Pulp and paper manufacturing is among the most energy consuming industrial sectors. For example, typical Kraft mill in Canada consumes 25 GJ/adt1 of pulp produced and a recently built mill consumes 12 GJ/adt (Browne, 1999). In the Kraft pulping process, the lignin which binds the cellulosic fibres in the wood chips is solubilised by a strong alkaline solution at high temperature and pressure (Smook, 2002). This operation is conducted in reactors called digesters. The spent liquor from the digesting step, the black liquor, is concentrated in a series of evaporators and burned in recovery boilers. In the combustion process, a smelt of sodium sulfide and sodium carbonate is produced and recovered. These chemicals are then recausticised with lime and recycled back to the digester. The cellulosic fibres are separated from the black liquor before it is concentrated in a series of counter-current washers and to form the paper pulp. The pulp is bleached, dried to about 90% solids, cut in sheets and baled for shipment to customers. A simplified schematic of the complete Kraft process is given in Figure 6.
The mill supporting this study produces 700 ton of high-grade bleached Kraft pulp per day. The average steam requirement of the mill is 161 MW (19.1 GJ/adt) produced by four boilers. High-pressure steam is produced at 3034 kPa (T = 371°C) by two recovery boilers (64% of total load), a biomass boiler using wood residue (27%) and a bunker oil boiler (9%). About 25% of the high-pressure steam is used directly, 25% is throttled down to medium pressure (MP) at 965 kPa (178°C) and 50% to LP at 345 kPa (144°C).
Data were taken from an existing simulation of the mill to develop the composite curves (Figure 4); 61 streams were selected as potential heat sources or sinks (Mateos-Espejel et al., 2008b). A minimum temperature approach ΔTmin of 10°C has been chosen, which yields the pinch temperature of 71°C; where hot and cold pinch temperatures are 76 and 66°C, respectively. The maximum heating and cooling requirements of the process are 178 and 65.2 MW, while the minimum heating and cooling requirements (MHR and MCR) are 122.8 and 10 MW, respectively. It was shown in a different study that by maximisation of the internal heat recovery, there is an opportunity to reduce the steam consumption of the mill by 30 MW (Mateos-Espejel et al., 2008a).
Several feasibility studies of implementation of AHPs in the P&P Industry have been published (Abrahamsson et al., 1994; Gidner et al., 1996; Costa et al., 2009). None of them is based on the results of Pinch Analysis. Since the pinch-point temperature, which plays an important role for the correct positioning of AHPs, is not presented and different mills have different pinch-points, it is not easy to find if those implementations are well positioned or not. However, it can still be determined for which pinch-point ranges they would have been positioned correctly. For example, in the study by Gidner et al. (1996), an AHP type II is implemented in the evaporation section in which the evaporator, generator, absorber, and condenser temperature are 95, 69, 125, and 20–25°C, respectively. For the correct positioning of AHP, considering ΔTmin of 10°C, the pinch-point should have been between 90 and 120°C, which seems high for a Kraft process.
In a previous study by Costa et al. (2009) presented a preliminary feasibility study of the implementation of various AHP configurations in a Kraft pulping process using data from the same manufacturing mill. They have shown that the implementation of AHPs is feasible and cost-effective. Since the implementation of the pumps into the process did not observe the principles that are formulated here, in some cases the overall energy benefit is not as assumed in those studies. As an example, a double-lift AHP type II was installed to recover heat from the steam flashed at the discharge of the digester. That installation violated the pinch rule by using a stream at 96°C to produce LP steam, therefore, affecting a transfer of heat loads rather than a reduction of steam demand. The present case study shows how the proposed method can be used to avoid such errors and correctly position an appropriately designed AHP in the process.
Step 1: Table 1 shows the selected streams that are available heat sources and sinks in the process after maximisation of the internal heat recovery. HB1 stands out as an interesting heat source; its temperature (75.9°C) is just below the hot pinch temperature and it carries a high load (12.2 MW). There are five heat sinks available, CA1–CA5. However, it should be noted that because of high-pulp concentration, CA1 and CA2 will not be considered due to the difficulties of handling such concentrated streams.
