Sharing statistics on paper mill energy usage could help slash costs, but mills are failing to provide accurate, comparable numbers

by Gordon A Robb and Keith Grimwood

A standard approach to lower energy costs

T hermal energy is one of the main controllable costs in a pulp and paper mill. One energy management technique that could create opportunities for cost savings is the use of benchmark data. Comparing actual mill data with benchmark data can help mills track down the energy inefficient sections of their paper machine.

Before this method can be used with confidence, however, the benchmark data from the participating mills must be obtained using a consistent method. One of the problems at present is that different mills include different components in what they see as the same process.

To address this problem, the process energy standards subcommittee of the Canadian Pulp and Paper Association (CPPA) is developing methods to establish energy/ton standards for specific pulp and paper processes. The first step in developing comparable energy process standards is to develop uniform metering. CPPA is promoting a standard approach that states exactly which process energy requirements should be metered and reported.

Dry tale

The drying section on a paper machine is a prime example of where a standard approach can be developed and used. It is generally agreed that steam used for drying should include the steam heat used by the dryer cylinders plus steam used to heat air for dryer ventilation.

The ideal system would measure, indicate and integrate:

• steam to the dryer cylinders
• condensate from the dryer cylinders
• blow through condensing water flow and temperature rise
• steam to and condensate from the air heating coils
• and steam to the wet end (including steam direct to the wire pit, to the shower water and the steam boxes etc).

A point in time

The idea behind this management technique is for mills to monitor the energy consumption of their specific mill processes on a daily basis and then compare the results against an industry standard. Microprocessor systems are commonly available on most paper machines and they should be capable of calculating the number of BTU or GJ per ton being consumed based on point-in-time production rates.

Using this method may help identify wasted steam. For example, some papermakers may be using more steam than is justified by the benefits of higher process temperature. It can also be used to identify potential problem areas. For example, problems in a heat recovery system may be indicated if the ventilating steam heat requirement exceeds the standard level.

Cash conscious

Operators can be motivated to make energy-conscious decisions by making them responsible for the difference between actual controllable energy costs in $/ton and standard costs in $/ton. The cost accounting system should focus on variances from standards or targets and the variances should be expressed in both energy units per ton as well as in $/ton.

The system should also be based on direct costing or marginal analysis, rather than average costs. This is because the average cost of thermal energy is often far below the incremental or marginal cost, especially if steam is available from black liquor or thermomechanical pulp (TMP) steam recovery systems.

A change in the quantity of steam being used can be converted to $/ton by using standard fuel prices. But the system must be able to accurately link steam heat quantities produced with fuel quantities used for steam generation. Mill management should concentrate on the quantity variances converted into dollar terms so they can be compared with other cost variances.

Often an adverse variance can be explained by looking at paper machine efficiency. That is why the point-in-time approach is so important. Stating the machine efficiency at the time the measurement is made helps a lot.

Where the quantity variances in terms of $/ton indicate that action is justified, the point-in-time difference between the MJ/ton used by individual processes and the targets should indicate what should be done. This approach avoids the distortions inherent in numbers based on periods of time.

Counting cost

Problems can arise when those responsible for energy information systems, energy management and cost accounting do not fully understand the thermodynamics taking place between the steam flowmeter and the point at which boiler fuel is priced - the gas meter, the oil tank or the coal pile.

In a thermal energy cost accounting system the aim is to allocate fuel dollars to specific processes. In effect, fuel is bought in $/million BTU or $/GJ on the basis of higher heating value (HHV). The higher heating value (or gross caloric value) is the heat liberated when fuel is completely burned in air and the water vapor cooled to the initial calorimeter temperature. Another type of measurement is low heat value (or net calorific value) which is similar to HHV, but assumes that the moisture vapor remains in the gaseous state. The difference between the two is therefore the latent heat of vaporization of water.

The use of HHV in North America has been taken for granted in the past. But gas turbine suppliers often provide performance data in terms of lower heating value (LHV). So as gas turbines become increasingly used in combined heat and power (CHP) systems at pulp and paper mills, the LHV data should be converted to HHV data. This is because condensing heat recovery systems, where natural gas is the fuel, can recover this latent heat and use it to heat boiler feed water makeup and heat shower water etc. It can also be used to heat air, perhaps via a water/glycol system. This low-grade heat is particularly valuable in cold climates.

