A Typical Case Study: Solid Waste Management in Petroleum Refineries

The current environmental concerns have forced developed and developing countries to reduce air, water and land pollution for sustainable growth. Solid refinery waste is cocktail of hydrocarbons, water, heavy metal and fine solids and is substantial in quantity. The principal processes of waste management focus mainly on waste source reduction, reusing, recycling, composting, incineration with or without energy recovery, fuel production and land filling. Waste management models have a common approach of assignment of generating sources to landfills, transfer stations sitting, site selection for landfills, etc. but recently new integrated models have been developed and applied. Waste management system in an industrial complexes uses multi-objective mixed integer programming approach for the running of existing facilities in dynamic net work flow models with nonlinear costs of management. The latest multi-objective mixed integer programming techniques are applied to resolve the potential conflict between environmental and economic goals and to evaluate sustainable strategies for waste management. In this approach, material recycling in an economic sense exhibits huge indirect benefits, although the emphasis of environmental quality as a major objective in decision-making drives the optimal solution toward pro-recycling programs. The use of grey and fuzzy system theories as uncertainty analysis tools as an enhancement of this modeling analysis proves to be highly profitable. A multi-objective optimization model based on the goal programming approach has been developed and tested in this study for passable management of solid waste generated by a typical petroleum refining industry in the state of Kuwait. The analytic hierarchy process, a decision-making approach, including qualitative and quantitative aspects of a problem, has been integrated in the model to prioritize the conflicting goals existing in the waste management problems of the petroleum industries. The optimization model is developed based on the goal programming technique that attempts to minimize the set of deviations from pre-specified multiple goals, which are considered simultaneously but are weighted according to their relative importance. A set of data from local petroleum refining industries is obtained and is analyzed using the unique presently developed multiobjective optimization model based on goal programming technique with all treatment options and economic constraints. The successful application of this model has provided the most economically viable solution to be applied for the specified objectives to be accomplished for the management of solid waste generated from petroleum industries.


INTRODUCTION
There is a four fold increase in natural calamities in last two decades reported in press due to increase in pollution levels that is major concern for the future sustainable growth for man and its habitat. The mammoth growth of petroleum products processing have resulted in the generation of enormous amount of waste that burdens the petroleum industries to resolve this burning issue of waste management. The level of natural resources reserves and increase in population with increasing living standards have reinforced the need to utilize the reminder in sustainable way. Thus refinery wastes are regarded as valuable asset (high energy potential) as far as resources are concerned and its management is of great importance. Fuel production, by-product processing, ancillary operations and waste management are the major operations in any oil refining industry. Fuel production encompasses those operations which manufacture petroleum products such as gasoline, polymers and coke. By-product processing covers refinery operations that convert used materials and/or undesirable petroleum constituents into saleable or reusable end products. Ancillary operations are those activities which support refinery functions and recover energy. Finally, waste Management deals with the recovery of useable materials from refinery waste streams, the disposal of solid and hazardous wastes and the treatment of wastewaters generated by refinery operations.
No matter today waste management is difficult and costly by the increasing volumes of waste produced, by the need to control potential serious environmental and health effects of disposal still it is the prime objective of industry for its sustainable operations. Many mathematical models have been developed to resolve the rising hazardous waste treatment problem by physical, chemical, thermal and biological processes. Additionally, mathematical programming techniques such as linear programming, dynamic programming and network models have been introduced to aid in managing the logistical aspects, such as finding the optimal location and size of facilities, of hazardous and non-hazardous wastes. In managing and planning the logistical aspects of hazardous waste systems, multiple goals, such as community and environmental control goals, those have different priorities have not to be ignored and have to be properly addressed.
In the late sixties in California, USA, an economic optimization for the system planning of solid waste management was first applied [1] . The issue of increasing environmental concerns and the emphasis on material recycling have gradually changed the focus of solid waste management in the following two decades. Recent research programs into solid waste management system planning frequently emphasize that both socioeconomic and environmental considerations have to be evaluated simultaneously to provide a set of total solutions regarding waste recycling, facilities sitting and systems operation.
