In the following, results from experimental investigations, calculation of mass and energy balances, and gas burner tests will be presented and discussed.
Steam gasification of bark using different bed materials in the advanced 100-kW
Results with the standard fuel softwood (SW) are presented as a comparison to bark (BA). Three different bed materials are applied olivine (OL), quartz sand (QS), and limestone (LS). The main operating parameters for the test runs together with several KPI are presented in Table 4.
1Location of measurement points see Fig. 2
Figure 3 and Fig. 4 show the main dry product gas composition as well as tar content and tar dew point comparing the three different bed materials and bed material mixtures, respectively. For all six experiments, comparable product gas qualities could be achieved. During the experiments, an already reported correlation between increasing content of limestone and increasing H2 content could be verified. For the case of softwood as input, the H2 content could be increased from approx. 39.3 with olivine to approx. 47.5 Voldb% when limestone was used as bed material. The experiment with quartz sand and 10% limestone also shows the impact of using discharged limestone from the lime cycle of the pulp and paper mills as an additive in the gasification reactor (Fig. 3). This is also supported by the results of the tar measurements. The addition of limestone significantly reduces the tar content in the product gas and also, subsequently, reduces the tar dew point in comparison to the experiment with olivine as bed material (Fig. 4). It is especially noteworthy to mention that the GC–MS tar content (excl. BTEX) was 7.7 g/Nm3db in the experiment with bark using quartz with 10 wt.% limestone in comparison to 18.1 g/Nm3db in the experiment with olivine. Similar findings were gained for the gravimetric tar content with a decrease from 5.3 using olivine to 1.7 g/Nm3db using quartz with 10 wt.% limestone as bed material. Quartz is a non-catalytic material in comparison to olivine after activation by ash layers, which has been reported to show a certain extent of catalytic activity towards tar reduction. Therefore, it is noteworthy that the addition of limestone to the quartz bed material significantly influences the tar reduction in the reactor and subsequently improves the product gas quality.
In summary, the results from the gasification experiments in the 100-kW advanced DFB pilot plant have shown that bark from pulp and paper mills can be used to obtain a valuable product gas as a secondary energy carrier, which is a potential substitute for fossil-derived combustible gases. If quartz sand is used as bed material, the addition of discharged limestone from the lime cycle leads to an enhancement of the catalytic activity resulting in an improved product gas quality concerning the main product gas composition as well as the tar content. This increase of the catalytic activity is due to the Ca enrichment in the reactor, which also further leads to an increase of the Ca availability during ash layer formation on bed particles. Subsequently, this results in a long-term availability of the catalytic effect in the reactor, as reported in previous publications [23, 37, 43].
Gas cleaning and its impact on the separation of non-process elements
In the following, different gas cleaning setups for the separation of NPE from the product gas of bark gasification are discussed. Table 5 shows the gas cleaning equipment used for the different scenarios.
The considerations leading to the selection of a gas cleaning setup needs to include two main aspects which contradict each other to a certain extent: energetic optimisation vs. maximal reduction of NPE. Figures 5 and 6 show process flow sheets in which the equipment already existent in a pulp and paper mill is displayed in grey colour, whereas the gasification and subsequent gas cleaning are displayed in green colour. Here, the authors assume a paper mill where a biomass boiler for combustion of bark is already installed as a CFB boiler. This is not valid in general but represents a typical modern paper mill. The existing units of the process are not shown in detail. Figure 5 shows the energetically favourable process when focusing on the temperature levels. In this first setup, the particle separation after the gasification reactor is realised by a multistage cyclone. The product gas temperature is regulated to approx. 750 °C after the multistage cyclone. Therefore, all organic components in the product gas, including higher hydrocarbons—referred to as tar—are burned in the lime kiln. The sensible heat of the product gas with approx. 1550 kW is comparably high in this configuration. However, no internal energy is transferred to the steam necessary for fluidisation of the gasification reactor because no product gas cooling is installed. Therefore, during the operation of the gasification reactor, externally generated saturated steam needs to be continuously supplied. Furthermore, only a simple particle separation is integrated into the form of a multistage cyclone.
Figure 6 shows the process flow sheet of two other possible configurations, which differ in the temperature level of the product gas filter and the degree of the particle separation compared with Fig. 5. In the second setup, the product gas after the gasification reactor is cooled down to approx. 450 °C. Thus, the tars in the product gas do not yet condensate and are therefore burned in the lime kiln. In this configuration, the sensible heat of the product gas is approx. 767 kW, which is lower in comparison to the first process setup. However, as shown in Fig. 6, energy is transferred from the product gas to the steam supply of the gasification reactor in the product gas heat exchanger. The third setup, also displayed in Fig. 6, is similar to the second one with the difference, that the product gas is cooled down to below 200 °C, so that a conventional fabric particle filter can be used instead of a ceramic filter. In this setup, the sensible heat of the product gas is approx. 340 kW, which is significantly lower than in the first and second setups. In this third process setup, more surplus steam is produced in comparison to the second setup, which can be used as a fluidisation agent in the gasification reactor or in a biomass dryer.
