The oil and petroleum industry is among the most energy-intense industrial sectors. Energy consumption in this sector in 2012 in OECD countries represented more than 7% of the total from industry [1
]. The available literature states that a 10–20% energy consumption reduction is feasible in U.S. refineries with economically acceptable parameters [2
]; in Russia, the situation is similar [3
]. Heat recuperation intensification [4
], modern electromotor installations [6
], advanced process control implementation [6
], improvement of fuel gas management and fuel switching [5
], and the use of new conversion technologies [8
] are among the most promising measures to achieve the energy intensity reduction goal. An additional benefit, although less tangible than direct energy cost reduction, is an associated cut in the GHG emissions [9
] emitted from process heaters. Analogous measures can be applied to combined heat and power plants [11
], which often serve as steam suppliers for industrial facilities. The resulting benefits can be further enhanced if oxy-combustion is applied [14
]. The reduction of fire and explosion risk should also be taken into account as an additional reason for such an investment.
Direct fuel consumption and associated emissions, which contribute a major part of the total energy consumption in a refinery, are attributed to process heaters and steam boilers [17
]. Heavy fuel oils and refinery fuel gas represented the majority of fuels used in both cases [19
]. The scheduling of a fuel gas network represents a challenge for a refinery due to both its multiple sources and its variable composition [21
], as well as its variable production and demand [23
]. The flexibility to switch between various types of fuel consumed seems to be an attractive economic route to deal with this challenge [25
]. As discussed below in more detail, this flexibility may include natural gas or hydrogen-rich gases [26
], and can be exploited on a national or a global level [28
]. The associated decrease in emissions [29
] and possible change in the refinery electric energy balance if fuel gas is used for cogeneration purposes [30
] may represent additional drivers for fuel management improvement. Fuel and flare gas management system improvement has been introduced in various studies; producing reformatted [31
] or recovered material [33
] is another viable option for exploiting the material value instead of energetic refinery fuel gas content. Several studies investigated the possibility of replacing refinery fuel gas with hydrogen, thereby increasing the fuel system flexibility and decreasing the GHG emissions [35
]. These studies show that the fuel consumption by industry may be subject to single- or multipurpose optimization, and that it can decrease its environmental impact and still be economically beneficial.
Weydahl et al. [35
] explored the possibility of replacing fuel gas in a process furnace with hydrogen produced by steam methane reforming. Similarly, Lowe et al. [36
] investigated the technical details of fuel gas to hydrogen combustion in fired heaters at one of Chevron’s refineries, especially regarding the change in heat flux and the required burner adjustments. They concluded that it is technically feasible, but associated with high investment costs. Higher NOx
emissions are to be expected as well, due to a higher flame temperature. Lee et al. [26
] proposed and tested natural gas switching to hydrogen-rich tail gas in a fired heater located in a hydrotreater unit, which resulted in flame shortening, a NOx
emissions increase, and a slight thermal efficiency decrease. Fonseca et al. [19
] proposed changing the fuel in the central CHP unit in a refinery located in Portugal, coupled with gas turbine installation, and claimed to have improved the energy efficiency of the CHP unit and reduced SOx
emissions. Gabr et al. [27
] addressed the possibility of pollutant emissions decreasing in a hydrotreater unit by switching from fuel oil to natural gas in the fired heater. This experiment provided heat integration improvement and substantial financial savings, along with emission reduction. Rehfelt et al. [6
], in their bottom-up study, presented a systematic approach to suitable fuel selection in fired process heaters, including the technological details of particular processes. Fuel price, carbon dioxide emissions, and technological infrastructure are the key factors influencing the fuel choice. Cogeneration potential has not been studied, however.
The above literature survey indicates that significant attention has been paid either to process furnaces or to CHP units where fuel switching can be performed directly, e.g., fuel oil to natural gas, or fuel gas to hydrogen. Carbon dioxide emission decrease can be seen as the major driving force in fuel switching process studies. Situations where water steam acts as the heat carrier, enabling a simultaneous fuel consumption change in refinery furnaces and CHP units, require a more complex approach, which, as we learned from our literature survey, is missing from the current literature. This approach should include the impact on emissions, possible energy production changes, and the resulting project economics, following multiple optimization criteria.