Table 1. Selected streams
|Section||Description||State (C or H)||Tstart (°C)||Ttarget (°C)||Flow (kg/s)||Heat load (kW)|
|Heat sinks above the pinch-point (CA)|
| Bleaching||Pulp through vapormixer 3||C||68.6||75.0||58.8||1471|
| Bleaching||Pulp through vapormixer 4||C||71.9||85.0||51.7||2626|
| Evap. #3||Liquor through HXC||C||104.9||132.2||14.0||8907|
| Evap. #2||Liquor through HXC||C||101.4||127.8||14.5||7363|
| Dearator||Condensate and fresh water||C||66||100||67.6||9653|
|Heat sinks below the pinch-point (CB)|
| ||Fresh water from 4 °C (winter) or 20 °C (summer) to any temperature below the pinch-point|
|Heat sources above the pinch-point (HA)|
| Boilers||Flue gas from CR2||H||164.0||105.8||63.5||4314|
| Boilers||Flue gas from CR3||H||199.0||117.6||47.6||4625|
| Boilers||Flue gas from bark boiler||H||182.0||126.5||47.7||2863|
|Heat sources below the pinch-point (HB)|
| Evap. #2||Vapor condensing||H||75.9||74.9||5.3||12 211|
| Bleaching||Effluent, washer 1||H||58.1||33.0||58.8||6150|
| Bleaching||Effluent, washer 2||H||68.6||33.0||67.0||9911|
| Bleaching||Effluent, washer 3||H||68.2||33.0||18.8||2768|
| Bleaching||Effluent, washer 5||H||66.8||33.0||9.5||1338|
| Bleaching||Effluent washer 4||H||73.1||33.0||36.0||6027|
Step 2: As a result, the combination map is HB1& CA3, HB1& CA4, and HB1& CA5.
Step 3: In the next step the temperatures are estimated. Considering the temperature of the heat source stream (75.9°C) and ΔT = 10°C, the maximum evaporator temperature should be 65.9°C. The minimum condenser/absorber temperature or absorber temperature depending upon the type of HP that may be utilised (AHP type I and type II) are estimated as 142, 138, and 110°C for CA3, CA4, and CA5, respectively. In this example, six different configurations (single-stage, double effect, and double-lift AHP type I or type II) for the two common working fluid pairs (LiBr-H2O and NH3-H2O) are considered; this gives 36 different combinations. Table 2 shows the thermodynamically feasible configurations for the resulting combinations. Those that are not presented in the table are too remote from feasible conditions; some would lead to excessive generator temperatures for an AHP type I, some to very low condenser temperatures for an AHP type II others would fall in the crystallization zone. The estimated generator temperature for the AHP type I cases and condenser temperatures for the AHP type II cases are presented for the 20 remaining feasible configurations.
