Pressures Build On Gas Pipelines To Automate Systems

Randy Bomar
Advanced Control Engineering Services
Houston

Among the visible products of pipeline automation are custom-designed panels such as this, with the LCD screen shown at top.

This is a typical OIT station overview screen for an eight-engine, one-turbine compressor station. [12,408 bytes]
The electro-pneumatic control panel shown here is of the type installed about 1960. [14,768 bytes]
A typical station ethernet communication LAN (local area network) for four integral compressors and two gas turbines has modem and satellite access as well as remote monitoring capabilities. [9,591 bytes]

Automation is the driving force of efficiency for the next century in the natural gas pipeline industry. But what is automation?

Every company seems to have its own definition. To one company it may be as simple as speed control with torque monitoring. To others, it might mean specifying a 90 w, redundant 16-bit, 133 megahertz processor with 1 megabyte of RAM, EEPROM backup, floating-point math instruction, local and remote distributed high speed input/output (I/O), and a 100 megabaud fiber optic TCP/IP ethernet LAN with RS-423 serial link failover.

Regardless of the definition, automation offers competitive advantages that the industry can no longer afford to ignore.

Where it started

It is important to consider how pipelines operated in the past and to recognize the potential of the latest automation technology.

Following World War II, the first major U.S. gas transmission lines converted marine and power generation engines to run on, and compress, natural gas. This was a relatively new concept. Technology was young, manpower was abundant, and operating costs were not a great concern. As a result, efficiency was not necessarily a priority. Pipelines were their own largest customers.

Control and monitoring the operation of these engines were very simplistic. Operator control panels on most units consisted of a metal plate with a small selection of gauges to monitor operating pressures and temperatures. And enough workers were available to continually watch these gauges and thermometers. Starts and stops were completely manual, with large handles required to open a starting air valve while slowly adding fuel with another valve and attempting not to overspeed the engine.

Under normal operating conditions, it might take two or three individuals to start and load an engine. Fixing a mechanical problem could mean as many as three to five people to get an engine running, provided they could find and identify the actual problem.

These engines required only general protection and the three original safety shutdown devices (SSDs): low lube oil pressure, high jacket water temperature, and engine overspeed. General maintenance, overhauls, tune-ups, and oil changes were performed on a regular schedule, whether the unit needed them or not. Operation crews worked three shifts a day, 7 days a week. This was the standard operating procedure throughout the early 1950s.

Bigger engines

In the mid-1950s, increased demand for natural gas created the need for additional horsepower at compressor stations. This required more, and bigger, engines. Compressor stations and gas processing facilities that previously operated with 2-4 engines might now have 6-10 engines.

Of course, additional engines required more space. This meant lengthening existing buildings or constructing new ones. It quickly became evident that the time required for an operator to walk to the farthest engine, possibly a quarter-mile away, to operate and monitor a single engine, required too much time. It simply was not efficient.

This led to the development of pneumatics for remote starting, loading, and stopping. Large metal control panels-some up to 8.5 ft high, 6 ft wide, and 4 ft deep-were constructed to hold these pneumatic components and the several miles of copper tubing.

Relay logic added to the capabilities of pneumatics with the ability to sequence an engine through start, load, unload, and stop operations while monitoring a wide variety of SSD end devices with go/no go indicators. These indicators were tied in to the sequencing pneumatics of the control panel and would automatically stop an engine operating outside its normal limits.

The individual devices could more closely pinpoint problems. Sequence control could automatically take care of closing valves, post-lubing the engine, blowing down high pressure gas, and notifying station personnel. Reading and recording the values on the gauges for pressure, thermometers for temperature, and site gases for levels were still the responsibilities of the operator. However, this early relay logic was the first sign of automation.

The 1960s brought great strides in electrical components, and a rating for safe use in a hazardous area allowed engine manufacturers to incorporate more and more electro-pneumatic, electrical, or some of the new electronic end devices to the control systems. This was a critical development. Pneumatic relay logic gave way in most cases to electrical relay logic as wiring in the new engine control panels took up considerably less space than the spaghetti of copper tubing. But even with electrical functionality, an operator could still only monitor safety points on an engine or identify the general area of a mechanical problem.