The goal of our study is a complex assessment of the scope of process heat source switching, comprising all of the abovementioned factors. As documented in the relevant literature, such a broad approach to this problem and the resulting general methodology have not yet been presented.
The novelty of the developed methodology, as presented in this paper and tested via a suitable case study, is its systematic approach to fuel switching assessment. It therefore has the potential to convey relevant findings to both academics and industry. The employed mathematical model includes the material and heat balances of the process itself and its impact on the CHP unit operation, both from a fuel consumption and a cogeneration potential exploitation point of view. Seasonal CHP unit operation modes were taken into account to present an even more realistic assessment. The emissions changes, in refineries as well as outside them, caused by electricity balance changes were also included, following the guidelines set by valid EU legislation.
The aromatics fractionation process was chosen as a suitable case study as it is among the most energy-intensive processes in a conventional refinery [37
] due to the close boiling points of the individual fractions and compounds. A conventional aromatics extraction and fractionation process includes fractionation columns that may be reboiled, either by water steam condensers or by furnaces fired by refinery fuel gas or fuel oil. Usually, an industrial CHP unit [38
] is a marginal source of water steam consuming various types of fuel to produce high-pressure water steam for electricity and heat cogeneration in steam turbines. Introducing water steam as a heat carrier can, among other benefits, help to reduce the fire and explosion risk in the refinery.
In this paper, the problem’s superstructure is defined first, followed by a general mathematical model covering all relevant aspects addressed in the introduction. The selected aromatics plant is then presented in a more detailed manner, including further model equations and assumptions along with the available process data. Because of industrial data confidentiality, the raw data are not provided. Calculation and sensitivity analyses’ results are discussed, and the relevant findings summed up, in the conclusions section.
3. Results and Discussion
3.1. Model Verification
Process heat duty calculation model verification (Equations (1)–(11)) is presented in Figure 6
. As can be seen, the process heat duties calculated from the process and furnace side correlate well, with only a few exceptions. A discussion with plant operators revealed that major discrepancies correlate with process flowmeters’ (reflux flow) calibration. A better match can be seen during higher process heat load periods. The highest observed heat loads were multiplied by a factor of 1.1 to 1.3 after discussions with plant engineers to provide some margin for possible future plant intensification. The adjusted maximal heat loads gave indications for new reboilers’ sizing.
Verified process heat duties were used to estimate HPS consumption in individual reboilers. An average total steam consumption of 33 t/h was used in the following economic calculations, which corresponds to the situation in the first months in 2017 and is expected to occur much more often in the future. A peak HPS consumption of 38 t/h or over can be anticipated to occur from time to time, and consequently the steam pipeline capacity has to be revised. HPS is delivered to the considered production units via a steam pipeline with an internal diameter of 150 mm, which is insufficient for the transport of even the average HPS flow to new reboilers. A new pipeline with an internal diameter of 200 mm has to be installed to ensure HPS delivery to new reboilers without excessive pressure loss; the approximate pipeline length is 100 m. The costs associated with the extra HPS pipeline are included in the TIC calculation, discussed below.
3.2. Investment Scope and Cost
details the TIC estimation procedure for scenario B. The resulting TIC estimate of over €4 million indicates that to reach a payback period appropriate for the refinery (less than four years), the yearly benefit has to exceed €1.1 million/year. The Lang factor value for “Piping installed” of 80% was higher than the recommended value of 68% in [48
] to cover the extra costs associated with the new HPS pipeline instalment.
The TIC estimation for other scenarios proceeded similarly and the results are presented in Table 10
. Scenarios A and C produced an identical TIC, lower than in B due to the absence of condensate coolers. Scenario D requires a higher TIC than B as it requires a new condensates return pipeline to the CHP unit with a length of several hundred meters. Its cost was estimated to be €450,000 after plant engineers considered the past investment costs of similar projects.