Table 2. Thermodynamically feasible configurations
|Case||H.Si* stream||Configuration||Working pair||COP||QG (MW)||QE (MW)||QA (MW)||QC (MW)||TC (°C)||Low-TH.Si||TG (°C)||High-TH.So|
|1||CA3||Type I SS||H2O-LiBr||1.72||5.17||3.73||5.17||3.73|| || ||250||HP|
|2||DL||H2O-LiBr||1.3||6.85||2.05||3.97||4.93|| || ||190||HP|
|3||NH3-H2O||1.23||7.24||1.66||4.41||4.49|| || ||200||HP|
|4||Type II SS||NH3-H2O||0.42||5.12||7.08||5.12||7.08||10||NF|| || |
|5||DL||H2O-LiBr||0.29||7.44||4.76||3.54||8.66||30||FW|| || |
|6||NH3-H2O||0.27||5.25||6.95||3.29||8.91||30||FW|| || |
|7||CA4||Type I SS||H2O-LiBr||1.72||4.3||3.1||4.3||3.1|| || ||220||HP|
|8||DL||H2O-LiBr||1.3||5.69||1.71||3.3||4.1|| || ||190||HP|
|9||NH3-H2O||1.23||6.02||1.38||3.66||3.74|| || ||190||HP|
|10||Type II SS||NH3-H2O||0.42||5.12||7.08||5.12||7.08||10||NF|| || |
|11||DL||H2O-LiBr||0.29||7.44||4.76||3.54||8.66||35||FW|| || |
|12||NH3-H2O||0.27||5.25||6.95||3.29||8.91||30||FW|| || |
|13||CA5||Type I SS||H2O-LiBr||1.72||5.61||4.04||5.61||4.04|| || ||160||HP|
|14||NH3-H2O||1.55||6.23||3.42||6.23||3.42|| || ||170||HP|
|15||DL||H2O-LiBr||1.3||7.41||2.23||4.31||5.34|| || ||140||HP-MP-FG|
|16||NH3-H2O||1.23||7.85||1.8||4.78||4.87|| || ||150||HP-MP-FG|
|17||Type II SS||H2O-LiBr||0.47||5.73||6.5||5.73||6.5||30||FW|| || |
|18||NH3-H2O||0.42||5.12||7.08||5.12||7.08||25||FW|| || |
|19||DL||H2O-LiBr||0.29||7.44||4.76||3.54||8.66||30||FW|| || |
|20||NH3-H2O||0.27||5.25||6.95||3.29||8.91||30||FW|| || |
Step 4: Once the COP value has been estimated by the Ziegler & Alefeld method (Ziegler and Alefeld, 1987), the load for each component can be computed and their values are given in Table 2 (Q's). For all AHPs type II cases, fresh water could be used as low temperature heat sink; except in cases 4 and 10, where the condenser temperature is too low and that will not be considered. For all AHPs type I, high-pressure steam could be used as the high-temperature driving energy. For some of them, MP steam could also be used (cases 15–16) and for others, part of the required heat load could come from flue gases (cases 15–16) (step 5).
Step 5: It consists of eliminating additional combinations by considering design and operating constraints. Some of the cases are clearly impractical and are readily eliminated. Considering the corresponding working pair equilibrium diagram; cases 1, 2, 3, 7, 8, and 9 are eliminated because of the high-generator temperature (they have generator temperature at 190°C and above), cases 4 and 10 are eliminated because 10°C would be required for the condenser temperature. Cases 1 and 7 could also be eliminated because of the high risk for crystallization and Case 14 is eliminated because of the high-operating pressure of its cycle.
Figure 7a and b illustrate why cases 1 and 10 are eliminated. In Figure 7a, the condenser/absorber and evaporator temperatures are used to illustrate the thermodynamic cycle of case 1. The estimated generator is clearly too high for current AHP technology (250°C). Also, because of the high-LiBr concentration in the solution side, there is a potential risk of crystallization during operation. Figure 7b shows that case 10 is eliminated because of the low condenser temperature. Reaching a temperature of 10°C in the condenser with a ΔT of 10°C, would entail a heat sink at or below 0°C, which is not practical. It should be mentioned that it has been verified by observation of the site mill layout that all selected streams are relatively close to each other and will not require extensive piping work. Another reasonable constraint to be considered to eliminate additional cases is the fact that as much as possible of the available energy of the selected heat source (12.2 MW) should be upgraded. Cases 13, 15, and 16 are thus eliminated, because they use only 33%, 18%, and 15% of the available 12.2 MW, respectively. At the end of this elimination procedure, 8 cases (cases 5, 6, 11, 12, 17, 18, 19, 20) remain from the 36 identified originally; they are all single-stage or double-lift AHPs type II. They will now be compared on the basis of economic and thermodynamic criteria for the final choice.
Figure 7. Eliminated cases in LiBr-H2O and NH3-H2O equilibrium diagram; (a) case 1; high-generator T and risk of crystallization, (b) case 10; low condenser T.
