulator 12 to verify that a desired set point is maintained correctly. In order to carry out these functions, it will be apparent that an electrical power source is also required. In summary, the means for adjusting the outlet pressure of the district regulator 12 to maintain a given set point comprises local feedback comprising the inlet pressure sensor of the district regulator 16, the outlet pressure sensor of the district regulator 17 , which are operatively connected to the intelligent controller 15, and to the motor 14 and the pilot regulator 13, placed between the intelligent controller 15 and the district controller 12 to adjust the opening and closing of the district regulator 12 according to the instructions of the intelligent controller 15. In addition to the means for adjusting the output pressure of the district controller for maintaining a given set point, a means is also required to generate an appropriate set point. The generation of the set point additionally requires a clock 20 with the time of day, an ambient temperature sensor 18 to monitor the temperature near the district regulatory station 11 and, therefore, the district controller 12 and a link of communication 19 next to the district regulator 12. Therefore, the time and temperature are parameters that are always available locally at the district regulatory station. To generate an appropriate set point, it is also necessary to obtain pressure and temperature data from low pressure points within the distribution system that are remote from the district controller 12, said data are only required during the adaptation or training period of the algorithm self-adaptable placed near the 11th district regulatory station.

The remote low pressure monitor 30 comprises an intelligent controller 33, a power source, a remote pressure sensor 34, a remote temperature sensor 31 and a communication link in communication with the communication link of the controller of the district controller . The pressure sensor 34 is connected to the gas distribution pipe 10 at a remote point to any district regulator. A pressure set point and tolerance are programmed into the remote low pressure monitor 30, the set point plus or minus the tolerance within a dead band within which it is desired to maintain the pressure of the fluid distribution system. The remote low pressure monitor 30 records the gas pressure in the fluid distribution pipe 10 and the ambient temperature near the remote pressure monitor 30. However, no communication is established through the communication link 32 to the communication link 19 of the controller 15 unless the pressure measured by the remote pressure sensor 34 exceeds the limits of the programmed deadband, at that time the remote low pressure monitor 13 transmits its temperature and pressure data to one or more District 15 smart controller drivers.

In this manner, the remote low-pressure monitor 30 provides the intelligent controller of the district controller 15 with a set point during the adaptation of the intelligent controller 15 according to the auto-adaptive algorithm. This arrangement provides an external cascade feedback loop with the local loop around the district controller 12. Upon receiving data from the remote pressure monitor 30, the controller of the district controller 15 adjusts the output pressure of the district controller 12 in order to return the pressure of the remote distribution system within the dead band programmed. This interaction between the remote pressure monitor 13 and the controller of the district controller 15 constitutes the feedback control. Concurrent with the feedback control, the adaptive portion of the algorithm processes the data received from the remote pressure monitor 30 together with the local time, room temperature and output pressure data at the district 11 buffer station to generate a pressure of the output of the district regulator predicted. This predicted district regulator outlet pressure is compared to the district regulator outlet pressure derived from the feedback control measured by the outlet pressure sensor of the district regulator 17.

