Digital Weighing System

U.S. Steel Goes Digital

Reprinted with permission of:
Weighing & Measurement, "U.S. Steel Goes Digital", August, 1999 Issue

The recent installation of a fully digital weighing system at U.S. Steel facility in Fairfield, AL. underscores the benefits of digital technology in dramatic fashion.

Digital technology has been a latecomer to the weighing industry. The recent installation of a fully digital weighing system at U/S. Steel facility in Fairfield, AL underscores the benefits of this in dramatic fashion. In addition to ease of installation, calibration and diagnostics of the new “plug and play” controller, the scales provide resolution and repeatability that is absolutely unachievable with traditional weighing technology. MTI Weigh Systems of N. Kingston, RI, designed and implemented this charter system using Revere’s new fully digital load cells in concert with MTI’s new Millennium D1000 Weight Controller. The results clearly suggest that digital technology has arrived!

In the early 1980s, U.S. Steel (USS) commissioned a new finishing pipe mill in Fairfield, AL. This mill was designed to manufacture high quality seamless pipes in a variety of lengths and diameters. Manufacturing advancements and improvements are continuously adopted at this facility, making USS one of the world’s leading suppliers of seamless pipe, which primarily supplies the petrol-chemical industry.

Return to top of page

Weight used as billing criteria

With annual sales of over $150 million, an accurate and reliable system of measurement was needed to account for product shipped from the facility. Product weight was selected as the billing criteria for this purpose. Each individual pipe (coming off of two separate manufacturing lines) is weighed and this data is used to calculate customer invoice amounts. Consequently, the facility weigh systems have been "mission critical" since the very beginning.

Originally, low-profile compression type mounts were used with 2,500-pound capacity stainless steel, potted, shear beam load cells. Six weigh modules were symmetrically arranged and evenly spaced across two combined sections of 20-foot, 12- by 12- inch box beams. This arrangement enabled USS to effectively weigh pipes up to 50 feet long and ranging in weight from 650 to 2,500 pounds, with an accuracy of ± 0.5 percent. The weight information was transmitted from a digital weighmeter via BCD output through GE Genius Blocks to a main frame which processed the transaction for billing purposes.

This original scale configuration was employed until May 1999. To understand U.S. Steel’s motivation to upgrade, the historical performance of the original scale system was evaluated in the context of the primary role it played in their manufacturing facility.

MTI Weigh Systems, Inc. reviewed internal maintenance records for the past several years of operation. Attention was drawn to the large deviations on the graphs related to downtime. USS officials were quick to point out that numerous 8- to 46-hour downtimes were directly attributable to scale failures. Additionally, the data suggested that the scales proved to have the highest failure rate of all the facility’s equipment directly involved in the manufacturing process.

Each time one of the scales went out of service, a 50 percent in the plant’s capacity resulted. This downtime had an accompanying labor cost of $4,000 per hour, exclusive of list profits. With average sales of $500,000 per day at full capacity, one quickly6 realizes the magnitude of these system failures in the plant.

Over the years, and after countless calls, USS developed an accurate understanding of their scale system and its inherent problems. First, the load cells had two modes of failure. Typically, failures were caused by overloading or environmental factors such as introduction of hydraulic fluid, chemical solvents or steam cleaning. The primary concern was the time and effort involved in replacing a failed load cell. The majority of the time and effort in repairing the scales was spent in rebalancing/ distributing the load equally among six load points. To re-balance the scale, shim stock was used to adjust the elevation of each load point. Elevating a given load point in this fashion would effectively increase the load “seen” by this point. This method proved to be extremely difficult and time consuming.

Locked into a service contract, USS also had to contend with inconsistent response times from their service provider. In many cases, different scale technicians with various levels of skill would work on the scales adding countless hours to the repair process. Upon expiration of the service contract, USS quickly decided to bring the repair and maintenance of the two scales "in-house".

Return to top of page

Scale system needed to be improved

Subsequently, USS solicited suggestions and recommendations from a number of participants in the weighing industry. The goal was simple. The durability and serviceability of the scale system needed to be improved. All proposals were constrained by the geometric requirement of maintaining elevation of 5 inches between the main tube / weigh beam mounting flange and the base. This limited the selection of standard compression type mounts with mechanical adjustability.

MTI Weigh Systems, Inc.’s proposal first focused on the mechanical design of the load cell mount. Our design offered several advantages over that of compression style mounts, which included:

• easier replacement of load cells;
• complete isolation from side loading;
• entire weighbeam was placed in suspension (an ideal loading configuration); and
• quick elevation adjustability for re-balancing the scale.

MTI Weigh Systems, Inc.’s mount design incorporated Revere Transducers Model SSB-5K load cell. A recently developed 2000 Kg digital version of the SSB Transducers called the SBC was available in Europe. Douglas Postek, Vice President at MTI, commented, "We were very excited about the possibility of using the SBC (digital load cell) because we had spent a year on a new digital controller developed specifically for use with Revere Transducers’ digital load cells." However, due to timing of the first round of proposals the SBC was still pending release in the United States and was not yet available.