Download figure to PowerPoint
Step 6: Case 17, which has the highest estimated delivered heat (5.73 MW) as a single-stage AHP type II, and case 19, which has the highest estimated delivered heat (3.54 MW) among the double-lift AHPs type II, are chosen to be dimensioned and cost estimated, as they are likely to be the most attractive options.
Step 7: For this purpose, a simulation model for the single-stage and the double-lift AHP type II using the LiBr/H2O working pair was used (Bakhtiari, 2009). It calculates the internal mass and energy balances of each component as well as the heat transfer between external and internal streams in a steady-state operation. The parameters of a cycle such as output temperatures, pressure of each components, heat load of heat each exchanger and COP are calculated from input values such as input temperatures and mass flow rates. The results from this simulation were validated with data from the literature (Herold et al., 1996). Table 3 presents the basic design parameters and simulation results. The useful delivered energy in the absorber is calculated at 5.83 and 3.89 MW for cases 17 and 19. The power to be supplied to the generator and evaporator were then calculated at 5.76 and 6.55 MW for case 17 and 7.45 and 4.76 MW for case 19. The COP is also calculated as 0.48 and 0.32 for cases 17 and 19. Figure 8 shows the feasible cycles in the LiBr/H2O equilibrium diagram.
Table 3. Basic design parameters for selected cases
|Case 17 (SS AHP type II)|
| LMTD||°C||9.3||10.3||17.3||10.9||12.8|| |
| Pressure||kPa||26.3||3.9||26.3||26.3||3.9|| |
| U.A.||kW/°C||693.5||559.2||337.0||100.0||498.4|| |
| Q||MW||6.5||5.8||5.8||1.1||6.4|| |
|Case 19 (DL AHP type II)|
Figure 8. The selected cycles in LiBr/H2O equilibrium diagram; (a) case 17; single-stage AHP type II, (b) case 19; double-lift AHP type II.
Download figure to PowerPoint
The economic feasibility of those two AHP implementations was assessed. The installed cost of each device was estimated on the basis of output heat in $/kW. The following cost compiled by the US DOE (RCG/Hagler-Bailly, 1990) were used,2 581 and 656 $/kW for single-stage and double-lift, respectively (those costs are almost identical to those presented by Berntsson and Frank, 1997). The annual rate of operations, steam and cooling water costs were taken as 8640 h, 62.5 $/MWh, and 1 $/MWh, respectively as suggested by mill personnel. Maintenance and other operating costs were neglected. The estimated installed cost of the AHPs and the annual saving in steam and cooling water are presented in Table 4. The simple payback times (SPB) given by Equation (1) are 1 and 1.2 years for the single-stage and double-lift AHP type II.
Table 4. Economic evaluation of selected cycles
|Case||SS AHP type II (case 17)||DL AHP type II (case 19)|
|Installed cost [M$]||3.4||2.6|
|Steam saving [M$/a]||3.2||2.1|
|Cooling saving [M$/a]||0.11||0.11|
Life-cycle costs of the two case studies were compared using the net present value (NPV; Costa et al., 2009). The initial investment cost and the yearly energy savings were estimated over the life cycle, which was set at 15 years. A discount rate of 7%, which is usual for energy projects and a yearly escalation rate of 4% for fuel price have been used for the time span of the investment. Figure 9 shows the evolution of the NPV over 15 years; the abscissa gives the time elapsed from the date of investment and the intersection with the nil value line indicates the time required for the projects to become cost-effective. It is about 1 year in the two cases for the current steam price. The figure shows that the two options are interesting not only in the short term (quite short PBT), but also in the long term. The initial investment for case 19 is more attractive although both cases produce almost the same SPB. However, case 17 which produces high-yearly revenue perform best in the long term, even if the initial investment is higher. It should be noted that neither operating costs nor potential GHG emission credits were taken into account. The latter could be a significant factor of payback time reduction.
Step 8: Since two practically possible and economically feasible configurations have been already identified, the HEN reconfiguration is not considered in this case study (step 8). The authors presented another study in which the HEN reconfiguration and the effect of water closure on the pinch-point are also presented (Bakhtiari et al., 2010).