The integrated models are based on simplified descriptions of the system and are subject to many limiting assumptions: weak disaggregation of material flows, a processing option of each type, sites dedicated to a particular processing or land filling technology, only one time period, recyclables/organics collections rarely taken into account, poor (or no) description of markets for recyclables, a single waste generating source, insufficient user's control on the accuracy of the investment cost functions. Gottinger [2] proposed a dynamic network flow model with nonlinear costs for waste management and facility-sitting decisions. Shekdar et al. [3] described a dynamic goal programming model for the management of existing facilities in a waste system. A multi-objective mixed integer programming approach was proposed by Caruso et al. [4] for the study of a regional system over a single time period. An interesting dynamic mixed integer programming model incorporating a large set of technologies and dealing with financial and air pollution constraints was presented by Chang et al. [5] . The preceding model has been transformed into a multiobjective one by Chang and Wang [6] . It takes four different criteria into account, three of them being environmental functions. Revenues from sales to markets are taken into account in the dynamic mixed integer programming model of Baetz and Neebe [7] . The model has a limited choice of technologies and only one new land filling site may be developed. A multiperiod and multi-regional model developed by Everett and Modak [8] has some interesting distinguishing features. Amongst them, there is the consideration of aggregated and disaggregated flows of materials and of a number of collection options for the components of the waste stream. The model does not deal with capacity addition. A very detailed static nonlinear programming model, MIMES/WASTE, has been proposed by Sundberg et al. [9] to address municipal and regional waste problems. The main objective of the model is cost minimization but emission control is integrated in the model via explicit restrictions and fees. Recycling and energy production goals may also be imposed. The model of Ljunggren [10] is an extension of MIMES/WASTE to national problems.
The approach of optimal waste minimization in a petroleum refinery was addressed by Takama et al. [11] . Their approach was to reuse and make use of regeneration opportunities. Wang and Smith [12] discussed the minimization of wastewater in the process industries. They pointed out that there are three possibilities for reducing wastewater, reuse, regeneration and regeneration recycling. Fletcher and Johnston [13] and Harries [14] described a waste auditing approach that involves a detailed analysis of a company's processes and wastes aimed at minimizing, a meliorating or even eliminating discharges from unit processes to establish waste management. Duke [15] indicated that waste minimization played a key role in US planning for hazardous waste management. He examined the effectiveness of waste minimization policies and regulations. Extensive pollution prevention programs in the industrial sectors have been adapted to minimize solid wastes generation [16][17][18][19][20] .
Waste minimization can be achieved by elimination of solid and hazardous waste generation through changes in product design and manufacturing technology [21] . Keen [22] addressed new regulations that require a waste minimization program to be in place.

Petroleum waste:
The present research is focused on the development and testing of a multi-objective planning model based on the goal programming approach for the proper treatment and disposal of solid wastes generated by Kuwaiti oil and petrochemical industries. All of the oil and petrochemical industries are located at the Shuaiba Industrial Area (SIA) in Kuwait. The SIA is located about 50 km south of Kuwait City. It accommodates most of the large-scale industries in Kuwait. The total area of the SIA (both eastern and western sectors) is about 23 km 2 . Fifteen plants are located in the eastern sector and 23 in the western sector, including two petrochemical companies, three refineries, two power plants, a melamine company, an industrial gas corporation, a paper products company and, two steam electricity generating stations, in addition to several other industries. Currently, approximately 70% of the total land area in the SIA's eastern sector is occupied by industrial facilities. Approximately 30% of the total land in the SIA's western sector is also occupied by industrial facilities.
Al-Shammari et al. [23] have given a detailed account of total solid waste generated from SIA in year 2006. There are three main categories; thermal, chemical and physical treatment for treatment of solid wastes from the petroleum refining industry. Each of these technologies has some advantages over the others while there are disadvantages associated with all of them. In any industry, to achieve the required criteria for disposal, all wastes must be treated to minimize the impact on the surrounding in order.
Thermal treatment unit operate at very high temperatures, usually 400-2000°C, to breakdown hazardous chemicals. These units are designed to handle specific type(s) of waste to be treated. The final stream could be a less toxic waste aqueous stream which could be further processed to separate the liquid phase from the solid phase. Thermal treatment units usually consist of two sections, the incinerator and the adsorber. The incinerator provides the thermal energy while the absorber removes the contaminants from the flue gas.
There are also non incineration alternatives for thermally treating hazardous wastes. These processes involve oxidation, reduction and/or pyrolysis environments to destroy the organic component of the waste matrix, but generate significantly less flue gases than incineration. Some of the industrially available technologies include: Rotary kiln oxidation, fluidized bed incarnation and Liquid injection incarnation.
Most widely used chemical treatment technology today is Stabilization. Stabilization is generally used to extract leachable metals prior to landfilling. In the refinery solid waste environment, streams that may require stabilization include: contaminated soils and incinerator ash.
Physical treatment technologies employ gravity separation techniques in order to separate the liquid phase from the solid phase in aqueous environments. Some of these processes are capable of capturing some of the fine solid that are in the mixtures.
Solidification can be accomplished by a chemical reaction between the waste and solidifying reagents or by mechanical processes. Contaminant migration is often restricted by decreasing the surface area exposed to leaching and/or by coating the wastes with lowpermeability materials. The technologies are not regarded as destructive techniques; rather, they eliminate or impede the mobility of contaminants.