Since gasification is realised using steam as a gasifying agent, the product gas has approx. 45–50 Vol% water content. Furthermore, tar components are present in the product gas. On the one hand, tars, which are organic molecules, contribute to the heating value of the product gas. On the other hand, they can potentially lead to problems in downstream equipment. Condensation of tar leads to either liquid or solid deposition which upon accumulation can result in undesired operation shutdowns. Therefore, as additional gas cleaning step a liquid scrubber unit is considered. As the satisfactory separation of NPE already results from significantly cooling down the product gas including a liquid scrubber unit is a further reasonable step in achieving the condensation of steam and the reduction of tars from the product gas. Figure 7 shows this fourth process setup further elaborating on the gas cleaning. It is noteworthy to mention that in the presented process setups, the gasification route does not produce additional waste streams from the process. Separated particles in the product gas filter, typically referred to as fly char, as well as the emulsion obtained from solving tars in the scrubber, are recirculated to the bark boiler. There, they are burnt and thus inherently disposed of. Only ash from the flue gas filter of the combustion reactor needs to be discharged from the system in the same manner as for direct incineration of biomass.
This product gas cleaning setup of heat exchangers, followed by a fabric filter and a liquid scrubber unit using rapeseed methyl ester as solvent was installed in industrial-scale dual fluidised bed steam gasification plants in Güssing and Oberwart, Austria, as well as in Senden, near Ulm, Germany, with fuel power between 8 and 15 MW . Therefore, it is possible to derive the separation efficiencies for tars and water in the product gas from long-term experience in these industrial plants. Besides, reducing the dust freight the product gas filter also reduces the gravimetric tar content by approx. 80% and the GC–MS tar content by approx. 34%. The subsequently installed RME scrubber further reduces the remaining tar content by another approx. 50% . Heavier molecules, referring to tar molecules with a molecular weight of naphthalene or higher, are almost completely separated in the scrubber, whereas the separation efficiency decreases for lighter tar molecules. However, the lighter tar molecules have a higher dew point which significantly decreases the risk for undesired condensation in downstream equipment.
Besides the energetic considerations, the impact of the different particle separation units on the reduction of the NPE in the system needs to be included in the discussion.
Figure 8 shows the separation efficiency of the four scenarios. Scenario 1, based on a multi-cyclone, leads to the introduction of significant amounts of NPE into the lime cycle, as the separation efficiency only reached approx. 90%. Based on the fact that NPE accumulate in the lime kiln, this scenario needs to be excluded for further considerations. Scenario 2 and 3 result in similar separation efficiencies of 99.0% and 99.5%, respectively. Regarding particle separation efficiency, these two scenarios can be considered satisfactory in their performance. However, as there is no liquid scrubber installed in either of the setups, the product gas is still containing approx. 50% of steam, which would be introduced into the gas burner. Therefore, scenario 4 is selected as the most promising process layout, based on its practically complete separation of NPE, scrubbing of tars and condensation of steam in the scrubber.
A complete mass and energy balance for scenario 4 is displayed in Fig. 9.
Combustion of product gas as a substitute for natural gas
Product gas from bark DFB gasification shall be used in the lime kiln, but this product gas shows different properties from natural gas. Therefore, an experimental study has been carried out in two different burners for comparison of the flames on the one hand of natural gas and on the other hand of product gas from bark gasification.
Table 6 shows the comparison of the results between natural gas and product gas combustion obtained from a free-standing experimental test rig. Two different capacities of approx. 100 kW and approx. 200 kW are included in this comparison. One of the main issues with substituting natural gas with product gas from steam gasification is the necessary gas flow to achieve the same capacity in the burner. Due to the lower heating value of the product gas in comparison to natural gas, a higher gas flow of fuel gas is necessary. As shown in Table 7, the flue gas stream after the lime kiln is slightly increased when product gas is used as the necessary airflow for combustion is lower in comparison to the natural gas. By enrichment with pure O2 in the gas burner – which has also an impact on the combustion temperature as discussed further below – the flue gas stream can be decreased which leads to a relief (or downsizing) of the gas treatment downstream. The obtained results are in good correlation with previous experience from gas burner experiments [45,46,47].
Furthermore, due to the high H2 content in the product gas, the ignition of the gas occurs earlier than is the case for natural gas. Therefore, the temperature in the burner itself is significantly increased from 260–326 °C when natural gas is used to 676–739 °C when product gas is used. Also, the flame length is decreased from 70–80 cm in the case of natural gas to 50 cm in the case of product gas. However, it is noteworthy to consider, that the experiments were conducted using a natural gas/air burner with a ceramic tube for natural gas without adapting it by any means. The results therefore do not show a problem using product gas but elaborate on the different ignition characteristics resulting from the gas composition. This merely shows the necessity to e.g. adapt the gas flow in the burner by design changes. Using a respective gas burner would alleviate these observations.
aHighest temperature measured at the exit of the burner.
bHighest temperature measured at the middle of the burner.
Table 8 shows the results obtained from the experiments using a burner chamber. The oxygen content in the table refers to the oxygen content of the combustion air. Thus, some of the combustion air was replaced with pure oxygen to achieve 23 and 25% oxygen in the combustion air, respectively. The measured wall temperatures (T1–T5) indicated that the combustion of product gas resulted in lower temperatures inside the burner chamber. Also, the ceiling temperature was approx. 20 °C lower in the case of product gas combustion in comparison to natural gas combustion. Adding 2% pure oxygen increased the temperatures to a level slightly beneath that of natural gas. Increasing the amount of pure oxygen from 2 to 4% led to a further temperature increase resulting in temperature levels above that of natural gas. Thus, using the identical setup for both the combustion of natural gas and product gas leads to slightly different temperatures inside the burner chamber, which can be equalised out by the addition of pure oxygen.