Several analog signals, such as pressures and speeds, were showing up on system designs, but the information was for display only at the panel and could not be sent to remote locations. Regular maintenance schedules were still maintained, and overhauls were still decided based on total engine hours. The change from pneumatic to electrical was a step forward, but gathering and transmitting information and data were still somewhere in the future.

Circuit-board controllers

The 1970s saw the introduction of several hardwire-type circuit-board controllers that were very similar in operation to early electronic microprocessors-but they were extremely limited in functionality.

Changing the program meant rerouting the wiring on the board and resoldering components. Several manufacturers debuted new concepts in remote communications, but each company had its own idea of how it should work. Because there were no standards, these systems were not compatible with one another.

This continued into the early 1980s, when the natural gas industry embraced higher-end microprocessors. This time, the gas industry borrowed and modified successful systems from the petrochemical and automobile industries. But the processors were still big and still not rated for use in hazardous areas. In most cases they were remotely mounted in nonhazardous control rooms with extensive wiring and cabling running to marshaling cabinets at individual engines.

Manufacturers developed end devices capable of converting pressure, temperature, and frequencies into electrical signals that the new processors could read and monitor. This meant, finally, that the operating staff had the ability to gather data on process changes over a given period of time. But since processor reliability was not yet proven and knowledge of repair "how-to's" was very limited, the first processors were integrated to an existing engine control system.

This resulted in dual control systems for each engine. Most of the time, during the early years, the redundant pneumatic or electrical system was used for the operation of the engine with the microprocessor being used more for monitoring than control.

Several companies developed controllers to operate individual functions on an engine, while others developed other controllers that could manage all aspects of engine operation. Brand name controllers operated on proprietary software, which would not allow site-specific changes. Without communication drivers or data acquisition standards, the different systems could not communicate easily.

The present

The latest innovations in automation show extraordinary progress in four areas: new processors, distributed I/O, new interfaces, and "ethernet" communications. The combination of these factors has turned many heads in the industry by reducing operating costs while significantly improving pipeline efficiency.

New processors

Within the last decade, the industry has seen approval of the Class I, Division II High-Speed Processor.

The technology has become available for smaller, more-powerful, and more-functional microprocessors that still maintain classification for hazardous areas. These improved processors have significantly enhanced programmable logic controllers (PLCs).

PLCs gather analog and digital data using proven industrial I/O equipment. The features of the new processors allow utilization of this information for advanced control algorithms and calculations that can be communicated in real time to one or more locations for analysis and decision-making.

Distributed I/O

Advances in hazardous-area rated industrial I/O systems include increased selection, improved response time, and the ability to distribute the I/O modules throughout the compressor station.

This distribution of I/O modules greatly reduces construction and installation costs. Instead of pulling hundreds of wires back to a central location, the new systems simply require a single, twisted pair coming back to the main processor.

New interfaces

Flat liquid crystal display (LCD) screens with graphical user interfaces (GUIs) on the station control panel are eliminating buttons, knobs, switches, lights, meters, and gauges. Utilizing either touchscreens or membrane-keypad systems, these advanced screens can control the entire operation process with the touch of a button.

Ethernet communications

With ethernet communication capabilities, the technology that makes client/server network computing and the internet possible can be transferred to compressor station operations.

Similar to linking PCs in an office network, multiple compressor stations can be linked to a main control center and monitored remotely. Again, the result is continuous, real-time data gathering across the entire pipeline system that is accurate and reliable.

For example, from a centralized gas control location, operators are no longer limited to a phone call informing them that a station is off-line or that a particular engine at a location is out of service. Instead, operators can actually do all the trending and data-gathering from gas control and resolve the situation, without having to go on location.

The old way was to have personnel go to the engine, identify the problem at the station level, and make phone calls to corporate gas control officials and inform them of the problem.

As another example, a mechanical engineer may be doing design work that requires knowing the operating ranges of a specific engine at a particular station. With an advanced automation system, the engineer logs onto the computer and pulls up the last 6 months of trended data for the specified engine. The information is available instantly and without the possibility of human error resulting from handwritten logs.