USS informed MTI Weigh Systems, Inc. that they had gained an advantage over the competition due to their creative approach to the mounting configuration and the clear advantages this offered. MTI then inquired about how US Steel felt about digital load cell solutions. Their response was negative. Digital systems were perceived as difficult to set up and complicated to service and maintain. Simplicity was the key and proprietary digital systems were not favored.

"Paying close attention to the feedback we were getting on the digital issue, we knew that our Millennium D1000 weigh controller addressed all of the U.S. Steel’s concerns," said Postek. "It was now only a matter of getting Revere’s digital load cells delivered in the United States." Word finally came that Revere could release and deliver the required quantities of the SBCs within a six-week time frame. USS was now poised to implement a fully digital weighing system.

Return to top of page

Ready for a digital weighing system

Ultimately, USS was receptive to the advantages offered by a completely user friendly, plug and play, self-diagnostic digital weighing system that required no special training. Adding to the utility of the system, MTI offered a firmware that would pinpoint any load cell failure in the system while keeping the scale fully operational without substantial loss in performance.

Postek said, "MTI’s position on digital weighing solutions address the limitations of proprietary systems. Our goal is to offer complete systems that reflect true ‘open architecture’ design. Our controller is based on PC/104 standard industrial microcomputer and field bus. The D1000 weight controller has the fastest processor speeds, unlimited expandability and can easily communicate with any PLC or computer in any protocol. When U.S. Steel saw that our Millennium controller was designed with open architecture in mind, their systems engineers gained an additional level of comfort with our equipment."

Upon receipt of USS’s purchase order for two custom designed, fully digital weighing systems, MTI expedited their efforts in order to comply with a strict six-week delivery schedule. Engineering drawings were developed within two weeks for the mounting hardware and the machine shop manufactured the mounts within a week. On May 17, 1999, MTI field engineers arrived at USS Fairfield to begin the installation process.

With the expert aid of the Finishing Maintenance and the Electronic Departments at U.S. Steel and several days of preparation prior to MTI’s arrival, the first system was installed and operational in three days. The second scale system followed shortly after.

One of the many successes of the installation was the ease with which the operators were able to distribute load equally among the six load points. During installation it was determined that the weighbeam on both scales were warped along the longitudinal and lateral axis of the main weighbeam. This was a big concern because it would affect the mechanical repeatability of the scale.

After the scale was wired, an Auto Initialization was performed on the weight controller to detect the six new load cells in the system. An ID was then assigned to each load cell to identify its location in the transducers’ setup menu (i.e. "NE" North East, "CE" Center East, "SE" South East). Also, by assigning a display index position for each load cell the weight information could be viewed simultaneously on a 4 by 20 vacuum florescent display. The zero offset and full-scale output of the load cells in terms of raw counts was optionally entered into the controller as well.

From this point it was intuitive to adjust the elevations of each load point with a turn of a wrench. It was easy to quantify how the system shifted an applied load during adjustments by viewing each load cell’s output in real time. Because of this, installers were able to equally distribute the load within 10 percent of each load point in only 20 minutes. This was a considerable improvement over the previous method (using shim stock) which normally took 4 to 8 hours to adjust the load within 25 percent.

Return to top of page

Remarkable results

Once the scale was balanced and prior to any calibration with known standards, the scale was zeroed and loaded with U.S. Steel’s test standard for the first time. The standard is a 46-foot long test pipe with welded end caps and had been weighed three to five years ago by and outside testing facility and found to be 2,020 lbs. The first weighment on the digital system set at 1 pound increments yielded a measurement for the test pipe at 2,019 pounds. This was repeated several times to check for initial repeatability with success.

Not satisfied with this remarkable "out-of-the-box" result, installers began loading the scale with various sizes of pipe to investigate how well it repeated over a range of weights. Pipes weighing 661, 985, 1,024 pounds and the test standard at 2,020 pounds were used on the scale. Each pipe was weighed 10 times from both sides of the scale in increasing and decreasing order. With insignificant error, each pipe repeated its initial weighment value over the range of weights we tested. It was now time to formalize a more controlled test on each of the scales.

At this point it was agreed not to calibrate the scales to the 2,020-pound standard despite the nominal 1-pound deviation on the initial display. The reasoning behind this decision was the possibility that the pipe had lost a small amount of mass over the past three to five years. Also, considering the load cells were calibrated using an accurate standard, any error in the weighment would be will in the ± 0.5 percent error in each weighment specified in USS’s standard operation procedure (SOP).

Testing each of the scale’s repeatability over multiple tests was next on the agenda. Each scale was tested three times in runs of 20 with the2,020 pound standard. The results are as shown in the table below.

When considering the data in this table, it is important to note that a displayed resolution of 1 pound is one part in 26,455 divisions. What was really exciting was the mechanical repeatability of the system with a displayed resolution of 0.2 pounds, which is over on part in 132,000 divisions! Can the true accuracy of the scale be calibrated at such high resolutions? The answer is yes, with further testing. Is there confidence in the scales’ performance and relative accuracy? "You bet," says Postek.
Has anyone been able to get a scale to be as repeatable as this system is at these capacities?
"Probably not," said Postek. "Clearly the use of creative mechanics and digital performance have combined to yield such a satisfactory result."

Return to top of page
Request your quote