The predicted output pressure is improved iteratively with each communication of the remote low pressure monitor 30. When the difference between the feedback pressure from the remote pressure monitor 30 and the predicted output pressure set by the self-adaptive algorithm is made small enough, the adaptation of the controller of the district regulator is complete and the predictive portion of the self-adaptive / predictive algorithm assumes control of the output pressure of the district regulator. At this point, the communication need initiated by the remote low-pressure monitor 13 via the communication link between the remote low-pressure monitor 30 and the controller of the district controller 15 is reduced or eliminated. The general mathematical form of this self-adaptive / predictive algorithm used by the method and apparatus of the present invention is that of a finite adaptive response (AFIR) impulse filter. According to said algorithm, a system parameter, in the case of this invention, a change in temperature, is sampled in a regular time interval. This flow of sampled values is fed to the filter algorithm. A finite number of samples is always retained, the reception of the most recent sample removes the oldest sample. This set, or vector of samples, is multiplied by a set of coefficients and the sum of these products is taken. This sum of products, or products, point of the sample and vector coefficients is the prediction or estimate of the change necessary to the output pressure of the district regulator. The output pressure of the predicted district regulator is subtracted from the feedback derived from the output pressure to form an error term. This error term is then used to refine the values of the coefficients. After a sufficient number of iterations, the predicted value converges with the feedback value, the error term approaches zero. Due to numerical rounding and the effects of digital quantization, the error term can not be zero. However, the predicted value may converge sufficiently close to the feedback value derived from the output pressure to be used instead. In practice, the finite adaptive impulse response for temperature can be constructed according to one of two methods. According to the first method, the output pressure of the district regulator as derived from the control feedback and the temperature are used as inputs to the AFIR filter. The Filter outputs are a prediction of the outlet pressure of the district regulator. After the outlet pressure of the precicha district regulator reasonably converges with the feedback derived from the outlet pressure, the AFIR filter obtains control of the outlet pressure. Any additional feedback from the remote low pressure monitor 30, if this occurs, is added as a differential correction term to the predicted value. Alternatively, the change in pressure and change in temperature are used as inputs to the AFIR filter of an element, only when both inputs are not zero. The filter output is a differential term similar to that required for feedback from the remote low pressure monitor 13 and is added to the setpoint of the inner loop of the actual district controller station. This new setpoint is equal to the previous set point plus the differential terms of the feedback and the AFIR filter. The time of day information is separated from the temperature by monitoring changes in district regulator outlet pressure and changes in temperature. If the pressure changes, but not the temperature, the change in pressure is sent to a self-regression filter. The day is divided into N equally spaced intervals and an element in a numerical arrangement is assigned to each interval. A fraction of the change in pressure is added to the element of the real arrangement and to each of the preceding n elements, where n is greater than one and less than N. In this way, repetitive pressure changes will be anticipated by the elements of the arrangement before the event. Repetitive patterns of pressure changes result in nonzero values that accumulate in the array. Non-periodic pressure changes tend to average zero, leaving periodic values without hiding. As the current element is each element of the array, the value contained in the element of the current array is added to the setpoint of the inner loop of the district controller station. In addition, the current array element will not react excessively to random inputs. Upper and lower limits are set to the values contained in the array. From this discussion, it can be appreciated that the self-adaptive / predictive system of this invention requires communication between the intelligence modes distributed during the adaptation phase. However, after the phase has reached convergence and enters the predictive phase, the volume of communication between the district regulatory station and the low-level monitor Remote pressure is reduced or eliminated. The pressure observed by the remote pressure monitors must remain within the programmed deadband and no communication will be initiated. This approach eliminates the expense of continuous feedback communication and the work of manually compiling the prediction profile load. Low-pressure monitors remain in place to provide means to update the load profile if long-term use changes sufficiently to require it. In addition, low pressure monitors provide immediate notification of dangerous pressures in the system for any reason. In practice, all district regulators and low-level monitors of the flow distribution system communicate with a central computer at least once a day. This low level of communication verifies the integrity of the system. It will be apparent to those skilled in the art that there may be one-to-one correspondence between low-pressure monitors and district controller controllers. Alternatively, a pressure monitor may communicate with a plurality of district controller controllers. According to yet another modality, a regulator controller of district receives data from a plurality of low-pressure monitors. While the above specification of this invention has been described in connection with certain preferred embodiments thereof, and many details have been established for the purpose of illustration, it will be apparent to those skilled in the art that the invention is amenable to additional embodiments and that certain details described herein may vary considerably without departing from the basic principles of the invention.

Claims (2)