What it means

Utilizing these technological advances, today's automation systems boost the flexibility, reliability, and competitiveness of pipeline operations.

Flexibility Advanced monitoring and communication capabilities offer unparalleled operational control, from a single engine at a specific compressor station to the complete pipeline system. This increased flexibility allows pipelines to operate in a variety of modes.

For example, from gas control, these systems allow for adjustments in pipeline conditions for maintaining discharge pressure based on deliverability, setting suction pressure based on transmission and storage requirements, and controlling flow rate based on operational needs.

At the single engine level, the latest systems offer:

Horsepower-torque calculation on real-time pressure-volume measurements.
Electronic air-fuel governor and ignition timing control.
Remote start-stop and load-unload sequencing.
Alarm monitoring and safety shutdown.
Fuel-flow rate calculation.
Advanced temperature and pressure trending.

Benefits also accrue at the station level, including:

Ability to control all engine and station functions from a single PC.
Remote monitoring of multiple locations.
Continuous data gathering and information trending of engines and controls.
Monitoring of fire and gas detection systems.
Control for emergency shutdown and plant and station blow-down systems.
Flow, pressure, and horsepower commands for throughput.

Reliability

Today's automation systems make the entire pipeline operation much more reliable than before. With this augmented reliability come increased safety, enhanced throughput and engine efficiency, and less downtime.

For example, automation control systems increase preventative maintenance efficiency by providing a more accurate picture of engine performance than ever before possible. With ethernet networking, it is now possible to monitor individual engine performance, as well as individual station performance, from a centralized gas control location and to spot potential problem areas before they occur. Instead of shutting down and overhauling engines based on run-times, the engines will be out of service only when they require maintenance. This means cost reductions and extended intervals between overhauls.

Competitiveness

Obviously, the gas industry operates within a highly competitive, spot-market environment. Small price differences have significant ramifications. In such a situation, it is a distinct advantage to keep operation costs as low as possible. This is the strength of advanced automation.

Gas companies are still their own biggest customers. Fuel gas constitutes a large percentage of what flows through the pipeline. Increased engine efficiency resulting from automation significantly lowers the amount of fuel gas consumption, thus saving the company substantial amounts of money.

In the end, the more efficient the engines and station control operations are, the more competitive they can be. And the cheaper the utilities get their supplies, the better they can hold prices for end users.

Advanced automation can pass cost savings through the entire pipeline system-all the way from gathering and processing to the end user. In a simplified model, if capital expenditures to maintain the system increase, utilities must raise rate bases to cover the additional costs. However, with the new automation systems, these increases can be kept to a minimum because of actual reductions in operating costs and gains in efficiency.

Need for efficiency

Traditionally, cost-justification has been the main argument against automation.

With the new technology available, this argument is no longer valid. When properly designed, today's automation systems can demonstrate payback within 18-24 months.

The need for efficiency is forcing natural gas pipeline companies to recognize the value of automation. It is no longer a question of automating or not. It is now a question of when-and how long companies can afford not to.

The Author

Randy Bomar, president and chief executive officer of Advanced Control Engineering Services, has more than 18 years of direct pipeline experience. ACES, a gas pipeline automation specialist, has completed projects for many major gas companies and has developed partnership alliances with Williams Natural Gas, Northwest Pipeline Corp., Tennessee Gas Pipeline, and Columbia Gulf Transmission.

Case study: automating a pipeline's engine control system

Randy Bomar
Advanced Control Engineering Services
Houston

IN THE EARLY 1980s, SEVERAL major gas pipeline companies moved into the first phase of electronic microprocessor automation and control by installing industrial computers used by the gas industry for the automatic operation and monitoring of reciprocating engines and turbines and the remote control of compressor stations.

One of these early microprocessors was a PDP-11 computer manufactured by Digital Equipment Corp. (DEC). The unit was designed for use in the business industry and was never intended for the type of service requirements of a gas compressor station.

Even though these units were 110-v ac-powered, nonrated for use in a hazardous area, and had limited and cumbersome software to work with, several gas companies adapted them for use at their facilities. Individual rooms had to be added near the operator control area at each location to house the massive DEC systems. Overhead cable trays or underground conduits were installed to route the miles of wiring from the DEC room to each engine marshaling cabinet located in the compressor building next to each engine.

