Table I Specific Gravities of Radioactive Minerals Found in
Contaminated Soils
William S. Richardson, III
Auburn University at
Montgomery and S. Cohen & Associates, Inc.
Charles R. Phillips
S. Cohen & Associates, Inc.
Gerald H. Luttrell
Virginia Polytechnic Institute and S.
Cohen & Associates, Inc.
Clinton Cox
U.S. Environmental Protection Agency
ABSTRACT
Mineral processing technology has been in use for decades and can be applied to separation of radionuclide contaminants from their host soil matrix by exploiting differences in physical properties between them or these two constituents. A number of studies have been performed to assess the applicability and effectiveness of mineral processing technology for treatment of radionuclide-contaminated soils. Many of these studies have produced inconclusive results because of their incomplete scope and a lack of understanding of radionuclide contamination and its association with the host soil. Some of the past studies were limited by a lack of commitment of funds to develop the proper application of characterization and treatability studies to mineral processing and others by a lack of understanding of the necessary requisites for a definitive study. In order to select potential treatment process(es), a complete characterization study of both host soil and contaminant materials must be performed to determine their physical and chemical properties as well as their relative associations with each other. Full characterization must be followed by treatment studies to evaluate the performance of process units and to optimize separation by computer modeling. Process modeling can save time and is critical to the successful selection and design of a treatment process train. Most characterization and treatability studies performed to date on radionuclide-contaminated sites lack this completeness and applicability and, therefore, have not resulted in a successful process design.
This paper describes the approach to the design and conduct of characterization and treatment studies and their relation to the conceptual and engineering design phase of the treatment process. Four important themes run throughout the paper: 1) the tiered approach to process selection, 2) the application of proven physical processes for treatment, 3) the importance of economic and practical comparison to other potential approaches to remediation, and 4) the importance of performing a complete, integrated, iterative, and reliable study of the contaminant and soil and the separation processes that are suitable for treatment and remediation.
INTRODUCTION
The primary goal of soil remediation technology is to provide a treatment process to reduce the toxicity, mobility, volume, and/or concentration of radionuclides in a contaminated soil. The level of complexity of the remediation technology required to accomplish these goals depends on: 1) the type of radionuclide contaminants present, their mineral nature and their host-soil binding characteristics, 2) the type and complexity of the host soil matrix, and 3) the process(es) required for reduction of toxicity.
Application of soil washing to the treatment of radionuclide-contaminated soils exploits the differences between contaminant particles and host soil particles to provide a method for separation, concentration, and collection of these components by mineral processing technologies. These processes, which have for decades been the foundation of the mining industry, offer the possibility of separating virtually any soil component. Its successful application to site remediation, however, depends on development of a treatment process that provides a solution that is acceptable to stakeholders and regulators and one that is economically competitive with alternative remediation methods, including excavation and disposal. Development of the process requires, in turn, a complete understanding of the physical and chemical properties of the soil and contaminant constituents and the selection of process units that have been fully optimized to provide the most effective isolation of the contaminants. Attempts over the last ten years to apply soil washing technology to radionuclide-contaminated sites have been characterized, for the most part, by the limited application of these principles, primarily because of a lack of understanding of the necessary requisites for a successful study and a lack of commitment of funds and time to develop the proper approach to soil washing.
A reasonable, cost-effective application of soil treatment to site remediation employs a tiered approach. The step-wise protocol attempts to limit the study only to those tests necessary to develop a treatment plant or, alternatively, to terminate the study if tests reveal that soil treatment is not a competitive remediation alternative. The importance of performing a complete, integrated, and reliable study, however, cannot be over emphasized. The tiered approach applies a protocol that permits the study to follow promising results, but it does not imply compromising the study by bypassing or attenuating tests needed to obtain the data required to make the best decisions about the conduct of the study. Neither does it suggest that a more complex technology or a combination of technologies might not be successful solely because a more simplistic method, particle-sizing for example, has been rejected earlier in the study. Treatment decisions are greatly affected by the success of the separation processes and their ability to compete with alternate remediation methods. Unless complete reliable data is available to make these decisions, treatment processes might be prematurely eliminated from consideration or, alternately, given too much emphasis as a promising remedy. Both errors are very costly, time consuming, and negative to the development and use of treatment technologies that in other circumstances might make important contributions to remediation efforts at other sites.
Studies to evaluate the potential application of mineral processing technology begin with a characterization study to determine the basic nature of the contaminant and host soil particles that, in turn, are used to propose a plausible remediation scheme. The proposed scheme will exploit differences in the properties of particles to accomplish a separation process(es). Generally, the larger the property difference, the easier the separation. Characterization identifies the particle size, density, magnetic susceptibility, hardness, solubility, and similar properties of all soil constituents as well as the nature of their associations with each other. If differences in particle characteristics are discovered that can be reasonably exploited to achieve separation, a treatability study is then performed to select the most efficient mineral processing units for a treatment method and to design a competitive treatment train that will accomplish the separation and meet the cleanup requirements.
In the past, considerable confusion has existed about the difference between characterization studies and treatability studies. In general, characterization studies are directed at determining the attributes of the soil components while treatability studies are aimed at identifying processes for separating the contaminant from the host soil. As such, the characterization data are independent of the process(es) used to achieve the separation whereas treatability data are not. For example, information obtained form laboratory sieving tests are appropriately termed characterization data since they depend only on the physical properties of the soil. Treatability data can be obtained using a variety of particle-size separators such as mechanical screens or hydroclassifiers. The treatability data are different for each separator since the efficiency of each unit varies with the mechanical design and operating principle of each sizing device. In some instances, characterization studies, incomplete ventures by themselves, have been labeled treatability studies. Independently, characterization studies might indicate that soil washing is not a promising remediation alternative for a site, but they will not, alone, provide sufficient information to decide if soil washing is a competitive solution. Treatability studies are the next step in the tiered approach that is required to support that decision; and if promising results are achieved, process and engineering design with a thorough economic and regulatory evaluation are necessary.