1. CLAIMS. A self-adaptive / predictive control system for adjusting the outlet pressure of at least one district regulator of a fluid distribution system in order to satisfy a real-time demand of a fluid distributed by said fluid distribution system, which comprises: at least one intelligent controller means for controlling said at least one district controller comprising controller communication means for communicating with at least one intelligent monitor means placed in said fluid distribution system at a distance from said at least one a district regulator, controller pressure adjusting means for adjusting an output pressure of said at least one district regulator, controller input pressure sensing means for detecting the fluid inlet pressure of said at least one regulator of district, sensor means of pressure output from the controller to detect a pressure n fluid outlet of said at least one regulator district, ambient temperature sensing means controller for detecting ambient temperature controller next to said at least one regulator district, and time of day clock placed next to said at least one district regulator; said at least one intelligent monitor means comprising communication monitor means for communication with said controller communication means, monitor pressure sensing means for detecting a fluid pressure in said fluid distribution system proximate to said at least one intelligent monitor means, and ambient temperature sensor means of the monitor for detecting a monitor ambient temperature close to said at least one monitor medium; computer monitoring means for providing status checks and manual control of at least one intelligent controller means and said at least one intelligent monitor means operatively connected to said at least one intelligent controller means and said at least one monitor means intelligent; and a self-adapting algorithm that operates in communication with said at least one intelligent controller means, said self-adapting algorithm generates a prediction of fluid demand and a fluid outlet pressure setting of the corresponding district regulator. A control system according to the claim 1, wherein said fluid distribution system is for the distribution of natural gas. . A control system according to the claim 1, wherein said self-adapting algorithm comprises a self-adapting finite impulse response filter. . A control system according to the claim 1, wherein said at least one intelligent monitor means is a pressure monitor. . A control system according to claim 1 comprising a plurality of said intelligent controllers. A control system according to claim 5, wherein said at least one intelligent monitor means is in communication with a plurality of said intelligent controllers. A control system according to claim 1, wherein said at least one intelligent controller is in communication with a plurality of said intelligent monitors. In a fluid distribution system comprising a plurality of distribution pipes, at least one district regulator and control means for controlling said at least one fluid district regulator, the improvement comprises: an intelligent controller means connected operatively said at least one district fluid regulator to control said at least one district regulator; at least one intelligent monitor means for monitoring ambient temperature and fluid pressure in said distribution pipes operatively connected to said at least one of the distribution pipes at a distance from said at least one fluid regulator; a computer means for providing state checks or manual control of said intelligent controller means and said at least one intelligent controller means operatively connected to said intelligent controller means and said at least one intelligent monitor means; and a self-adapting algorithm operating in communication with said intelligent controller means said self-adapting algorithm generates a prediction of the fluid demand and a fluid outlet pressure setting of the corresponding district regulator. A fluid distribution system according to claim 8, wherein said intelligent controller means comprises controller communication means for communicating with said at least one intelligent monitor means, means for adjusting controller pressure to adjust an output pressure of said at least one district regulator, sensor input pressure sensor means to detect the fluid inlet pressure of said at least one district regulator, inlet pressure sensing means for detecting the fluid outlet pressure of said at least one district regulator, ambient temperature sensing means of the controller to detect the ambient temperature of the controller close to said at least one district regulator, and a clock of the time of day next to said at least one district regulator. A fluid distribution system according to claim 9, wherein said at least one intelligent monitor means comprises monitor communication means for communication with said controller communication means, monitor pressure sensing means for detecting a fluid pressure in said fluid distribution system proximate to said at least one intelligent monitor means and a monitor ambient temperature sensing means for detecting a monitor ambient temperature close to said at least one monitor means. 11. A fluid distribution system in accordance with Claim 8, wherein said self-adapting algorithm comprises a self-adapting finite impulse response filter.
2. 2. A fluid distribution system according to claim 8, wherein said at least one intelligent monitor means is a low pressure monitor. . A fluid distribution system according to claim 8, comprising a plurality of fluid distribution regulators and at least one intelligent controller for each of said fluid distribution regulators. . A fluid distribution system according to claim 13, wherein said at least one intelligent monitor means is in communication with a plurality of said intelligent controllers. . A fluid distribution system according to claim 8, wherein said at least one intelligent controller is in communication with a plurality of said intelligent monitors. . A method for controlling a fluid distribution regulator positioned in a fluid distribution system comprising the steps of: collecting actual fluid pressure data and ambient temperature data of said distribution system of fluid at a distance from said fluid distribution regulator; communicating said actual fluid pressure data and said ambient temperature data to an intelligent controller operatively connected to said fluid distribution regulator; processing said actual fluid pressure data and said ambient temperature data using a self-adaptive algorithm in operative communication with said intelligent controller together with the local time of the controller, ambient temperature of the controller and output pressure data of the fluid distribution regulator, resulting in the generation of an outlet pressure of the predicted fluid distribution regulator; comparing said outlet pressure of the predicted fluid distribution regulator with said actual fluid pressure data; iteratively repeating said processing of said actual fluid pressure data and said room temperature data to generate said outlet pressure of the predicted fluid distribution regulator and comparing said outlet pressure of the predicted fluid distribution regulator with said pressure data of the real fluid until the difference between said actual fluid pressure data and said predicted fluid distribution regulator outlet pressure is sufficiently small; and transferring the control of said outlet pressure from the fluid distribution regulator to said self-adapting algorithm. RESUM-N Ut, THE INVE.NCIÓN A method and apparatus for oldi a distribution device of Fug placed in a fluid distribution system in which a controlled! Intelligent (15) is connected in an operating manner with at least one fluid district regulator (12) to uoiitix the district regulator. At least one intelligent low pressure monitor (JU) is connected to the distribution system at a distance from the district regulator of f] \) 3 o and monitors the ambient temperature (31) and the fluid pressure (34) in the system of distribution. A comma ora is operatively connected to the intelligent controller and the intelligent monitor (30) provides status checks and manual overcontrols of the intelligent contxolacior and intelligent monitor. An adaptive algorithm close to the intelligent controller generates a fluid requirement prediction and a correspog district regulator fluid output correction adjustment with;; e in retro-integration from the intelligent monitor.