Extensive conduit, explosion-proof enclosures, and intrinsically safe isolation barriers were required to meet the National Electric Code (NEC) requirements for a Class I, Division 2 area. Pressure and temperature monitoring and trending capabilities were limited, and gas control set-point changes and commands to remote locations were slow and uncertain due to communication errors.

After overcoming the challenges associated with these first microprocessor systems, several companies began installing the DEC systems during the early to mid-1980s. Some of the systems are still in operation today.

During the mid-1980s, several companies internally engineered new automation control systems for their locations using various early-model programmable logic controllers (PLCs) or individual proprietary system controllers. This worked well for minimal control, but system flexibility and expandability were limited and the lack of a standard communication protocol eliminated multiple-system integration. Most companies that were moving in the direction of electronic control were accomplishing this by developing their own automation systems using processors of the day, such as the Allen Bradley PLC5 series, the Bristol Babcock 3300 series, the GE Series 6 and 90 series, and the TI-Siemens 505 series.

In late 1995, DEC notified all users of its system components that the manufacturing of PDP-11 processors, input/output (I/O) cards, auxiliary boards, and replacement components would be discontinued over the next 12 months. As several systems were beginning to falter due to the limited availability of spare parts or problems caused by the increasing age of field devices, several companies faced the problem of how to replace their automation systems and at the same time upgrade to the latest technology available. This problem faces most gas pipeline companies worldwide.

Replacement systems

In an attempt to begin replacement of the obsolete DEC systems, companies began engineering replacement systems for their locations, very much like the companies that started the in-house automation design of the mid-1980s.

As larger companies continued to engineer and improve on their in-house automation systems, others turned to specialized engineering companies for assistance as they did not have the internal technical resources available.

In 1996, Advanced Control Engineering Services was approached by a major U.S. gas pipeline corporation to design a universal engine control system, develop the PLC software, and develop the human-machine interface (HMI) software required to operate the reciprocating engines throughout its system. Due to contractual agreements with this corporation, the company will be referred to in this case study as the customer.

In most cases, gas pipeline companies have either standardized on a manufacturer of microprocessors to be used at compressor station sites for cost or communication capabilities or are in the process of reviewing the latest technology available. In the case of the latter, the more specialized engineering company is asked to consult and evaluate current systems to assist in the final decision.

When the customer agrees on a manufacturer, a universal control system and software can be developed specifically for the reciprocating engines and station control throughout its pipeline. The customer takes the lead as the prime contractor, purchasing equipment and coordinating engineering support services for system design and software development along with site-specific planning for instrumentation and electrical (I&E) construction installation. In some cases, the engineering service company can take the project turn-key; in others, its involvement is more as a consultant, software programmer, and provider of technical support for engine, turbine, and station operations.

In this case, the customer began the project in the summer of 1996 by requesting a system be designed for each engine type being operated at the 18 reciprocating compressor stations throughout its system. This consisted of a total of 57 integral compressors, 11 different models, and 3 manufacturers-Cooper Bessemer, Dresser Clark, and Ingersoll-Rand. The control panel and all associated components would have UL certification for use in a hazardous area and be rated for Class I, Division 2 service. It should also have the capability of redundant communication capabilities with ethernet as the primary and a serial connection as the secondary.

In an attempt to standardize on a panel design that could be used system-wide, an electronic panel previously designed for the customer's sister company was reused with minor modifications. A second panel would be incorporated in the design to allow for pneumatic interface devices; it could be stand-alone, if space restrictions applied, or attached to the new electronic panel to appear as one unit.

The new control system would be the same for each engine type with fewer components per panel based on the engine-compressor make and model. If a unit control panel was assembled and installed with software on a Dresser Clark TLA-6 in one state, each end device that either controls or monitors the engine operation would be terminated, connected, and programmed to the same location as an Ingersoll-Rand 412-KVS five states away.

