Spill control: assessing your situation: good spill control can help curb costs and calm public perceptions. Specific conductivity is the most successful parameter for monitoring black liquor spills.

Editor's Note: This is the first part of a two-part article. The second part will appear in a future issue of Solutions!

In most mills, improved control of spills of black liquor, white liquor, green liquor and paper coatings is the most cost-effective approach to reducing effluent discharges and lowering wastewater treatment system (WWTS) operating costs. Capital costs are generally low (from trivial up to a few hundred thousand dollars). Spill control operating costs are normally modest.

All mills discharge certain liquid effluents to remove substances that cannot be tolerated in the product, or in the recovery boiler in the case of kraft mills. These include cooling water, boiler blow down, bleach plant filtrates and many more. Other planned discharges, such as filtrates from brown stock screening, wet debarker effluent, and acid wastes from chlorine dioxide generation, can be eliminated by use of suitable equipment.

Planned discharges are generally predictable, and can be established quite accurately by a mass balance, particularly if modern process simulation tools are used. Some unplanned discharges are inevitable, however, and are usually contained by mill safety systems. These "spills" include leaks, off-spec product dumps, tank overflows, and equipment drainage prior to shutdown.

Unplanned discharges typically have a greater impact on WWTS costs and on public opinion than do planned discharges. In a modern mill with closed screening, oxygen delignification, ECF bleaching, condensate stripping, and a good WWTS, spills can cause up to half the color in the final discharge, and up to one-third of the operating costs of the treatment plant. Reducing spills will often allow a mill expansion without also expanding the WWTS, or having to apply for a modification to the effluent discharge permit.

U.S. RULES

The Cluster Rule promulgated by the U.S. Environmental Protection Agency (EPA) in 1998 includes requirements for mills to implement "best management practices" plans (BMP). These, in effect, are spill control systems for black liquor. The essence of the rule is that mills must record variability of effluent before treatment and take corrective action to reduce peaks. The rule leaves decisions on control criteria to mill management. It is worded in such a way that imaginative statisticians could derive BMP plans that comply with the law without improving effluent characteristics. Such plans become merely a paperwork burden on mill start; with no benefits.

Most U.S. mills are actively working on compliance with the spirit as well as the letter of BMP rules. While such projects almost always improve spill control, compliance with EPA guidelines alone does not guarantee optimal spill control.

WHAT TO MEASURE?

The prime target for most spill control programs is black liquor. Other substances, such as soap and turpentine, are a concern because they can cause catastrophic spills that drastically reduce the performance of the biological section of the WWTS. That is normally avoided by total closure in the same way as bunker C oil spills are prevented from reaching the receiving waters.

Specific conductivity is the most successful parameter in current use for continuous monitoring for black liquor spills is. It will also detect spills of white and green liquor, and usually soap (because stone black liquor normally travels with soap). Conductivity measurements will not detect paper coatings or turpentine.

Most other parameters that may appear to be of interest suffer from cost or maintenance complexity of the sensors (for example total organic carbon, immediate BOD, color, and sodium).

Spills have to be monitored instantaneously by conductivity or other continuous sensor, at multiple points, to initiate appropriate corrective responses. A longer-term assessment is also useful, particularly to benchmark your mill against other mills. COD and color are useful for such assessments.

Most mills consider that an effluent stream should be pumped back to the black liquor system when the conductivity exceeds 5000 micromhos/cm ([micro]mhos). This corresponds to a black liquor concentration of about 0.5%. Some mills have set points for recovery as low as 2500 [micro]mhos. You should determine and correct the causes of the high conductivity whenever possible.

PRACTICAL EXAMPLE

The three graphs at right (Figure 1) represent six months of daily color flows for three mills, expressed as kilograms of color per average ton of pulp production at the influent to the WWTS. The mills are similar. The principal difference is the execution of the spill control programs. All mills have two fiber lines. The mills have paper machines built before 1940 and are located on severely restricted sites with far from optimal layout of the equipment and sewer systems. All three mills have complete coverage with spill recovery sumps, and a number of conductivity monitors with data display and alarms on the operator's control panels. Management at all three mills believe they have excellent control of spills. In the author's view, the performance of the systems shown ranges from average to excellent.

[FIGURE 1 OMITTED]

Clearly the staff at mill S are the most effective at controlling spills. The data in Table 1 show that their color discharge is also lowest. (In all cases the color in the treated effluent is about 30% lower.)

Table 1 also shows that the data underlying the "better looking" graphs exhibit lower coefficients of variance (CoV), but the reader is cautioned against making a simple assumption that this value is a good measure of spill control. One major spill in a year would raise the CoV for mill S to well above the others, but it would still be a mill with excellent performance. Spectacular spills are normally due to known events that mills learn not to repeat.

The CoV is also misleading in a mill with high constant losses due to poor brown stock washing or other reasons, since a very poor spill control system is masked.

WHAT IS "GOOD" SPILL CONTROL?

At first sight, it seems obvious that a statistically based measure could be used to assess effluent data from departments or entering the mill's WWTS and provide a barometer of the effectiveness of spill control. However, nobody has yet developed a practical tool for this. NCASI bulletin 805, published in May 2000, provides an indication of the complexity of the issue, but is oriented towards assessing compliance with EPA regulations, rather than to optimal reduction of spills.

Low discharge of color or COD is generally indicative of good spill control, partly because mills with aggressive spill control systems usually also have tight control of planned discharges. For mills with oxygen delignification, discharges, before treatment, below about 40 kg color/ton or 30 kg COD/ton, suggest good control. If the mill does not have oxygen delignification, then both values would be about 10 kg/ton higher. These are only rough rules-of-thumb, since COD and color discharges depend on several aspects of mill operation.

Table 1: Statistical summary of color discharges at entry
to WWTS

                          Mill S   Mill C   Mill L

Mean color discharge,       39       52       59
  kg/ton
Standard deviation           6       19       25
Coefficient of variance     15%      36%      42%

"kg/ton" refers to kilograms of color measured each day divided by the
mill's average production rate, both measured over a six month period
in 2000.

This is one of a series of columns prepared by the Bleaching Committee of TAPPI's Pulp Manufacturing Division. For more information on TAPPI's next Process Closure Course, contact Tony Johnson by email at [e-mail] or Neil McCubbin at [e-mail]

Neil McCubbin is president, N. McCubbin Consultants Inc., Foster, Quebec, Canada. Reach him at [e-mail]