With pressure to minimize downtime at an all-time high, a properly applied SDA program could prove to be one of the most effective resources at your disposal

The winter maintenance turnaround will be, as always, a critical time for the modern cement plant. Cement market dynamics took an abrupt shift in the later stages of year 2000. Soft demand, energy concerns, and international consolidation will make for a different playing field in 2001. Still, U.S. cement demand will outstrip capacity by 25%. And now that the 1990s have brought some degree of sanity to cement imports, we continue to see a general increase in plant utilization compared to last year, with sold-out conditions in most markets. So challenged, management will demand that productivity and efficiency be maximized during 2001.

Universally, accounting protocols funnel the costs of last winter's maintenance shutdown into the operating budget of 2001. This impact is tremendous, if not for the financial impact but for the lack of clinker production during this time. Many plants find that over the year they never recover to budgeted levels should a prolonged outage befall them. Managers must do everything in their power to minimize this downtime through proper scheduling of labor, materials, and other resources, while at the same time providing the corrective repair and restoration that the equipment requires. A properly applied program of Statistical Downtime Analysis (SDA) will prove to be one of the most effective resources at your disposal.

Thomas Edison contributed his business sense and imagination to many industries, including portland cement manufacture. He was no stranger to the perils and risks of high-temperature pyroprocessing. During a moment of reflection, he stated, “Development of kilning equipment will always be hampered by the chaos brought by the equipment failure.” What did he mean?

When things went wrong in Edison's day, they went very wrong. The ability of Edison to reconstruct the failure and pinpoint the cause was hampered by the rapid sequence of disasters inherent in any kiln failure and the high-temperature damage to collateral systems once the original single system fails.

The simple physical limits of the construction materials of the time often were the culprit in such high-temperature disasters. The landscape is not much different today. The situation Edison found himself in demonstrates his respect for the importance of such a forensic analysis, recognizing that he could reallocate his resources for the greatest improvement once the offending piece or system could be properly identified.

To support the enterprise, the day-to-day objective of any plant maintenance operation is to minimize downtime. A simple demonstration of why downtime analysis is important is to first recognize the cash flow derived from the otherwise lost production. We will skip going into these calculations in-depth but show the key components:

First, recognize the lost production in any downtime situation (winter or otherwise) as incremental production. Incremental production is defined in the textbooks as the next ton produced after all fixed costs have been absorbed. Simplistically, it often is looked at as the final production run of the year, after budgeted tonnage expectations have already been made.

Incremental production creates a cash flow (or profit) equal to the selling price, minus the variable costs. Fixed costs are excluded. At many North American plants, incremental production generates $65 minus $20 = $45 per ton or more. This is the most profitable tonnage a plant can generate. The investment justified to achieve the incremental production — the money you can focus to correct the cause — also will be shown to be tremendous, perhaps the best return on investment you will make all year.

A cause is identified that stops the kiln at least once each year. For illustration, let's allocate to the correction two days plus four hours of repair work. The published cool-down and heat-up schedule adds another 48 hours. So 100 hours per year lost in a 100-tph kiln yields 10,000 tons or $450,000 lost per year. This is over the depreciation life of the repair, generally four to seven years. What investment will return $450,000 per year?

The five-year Internal Rate of Return (IRR) calculation shown in Table 1 can be punched out on the computer to show that $1.5 million can be invested, still yielding an acceptable return, provided it removes the cause of that downtime for the period considered. Applying this money should make you a hero; but if misapplied, you might wish you had never been born. So how does one determine where to spend it?

Operators apply SDA every day, often subconsciously, but it needs to be brought to its most sophisticated level by plant professionals. The cement industry, and the profits at risk, demand it. I'm not convinced that the concept is practiced within maintenance and engineering groups with the religious-like fervor that it should be. That is why the maintenance department exists, to keep the plant running. Simply stated, SDA is a complete and thorough analysis of what things shut a plant down: what caused the plant to go down, how many times the plant was down, and how many hours it was down (including the cool-off and warm-up penalties). If you don't have a program, one needs to be started. If you do have one, improve it. Except for developing the human resource, there is not a single more productive area to focus staff energies.

The intellectual tools needed for an effective SDA program are not great. Merely a summary understanding of statistical and financial analysis need be applied by the engineer for this effort to be effective. He or she also must have a good understanding of the production process. Corporate job descriptions for plant engineer or process engineer (either one) may include these responsibilities depending on the plant. In some cases, a maintenance superintendent is charged with the job. In other words, expect your staff to have the ability to do this. If the staff person that you have assigned to do your SDA doesn't have these tools, you've got the wrong person doing it.

The scope we work within here is bounded by two countervailing tendencies, and a casual departure beyond those boundaries will yield meaningless results. The first is too little segregation of the potential causes of downtime; the other is too much. This is admittedly a delicate balancing act and one often developed through trial and error.

So what are the components of an effective SDA?

1. Develop a list. The simplest list to develop is an equipment list, including all motors, fans, lube pumps, etc. Your Computerized Maintenance Management System (CMMS)-derived plant equipment list, correlating to your plant numbering system, would be an excellent place to start.