The efficiency of conversion from fuel to steam heat available for distribution to the mill must be known. The ideal calculation would be to use an input/output method treating the process as a black box. This method relies on accurate pressure/temperature compensated measurements of steam flow from the steam plant, condensate returned and makeup added. The problem is that most steam flowmeters are not temperature/pressure compensated and although compensation can be carried out using software in an energy information system, recent experience suggests that accuracy is not yet good enough.

In mills without accurate, temperature/ pressure compensated steam flowmeters, the steam plant output should be derived from the HHV of fuel input using a conversion efficiency. In other words the output should be determined from the fuel input and the conversion efficiency.

A way of calculating efficiency is via the loss method. In most cases, individual boiler steam flowmeters must be used together with estimates of steam plant losses and uses. This method of determining the efficiency of conversion from fuel to steam heat should be used from time to time because it is accurate and it points to where losses can be reduced.

CHP problem

Combined heat and power (CHP) plants introduce an interesting problem in allocating cost. If a non-condensing steam turbine is used, kWh's of electricity are produced as a byproduct of the process steam, without condenser loss. Using more process steam results in more cogenerated electricity if all the steam goes through the turbine. Similarly, the use of less process steam means less cogenerated electricity.

Without CHP, the purchased power price is normally related to the cost of electricity from a single-purpose fossil-fuel steam electric plant, where 48% of the higher heating value becomes condenser loss to the river. About 10 kJ of fuel heat are required to produce 1 kWh. CHP with a non-condensing turbine requires only about 4.43 KJ of fuel heat per kWh. This means that the fuel component of cogenerated power cost will be much less than that of purchased power. The cost of using an additional GJ of steam heat is partly offset by the CHP plant.

If a CHP system is well designed, more valuable electricity is generated as a byproduct of low-pressure steam than high-pressure steam. So using additional amounts of low-pressure steam heat will cost less than the equivalent amount of heat using high-pressure steam. One way to show this is to consider the CHP effect when translating the variance into dollar terms. The person responsible would then have an incentive to use low-pressure steam. The energy information system should indicate the steam pressure/temperature level.

Canadian comparisons

The Canadian newsprint industry has been comparing the amounts of drying steam it uses per ton of paper for decades, but the problem has been one of comparability. A CPPA report titled "Newsprint Machines - Comparison of Operations" includes most of the basic operating information. Of 137 newsprint machines surveyed in seven countries, 105 reported (Figure 1).

The accuracy of the measurements in the study are questionable, however, as the size of the range is unreasonably large. This is highlighted in the following analysis.

Theoretically it takes a minimum of 2,965 kJ to evaporate 1 kg of water (Figure 2). Assuming the moisture entering the dryers is 55% (which is about the lowest achieved in practice) and the final moisture content in the paper reel is 7%, this gives a minimum energy use of 3,162 MJ/ton (Figure 3).

Assuming that it takes a maximum of 3,665 kJ to evaporate 1 kg of water and that the moisture entering the dryers is 63%, the energy consumption becomes 5,547 MJ/ton. These numbers are based on the rate of steam heat consumption when the paper machine is operating normally.

The range of 2,200 MJ/ton to over 8,000 MJ/ton in the survey, extends far beyond the theoretical range of 3,162-5,547 MJ/ton. There are 13 machines below the minimum and 37 above the maximum reasonable number.

It seems safe to assume that there are measurement and/or calculation errors associated with the survey's numbers, which are outside the range of reasonable error. Some of the lower figures may include only steam to the dryer cylinders without considering pocket ventilation heat. Where the numbers are too high, the meters could be measuring more than just drying steam.

PM efficiency is another potentially important consideration. In the CPPA survey, the consumption figures are based over a period of time so paper machine efficiency inevitably influences the results. Steam heat consumption in MJ/ton will tend to be higher if PM efficiency is low for the reporting period. Saying that, the effect depends on the amount of time dry broke is being produced compared to wet broke. When a paper machine is making dry broke it uses just as much steam as when it is making saleable paper, but if breaks are at the wet end, much less steam is used. Machine efficiencies are included in the CPPA survey, which increases its potential usefulness.

Recent research

CPPA has developed a standard format for paper machine energy reporting (Figure 4). This format has been tested out on some of the few modern newsprint machines with good metering systems and has yielded some useful information. Data on a new groundwood specialties machine indicates a steam heat use of some 3,996 MJ/ton for dryer cylinders, pocket ventilation and a steam box. The month the results were collected was April (the season influences heating energy) and the moisture entering the dryers was 61.5%. If the room ventilation requirement is added the number becomes 4,297 MJ/ton.