The installation of new steam reboilers and shutting down old, inefficient furnaces reduced the future expenses associated with equipment maintenance and with the necessary adjustment of furnaces and burners to tighter emission limits. Another, though even less tangible, benefit is the contribution of this investment to a decreased fire and explosion risk in an aromatics production plant.
3.3. Emissions Balances
For scenario A the calculated annual pollutant balance in the refinery is shown in Table 11
. As can be seen, changes in all other pollutant emissions, except for CO2
, are fairly low, which seems surprising considering that refinery fuel gas is much “cleaner” fuel than the heavy fuel oil combusted in the CHP unit boilers. This accentuates the need to evaluate emissions from fuel combustion not just using the “clean–dirty” fuel concept, but also by taking into account the operation of individual thermal aggregates where fuel is combusted.
The following facts elucidate the observed small changes in pollutant emissions:
Furnaces F1 to F3 are of advanced age, including the RFG burners installed within. Their renovation is necessary to meet stricter emission limits.
Compared to CHP unit steam boilers, no flue gas cleaning is installed in the common flue gas ducts from F1, F2, and F3. CHP unit boilers are equipped with dedusting, deSOx, and deNOx systems, which significantly reduce pollutant emissions from the CHP unit.
Carbon dioxide emissions, on the other hand, increase significantly, probably due to the high specific HFO consumption per t of exported steam, as no steam condensates are returned to the CHP unit in scenario A. Lower CO2 emissions can be expected in other scenarios.
compares the pollutant emissions from individual scenarios and elucidates the influence of incorporating emissions from power generation in emissions estimation. Emissions from power generation were estimated using the specific emission factors listed in Table 8
. As expected, the CO2
emissions increase is lower in all other scenarios compared to scenario A. Scenarios B and D do not consume that much HFO in the CHP unit compared to scenario A and produce a comparable electric energy amount in the CHP unit, e.g., the refinery itself produces fewer emissions. Substantially more electric energy is cogenerated in the CHP unit in scenario C, which reduces CO2
emissions from the CHP unit in the summer operation regime (see Equation (29)). A CO2
emissions decrease was achieved outside the refinery in the CHP winter operation regime. Similarly, the SO2
emissions increase in scenario C was cut in half compared to scenario A for the same reason. This effect is less pronounced in other pollutants, but it still exists.
Applying marginal emission factors instead of the average ones from the chosen energy mix significantly improved the pollutant emission amount estimate accuracy [62
]. Slovenské elektrárne, a.s. operates several hydropower and nuclear power plants, which contribute to the low energy mix emission factors (Table 8
). Such power plants, however, cannot be considered traditional marginal power sources, e.g., power sources with flexible operation. This function can be met, for instance, by a flexible NG-based combined cycle power plant. Assuming a net electric efficiency of 50%, its CO2
emission factor is over 400 kgCO2
/MWh, i.e., it triples the value listed in Table 8
. Coal- or oil-fired power is a marginal power source with an even higher CO2
emission factor: up to 900 kgCO2
] and generally higher emission factors of other pollutants compared to those used in this study. Table 13
shows the impact of an emissions factor considered marginal for power production on the total CO2
balance in individual scenarios.
The results presented in Table 13
highlight the significance of correct power production emission factor selection: annual CO2
emissions decrease by up to 45% (scenario C) when using the coal power plant factor instead of that of Slovenské elektrárne, a.s. In other scenarios, the effect is less pronounced, but the difference can still exceed 15% (scenarios A and B). Scenario D is the least sensitive to the power production emission factor selection as it leads to the lowest power production increase in the CHP unit compared to the present state. Similar results can also be expected for other pollutants’ balances. This stresses the need for emissions from power production to be included in the total emissions balance, together with the proper use of the power production emissions factor.