2. On a daily basis, allocate the stop, plus the associated downtime, to the equipment or system. Again too much or too little specificity can be detrimental. However, if your clinker breaker, for example, is belt driven, clinker breaker drive belts, clinker breaker hammers, and clinker breaker bearings should be sub-components under the clinker breaker category. Use your imagination and use the power of the sophisticated statistical computing packages within your computer applications (database, sort, histogram, Pareto chart, etc.). It's important not to forget non-equipment causes, including mix parameters, weather, late deliveries, and human error. Be very cautious on the latter. Don't use it as a catch bin. Deal with the statistic with the same concern as any other condition on the list and pursue a proper correction. Needless to say, there should not be many of these showing up regularly.

3. Summarize on a monthly basis the accumulated events and hours this simple three-column list has developed into. This step is simply to provide summary data to the manager for inclusion in the monthly reports. Meaningful conclusions are not likely to develop at this stage.

4. The holy grail of an SDA program is saved for this step: A 12-month running summary of the accumulated data. This is the report that will show where to spend the money. Arrange a histogram chart like Figure 1, with the significant few big numbers on top and the numerous trivial causes decreasing thereafter. Then go after the big fish on top whenever you can. These corrections are of course not limited only to winter maintenance. Many situations can be dealt with during the brief stops for other causes. Develop an action plan and be ready to tackle appropriate culprits the next time your kiln goes down. Seek the opinions of specialists or vendors if necessary. This list also will become a key component of your Predictive Maintenance Program, a program that will cause you to be nominated for Executive of the Year by your storeroom staff.

   (The point of this article is to prove what you can do for your own operation. However, there are a few reliability consultants specializing in the cement industry. An audit by these groups can gain you much bang for your consulting buck if you have a full understanding of Table 1's valuation process.)

   A useful appendix to the 12-month histogram report is a statistical table showing the number of discrete outages that lasted for less than two hours, those from two hours to 24 hours in duration, and finally those that took more than 24 hours to correct. Figure 2 shows this applied to a “Best in Class” analysis, where sister plants are compared; and to a year-on-year evaluation, looking for trends within a specific plant or system.

5. This is the most important. Have the same person assign the cause during each review. Don't leave it to the person on shift. This should be the SDA administrator, because that will be the person who will have to compile the annual summary. The following is an example of one of my experiences.

I'm going back to my old example of the clinker breaker. Your kiln is down because a clinker ball has burned the drive belts off the breaker. The cause is a badly worn drive sheave on the v-belt drive. The correction is obvious and simple. However, I have seen the following reasons given for such a shutdown (in addition to the correct one, which I will title “Clinker Breaker — Drive”): Chunks, kiln slabbed, cooler, cooler plugged, ring fell out, mix, hammer mill, clinker breaker, worn hammers, out of kiln gun shells, etc. You get the picture.

The SDA administrator must make a thorough determination the next day and log his conclusion. He is likely to make the same conclusion the next time this occurs as well (building the statistical base, revealing the evidence). If left to three different operators, we may never know. Such a mish-mash of bogus, though well-intentioned, information would prove useless when a third party needs to seek the proper fix.

Another peril is categories that are too broad. They become catch bins where important evidence is lost, and a collection of correctable faults are accumulated as one. One I am seeing more and more lately is the “power bump.” What does this mean? For my purposes, I want to know:

Power interruption — external network. I'll hammer on the utility company.

Power interruption — internal network. Call in the electrical engineer, or maybe it's time to clean the insulators in the sub station.

Electrical Transient. Perhaps the solid state motor control center (MCC) is too finely tuned. Let's dial it back.

Motor starter trip. Overheating, arced contacts, dead otters in the drainpipe?

Finally, take a deep breath and step back from the data. Is there another way of looking at it? Is there something you may have missed? An important tip is to look at the frequency of the category rather than the total hours down. It is an especially important analysis for the kiln systems. These small fish are the “kiln piranhas,” because they can eat you alive. Any interruption or slowing causing a cooling of the burning zone or a shift in the location of the burning zone, will take coating (and a layer of refractory) with it. Before you know it, the brick is gone. The moral is not to ignore high-frequency, short-duration (under two-hour) stops, especially if they affect the burning zone.

Armed with this information, ask your staff “The problem is clear, what are we doing about it?” and you are more likely to get effective answers and cost-saving results. The old sign said:

SCHEDULE OF RATES

Correct Answers	$4
Answers	$2
Dumb looks	FREE

Clearly, in today's sold-out cement plant, the sign can be changed to say that dumb looks and incorrect answers are extremely costly. SDA is a very important tool to keep equipment on line and get equipment back online once it goes down. Review policies and procedures in this area; challenge the staff to make improvements; provide incentive; promote the program; track your progress; and celebrate your gains.

Kenneth J. Rone Jr. is vice president — manufacturing services, Ash Grove Cement Co., (+1) 503-293-2333.

TABLE 1: Five-year Internal Rate of Return Calculations — Annual Volume: 10,000 tons; Useful Life: 5 years

Year	0	1	2	3	4	5
Incremental Mill Net Margin		$450,000	$450,000	$450,000	$450,000	$450,000
Operating Cash Flow	$ —	$450,000	$450,000	$450,000	$450,000	$450,000
Initial Expenditure	$(1,500,000)
Cash before Interest & Taxes	$(1,500,000)	$450,000	$450,000	$450,000	$450,000	$450,000
IRR = 15%