General model: The model is based upon a general hierarchy of waste flowing a source to a thermal treatment plant or a chemical treatment plant or a third party. If a thermal unit is chosen, then the next tier is a chemical unit. Following a chemical unit is a physical processing unit. The hierarchy ends at a landfill which follows a physical treatment unit. This hierarchy is shown in the following Fig. 1  Based on available refinery data a model is developed and tested to minimize the transportation, processing, disposal and capital costs for the management of solid wastes produced from various facilities and having many processing and disposal routes. The objective function that must be minimized is composed of mainly four different sections. The first section is the transportation cost. The cost of transporting waste is given as dollars per mass unit of waste. This cost rate is multiplied by the total amount of waste that is transported would give the cost of transporting the waste. Transportation costs are incurred anytime there is a transfer of waste from one node to the other node. The second section of the objective function is the processing costs. Each facility that is operating will incur a processing cost. This cost is based upon utilities, man power and other operating costs. Disposal costs are the third section to the objective function. These costs are imposed when one is disposing of waste in a landfill. Third party costs are incurred when a decision is made to exercise a contractual agreement with a third party.
Lastly, capital costs are incurred when a new facility is opened. These costs are incurred only in the case of a new facility. Capital costs are based upon the facility type and capacity of the facility. The overall objective equation is given as follows.

Objective function:
ij ij ik ik jk jk km km ml ml jl jl j stp s6 s7 i A k scp i q j stpk scp k scpm spp m spp l sl l sl j s7 Transportation costs Minimize Processing costs Constraints: The design model is constrained on several parameters. Firstly, each node must satisfy a mass balance equation. This states that all the mass going into a node must equal an efficiency value multiplied by the output.
Material balance on thermal units There is also capacity limitation at each of the facilities which must be satisfied. Logic states that once the capacity of a facility is reached, then a decision must be made. First option is to open a new facility to handle the rest of the waste. The second decision is whether to neglect the first facility and just consider another facility.
As an extra limitation, there can only one type of each facility. This constraint was set to in order to achieve a non complex solution. The mathematical formulations of these constraints are as flows: At most build one thermal treatment unit, Typical case study: This study deals with refinery waste data where all available technologies and their processing and transportation costs and capital investment are considered. However, this model has an option to analyze fluctuating data that could occur in day to day operation.
The objective in this exercise is to determine the best method of treating these wastes in the most economical fashion. The model explores many possible combination of treatment technologies in order to achieve the required pacification to dispose of these materials.
The effectiveness of each treatment is a factor of the inherent capability of the treatment technology, the size and cost of the equipment. In the treatment hierarchy, twelve possible thermal treatment units have been defined. In addition, the model was asked to consider the possibility of allowing a third party to dispose of some of the waste. The model considered the following technologies: Continuing with the hierarchy, the model explored the best chemical treatment unit from a list that was provided. Similar to the thermal treatment units, the effectiveness of each treatment is a factor of the inherent capability of the treatment technology, the size and cost of the equipment. The following is the list of the considered chemical treatment technologies: • Organic extraction (OE1) • Solvent extraction (SE1) Physical treatment units were the following item on the hierarchy. The following are the list of the considered physical treatment units: • Encapsulation unit (EN1) • Stabilization unit (ST1) Lastly, the model was requested to consider a list of possible landfills and land treatment facilities. There were specific criteria for sending waste to either a landfill or a land treatment. This criterion was a factor of the waste stream and the amount of treatment it has received. The following are the list of the landfills and land treatment available facilities: The model was asked to determine the optimized route from the waste streams to the landfills/land treatment. The model was run using GAMS. The   FB2  FB3  FB4  MG1  F037  70  71  72  74  197  76  F038  76  79  77  84  177  78  K048  79  73  77  78  95  78  K049  72  75  79  74  79  76  K051  74  72  79  73  90  78  K052  72  77  70  80  81  71  K062  70  79  77  76  70  70   WO1  WO2  WO3  WO4  RK1  RK2  F037  78  79  87  92  197  182  F038  72  78  66  96  177  72  K048  77  69  86  62  87  69  K049  78  89  87  82  167  174  K051  71  68  81  61  188  181  K052  79  70  61  96  177  62  K062  73  81  86  80  182   The above data are used as input to the developed model and define the problem fully and clearly describe the structure of the refinery waste under study. The flow scheme for this integrated solid waste management system is shown schematically in Fig. 2.
The results of model solution provides the management with information about the extent of solid waste removal from the various sections. The model predicts the optimum route an provide a level of savings in financial resources allocated to run the transportation fleet and operates the solid waste treatment facilities, the extent of facilities utilization, energy production and level of recycling.