3.4. Economics and Sensitivity Analysis
An economic assessment of individual scenarios is provided in Figure 7
and Table 14
. Annual benefit changes with heat in steam price to heat in RFG ratio (herein referred to as the Ratio), as well as with electric energy cost, are investigated in Figure 7
. These two factors mostly influence the economic attractiveness of individual scenarios. The annual benefit of scenario A is negative in all cases with a Ratio value of 90% or over, whereas all other scenarios provide a positive annual benefit under these conditions. This effect can be attributed to the decreased HFO consumption compared to scenario A (scenarios B and D) or to the much higher power production in the CHP unit compared to scenario A (scenario C). Therefore, scenarios B and D are much less sensitive to electric energy price changes than scenario C. Of all the scenarios considered, D is the most robust one, leading to positive annual benefits in all cases.
The influence of other factors’ changes on the annual benefit is shown in Table 14
. The cost of carbon dioxide emissions can be identified as the third most important factor. Its increase from €10 to €20/t reduces the annual benefit in all scenarios; A and C are the most sensitive to this change.
The annual benefit of scenario A reaches a negative value under these conditions and the simple payback periods of scenarios B and C increase to four years or more. Only scenario D is sufficiently attractive, even with a carbon dioxide emissions price of up to €30/t, leading to an annual benefit of around €1.2 million/year and to the simple payback period being slightly longer than four years. The sensitivity of all scenarios to carbon dioxide emissions’ price increase has to be accounted for, especially since refinery managers expect emissions prices to increase to €30/t or above in the near future.
As can further be seen, the RFG cost increase and chemically treated water cost decrease are beneficial in all scenarios, with scenario D being the most sensitive to RFG cost change and scenarios A and C to most sensitive CHTW cost change. RFG cost is an important factor influencing the economics of the scenarios; RFG cost changes of ±10% are quite common within the time span of a few months. CHTW cost changes are less frequent and less pronounced, and therefore this factor can be considered the least important.
Examination of other pollutant costs’ variation in terms of the annual benefit of the considered scenarios seems to be meaningless. Their total increase is below 100 t/year (scenario A), which means that even doubling their cost (from €66 to €132/t) decreases the annual benefit by less than €10,000/year.
The complex framework of process heat source switching in refinery conditions developed in this study allows for its objective assessment from energy, environmental, and economic points of view. The effect on other refinery parts as well as on emissions from the power production process outside the refinery was also taken into account.
The method of process heat source switching was tested on an aromatics production plant, employing higher boiling hydrocarbons fractionation in columns reboiled by refinery fuel gas-fired furnaces. Reboiling by condensing the high-pressure steam produced in the CHP unit in heavy fuel oil-fired steam boilers was proposed. The scenarios considered reflect the possibilities for improved heat conservation in the plant itself, condensates return to the CHP unit, as well as possible future high-pressure steam throttling mitigation.
The annual benefits resulting from individual scenarios exhibit various sensitivity to the key economic parameters change, which further stresses the need for complex method application in the process of heat source switching. Apart from fuel and steam costs, which are the most important economic parameters, energy costs and carbon dioxide emission costs were identified as other parameters to which the annual benefit is sensitive. The most robust scenario incorporates both heat conservation and condensates return to the CHP unit, and, despite it having the highest estimated total investment cost (of over €4.5 million), it offers acceptable, simple payback periods in most energy and media costs combination cases.
An environmental evaluation revealed that taking into account the emissions generated in the power production process outside the refinery can substantially change the estimate of the total generated emissions. Apart from CO2, the emissions of other pollutants did not increase as significantly as expected when switching from refinery fuel gas to heavy fuel oil. The reason is that refinery fuel gas-fired furnaces are small and medium-sized thermal aggregates with no flue gas cleaning installed, while the CHP unit steam boilers are equipped with multistage flue gas cleaning. The operational conditions and technical state of thermal aggregates are thus as important as the fuel type they consume in a complex emissions generation assessment. These findings are relevant and should result in more complex emissions generation evaluation in industrial process heat source switching projects.
It can be concluded that complex multi-objective evaluation is necessary for industrial heat source switching projects and the use of “cleaner” fuels is not the only goal to be pursued. The presented method is a suitable tool for such evaluations and can thus be applied in the industry generally to aid engineers, energy managers, and industrial policy makers with their decision-making.