Theoretical data showing the effect of moisture entering the dryer section on the steam heat requirement is shown in Figure 5. The practical data from the groundwood specialties PM confirms the reliability of Figure 5, where the lower theoretical figure of 2,965 kJ is used as the amount of energy needed to evaporate 1 kg of water.

If more mills provide data using this standard reporting formula, CPPA will be able to produce proper benchmark figures.

CPPA blueprint

The energy committee at CPPA has prepared a detailed set of instructions for the complete paper machine. There is too much detail for a complete analysis, but an explanation of what to include in the major blocks, which make up the PM, is shown in Box 1.

The blocks can be grouped together. For example, drying steam could be defined as blocks E+G or E+F+G or E+F+G+H. Blocks A+B+C could be combined under the heading wet end steam. The main thing is that the user should know exactly what information is contained in the number.

The overall aim should be the inclusion of thermal energy inputs and calculations in the distributed control (microprocessor) system. It would then be possible to calculate the MJ/ton from that system. If it is doing that, filling in the forms will be easy and reliable benchmark figures will be one step closer.

Block A - Stock preparation

Include kWh and MJ from the machine pulper up to and including the machine chest pump. Include additive systems from storage to the point of application.

Block B - Forming

Include MJ to maintain headbox temperature after the machine chest. MJ to deculator steam jets, showers etc up to and including the couch. Include kWh/ton to drive any device where stock passes through or over. This includes fan pumps, white water pumps, shower pumps and deculator vacuum pumps. Include kWh/ton for the former drive if measurable separately. For steam added directly to stock or white water allocate kJ/kg steam-kJ/kg cold boiler feedwater makeup.

Block C - Pressing

Include MJ added to showers, felt conditioners etc. Also include kWh to vacuum pumps and (if possible) to press drive.

Block D - Machine drive

KWh that are not allocated to individual blocks should be allocated to block D. For a steam turbine electric drive allocate 3.9 MJ/kWh to account for the heat removed from the steam as it passes through the turbine. For a line shaft drive the kWh must be estimated.

Block E - Dryer cylinders

If all condensate is returned to boilers allocate MJ/ton of steam-MJ/ton of condensate. Where condensate is lost use MJ/ton steam-MJ/ton boiler feedwater makeup. Metering condensate flow makes allocation to this block easier. Condensing rate tests can be conducted based on the time for a tank level to increase a measured amount with the outlet shut off.

Residual (non-recirculated) blow through would be included in the steam flowmeter reading. This should be allocated to block F, not block E. If E and F cannot be separated they should be combined to avoid double counting of blow through.

Enter the number of MJ in the block where the steam heat is first used (where the steam is condensed). Indicate in the diagram on the right hand side where the residual blow through goes, but put it in block F not the destination block.

Include energy supplied to air caps for yankee dryers, through dryers, on machine coaters, etc. There should be separate numbers for steam heat as opposed to direct firing heat. Use an additional block where appropriate. Allocate kWh for drives, which may reflect the amount of condensate in the dryer cylinders as an accumulation of condensate increases dryer load.

Block F - Dryer blow through condenser

The steam flow to the blow through condenser is the total steam flow to the dryer cylinder minus the amount of steam condensed. Alternatively put a flowmeter in the condensing water line and measure temperature rise. Indicate in the diagram to the right of the blocks where the condensing water goes (is it used as shower water, does it go to a warm water tank etc.) If the residual blow through is condensed by direct contact (quenched) and the condensing water mixed with condensate returned to the boiler, the amount and temperature of quench water should be measured to allow calculation of the amount of steam condensed in the dryer cylinders. In this case, the residual blow through heat is returned to the boilers. The heat has been degraded, however, so there remains a cost associated with excessive blow through.

Block G - Dryer ventilating system

Include the steam heat added to the air, which is added to the dryer ventilating system. Also include natural gas heat added for direct heating of this air. Do not include waste heat used to heat air. Use separate numbers for steam heat as opposed to direct firing heat.

Block H - Calender

Allocate steam heat used to heat water for calender heating (kJ/kg steam-kJ/kg condensate if returned). If condensate is not returned, substitute the enthalpy of the boiler makeup for the condensate enthalpy. This metering may require combining with drying. Include drive electricity in kWh/ton.

Gordon Robb and Keith Grimwood work for Thermoshare in Ottowa in Canada. This paper was presented at a recent Pira conference in the UK