RESULTS AND DISCUSSION
In this case 23,866.5 t year −1 of various types of solid wastes have to be treated before it could be safely sent to landfill. The first step in treating all the solid wastes is the thermal treatment where various thermal treatment technologies are available to the model namely liquid injection, fluidized bed process, molten From various options the model calculates that fluidized bed process and wet oxidation process are most economical choices. In fluidized bed process FB1, FB3 and FB4 are available for use. The availability of FB2 is subjected to very high processing cost (718 $/ton) compared to other fluidized bed process. Among available four wet oxidation processes only WO4 is used as it has lowest processing cost of 119 $/ton compared to all other wet oxidation processes. However all the catalyst wastes are treated in high thermal treatment process (HTT) as the model has very restrictive option for treating the catalyst waste.
All the waste (3,351 t year −1 ) from F037 is treated in FB4 where as waste from F038 used four different thermal treatment units. It uses FB1, FB3 and WO4 to its full capacity and the rest is sent to FB4. Therefore, the amounts of wastes from K038 treated in FB1 are 795 t year −1 , FB3 is 295 t year −1 , WO4 is 5,400 t year −1 and FB4 is 2,996 t year −1 .
Waste from K048 uses both fluidized bed process and wet oxidation process. Nearly 1,422 t year −1 of waste is treated in WO4 and the rest 478 t year −1 is treated in FB4. All the waste from K049 (3,540 t year −1 ) is treated in WO4. For K052, it is more economical to use FB3 process compared to WO4, thus it uses FB3 to its full capacity (295 t year −1 ) and the rest of the solid waste (438 t year −1 ) is treated in WO4. Solid wastes from K052 is processed in FB1 and FB4. FB1 is used to its full capacity and processes 795 t year −1 of waste and the rest 105 t year −1 of waste is processed in FB4.
Total amount of waste fed to thermal processes was 23,866.5 t year −1 and is reduced to 17,040.3 t year −1 after thermal treatment. For further treatment the processed wastes are sent to chemical treatment section. The model has to choose between either solvent extraction or organic extraction process or both for processing 17,040.3 t year −1 . Processing cost of Solvent extraction process is 116 $/ton compared to the processing cost of 948 $/ton for organic extraction process. Although the capital cost and efficiency for organic extraction process is slightly favorable but processing cost plays a major role in deciding the usage of chemical treatment process and hence all the waste is processed in solvent extraction process. In this process 5,112.1 t year −1 of waste is rejected and hence 11,928.2 t year −1 of waste is send to physical treatment section.
Similar to chemical treatment process all the 11,928.2 t year −1 of waste is treated in encapsulation unit as it was found to be slightly more economical compared to stabilization unit. In this process 1,192.8 t year −1 of waste is rejected.
Finally the treated 10,735.4 t year −1 of waste coming from encapsulation unit and all the catalyst treated in high temperature thermal unit are send to landfill2 as its capacity is high enough to accommodate all the waste. The results of this model (Fig. 3) proves that it can be used to address many of the waste treatment and disposal problems and issues associated with the management of solid waste systems such as the need for solid waste removal from the various petrochemical plants, the efficient utilization of facilities, systems cost control and the control of environmental pollution.
The use of the model has been demonstrated, through this real problem, by showing how it can be utilized to assist in the management of solid waste generated by a petroleum industry. The results obtained show that the model is a viable tool and can be used to assist in making appropriate decisions regarding the management of solid waste.

CONCLUSION
The present model provides the most cost effective petroleum waste disposal details based on the choice of treatment processes, their capacities and appropriate routing of waste streams from a local oil refineries in Gulf State.
The multi-objective optimization model based on goal programming approach provides the efficient use of all available waste treatment units based on the control of environmental pollution and the most cost effective management strategies for the selected petroleum industrial complex.
The available data from typical petroleum refinery with all the accessible waste management facilities, the application of model provides the most efficient intermediate handling units, fluidized bed processes and wet oxidation process, which make the waste management highly cost effective. API separator waste can either be treated in solvent extraction unit or organic extraction unit but extraction unit is preferred based on space velocity/residence time. For a known waste quantity, the capacity and operation time is used to gauge the cost effectiveness for this chemical treatment process. In physical treatment process, the model provides the optimum use of the encapsulation unit with high cost stabilization unit to satisfy all the defined constraints in the industrial waste management exercise. The computed results are in concordance with other published refineries waste management results validating the application of the present model. This can be concluded that the present model is a viable tool and can be efficiently applied to solid waste from petroleum industries. The model output can be used to assist in making appropriate decisions regarding the petroleum industries solid waste management (minimization, processing, landfill etc.).