For example, the Engine Lube Oil Pressure Transmitter is labeled PT-1006; terminated in the panel at the field wiring termination block TB-6, terminals A, B, and G; wired internally to the Allen Bradley PLC5/40E Rack 1, Slot 1, Channel 6, Card Connection 13, and 14; and programmed in Rockwell Software (RS) Logix 5, integer register N16:9, and memory map location F8:9. This would be the standard on all 57 engines.

GUI specified

In an attempt to reduce I/O cards required to operate panel display devices and to eliminate the extensive metal fabrication required for equipment cutouts, a graphical user interface (GUI) was specified as a replacement to the many buttons, knobs, lights, switches, meters, and gauges previously used on most engine control panels. The GUI was to also have a Class I, Division 2 hazardous location rating, the same as the PLC, and be able to operate the same HMI software as the central station operator interface terminal (OIT).

In addition, ethernet communications would be required from each GUI so that all engines at a site could be monitored or operated from any given panel.

In order to meet these requirements the GUI was required to have an industrial PC with a Pentium processor running Windows NT with an ethernet communication card and a color liquid crystal display (LCD) touch screen. This would allow the station OIT and each GUI to run the same software and eliminate additional development time required to generate display screens for multiple formats. The same engine overview, monitoring and control screens could be used on both platforms.

The only way to design a control panel and develop both PLC and HMI software that would apply to every engine throughout the customer's system was to review the points to be monitored and to select the largest engine on the system and use it as the design model.

The engine used as the prototype model was a Dresser Clark TCV-16, with 16 power cylinders, 2 turbochargers, 8 compressors, 44 unloaders, 17 pocket steps, 1 jacket water temperature control valve, 1 auxiliary cooling water control valve, 1 cooling fan speed control valve, 2 suction valves, 2 bypass valves, 2 discharge valves, and 2 vent valves.

With other locations having several monitoring and control points that were not used on the TCV-16, one pony turbo control, a second and third cooling fan speed control valve, and a third discharge valve were added to the I/O list for the prototype universal system design.

Due to engine type differences such as two-cycle or four-cycle and turbocharged or mechanical blower, three different modules would be developed in the automatic and manual start, run, and stop sequences of the universal software to cover the operational differences of each engine type. These modules could be turned on or off depending on the type of engine operating at a specific location.

System prototype

Engineering design and software development began on the prototype of the universal system in August 1996 and were completed during the first quarter of 1997. In an effort to evaluate the operation of engine control software and test communications from gas control to each new engine and station operation, a development center was set up at the offices of ACES in Houston to simulate engine, turbine, and station control for overall pipeline operation. An engine, turbine, and station PLC, complete with simulated I/O, was temporarily set up and connected to a GUI with active TCV-16 display screens. A central station OIT was also included in the equipment so that the entire system could represent a single engine and turbine compressor station.

The system was connected to the customer's actual gas control via local phone company high speed data lines and the parent company's intranet communication system. This development and testing facility was used to check all monitoring and control software as well as to verify communication from gas control headquarters to specific compressor stations.

During the last several months of the prototype phase, work was begun on the first location to receive the new control system. The engine control panel assembly began in January 1997 with the construction phase starting in February. The construction installation was completed on engine No. 1, a Dresser Clark TLA-6, in late March, 1997, with start-up and commissioning in first week of April. This was a nonpressurized start as the station piping was blown down for construction safety.

The I&E construction was completed on all four engines by the end of April 1997, and the station piping was repressurized the first of May. Three Dresser Clark TLA-6 engines and one Dresser Clark TCV-10 engine were checked out, started, and commissioned during May. Commissioning of the station control software followed within the next month. This meant that from start to finish, the automation retrofit of four reciprocating integral compressors, engine and station auxiliary controls, and central station monitoring and operation was completed within 7 months. This would not have been possible at this location had the time not been spent up-front to design the universal system and completely check it out prior to installation.

To date, the customer has installed the new engine control systems on 29 reciprocating integrals at 7 compressor stations. At each site, additional "smart" station control and updated turbine control were included in the upgrade.

There are currently 4 locations, with 12 engines, under construction with 7 stations and 16 engines remaining.

The pipeline-wide engine and station automation upgrade is scheduled for completion in 2000.