In treatability studies, individual units as well as the complete process train are evaluated using computer process simulations in conjunction with bench-scale processing units whose operations are optimized by laboratory efficiency studies. These tests are complementary, providing guidance to each other for adjusting operational parameters in an iterative process to obtain optimal separation and to convert the ideal characterization data into actual performance data. Computer simulation also permits the examination of a wide variety of processing alternatives and process-train configurations to provide a cost-effective approach without unnecessary, more costly laboratory tests.
After the treatability study, a conceptual process design from the treatment data is prepared using the most efficient process units selected by evaluation of data from the treatability study. The design is optimized in another similar iterative study using computer modeling with uncertainties resolved by bench-scale and pilot-scale testing. A final flowsheet design is prepared from these studies, the plant is sized for the site, and an economic evaluation is performed and used for design optimization. The last step is engineering design, the construction, testing, and optimization of the semiworks plant, a full-scale pilot plant, followed by construction of the final operational plant to be used in the overall remediation plan for site cleanup.
CHARACTERIZATION STUDIES
Characterization has three purposes: 1) to identify the mineral constituents of the host-soil matrix and its contaminants and define their intrinsic properties, 2) to identify the relative association of these components, and 3) to provide data to estimate the potential for separating the contaminants using an ideal mineral separator that is perfectly efficient. Intrinsic properties include particle size, density, magnetic susceptibility, hardness, and solubility; they depend solely on the inherent properties of the soil constituents and are independent of any process that might be used eventually to achieve the desired separation. It is important to understand, however, that it is never possible to achieve this ideal separation because of natural inefficiencies in real separation processes. Characterization data are collected, then, to provide a plausible remediation scheme and to estimate the ideal remediation potential that can be expected for the given separation process(es) and the potential for realizing the cleanup criteria established for a site.
The nature of the host soil and the manner in which radionuclide contaminants are associated with this host material will ultimately determine the applicability and dictate the type of treatment technology that might be appropriate for remediation. Contaminants that are well liberated and uniquely distinguishable in terms of one or more physical properties (e.g., density, size, etc.) are relatively easy to separate in a single processing operation. In contrast, contaminants that are intimately associated with the host soil or are virtually identical in terms of their physical properties require the complete solubilization of the soil matrix to separate. Metallic radionuclides are associated through absorption, ion exchange, precipitation, coprecipitation, ligand and chelate exchange, and occlusion (1). Soil minerals that are significant to radionuclide contamination by these processes include clay minerals, hydrous oxides, carbonates, and humic substances (2). These association types and soil material categories combine to produce a wide spectrum of soil-contaminant interactions.
The nature of the host soil and contaminants and their distributions are generally different for each site and within a given site. It is critical to characterization and treatability studies, therefore, that the site itself is properly characterized and the samples used in the study are representative of the site. The cleanup criteria must be also firmly established before the study begins. Site characterization should include a sampling plan that permits collection and analysis of a sufficient quantity of representative samples. It is often necessary to excavate trenches to obtain samples that are representative of the vertical radionuclide distributions. This process also allows for collection of samples of sufficient volume to perform remediation studies. Complete characterization and treatability studies performed on samples that adequately represent the material to be treated provide a realistic estimate of the magnitude and extent of contamination and suggest approaches to soil removal and treatment. Incomplete or erroneous information at this stage or absence of well-defined cleanup criteria will be detrimental to the entire remediation study and lead to very costly errors in evaluation, planning, and remediation attempts. Based on the cleanup criteria, site characterization, and subsequent characterization and treatability studies, soil from the site might be divided, ultimately, into three types: 1) soil that meets the cleanup criteria as it exists and could remain on the site, 2) soil that cannot be cleaned up by any treatment method and will require disposal, and 3) soil that can be treated by a separation technology to meet the criteria. The relative amount of soil in each of these categories will greatly influence the options and costs of site remediation.
Characterization studies begin by wet sieving the samples to determine a particle-size and contaminant distribution profile. Soil particles are typically sized, from larger to smaller particles, into five categories: cobbles, gravel, sand, silt, and clay. The particle-size and radionuclide distributions affect the choice and effectiveness of treatment technology. If the contaminants are in a specific size fraction, isolation of that size fraction from noncontaminated material can provide a simple treatment solution. Likewise, removal of the contaminated size fraction in a preliminary treatment step will reduce the amount of soil requiring more extensive treatment. The size distribution within the overall soil material also influences the type of treatment process that might be applicable. For example, clay material often forms an aggregate clump in the presence of water, and soils with large amounts of clay-sized material might be difficult to treat by some mineral processing units. In some cases, a chemical dispersant (such as sodium silicate) might be necessary to improve the dispersion of clay. Soils with contaminants in larger-sized sand fractions, alternatively, are usually easier to process than those with large amounts of very small clay-sized particles (provided that the contaminants are not locked together with uncontaminated particles in the host soil). Soil classification is also indicative of certain mineral compositions that absorb radionuclides. Clay particles contain silicate minerals that carry a negative charge on their surface that selectively attracts radionuclide cations. Combined with the large surface-to-weight ratio, clay material has a high, strong absorption capacity for radionuclide contaminants.
Characterization continues with a preliminary examination of the mineral nature of the soil constituents, usually by scanning electron microscopy (SEM) and often complemented by traditional petrographic examination by polarized-light and stereomicroscopy. Preliminary data provide an indication of the specific nature of the host soil and contaminants that is valuable in planning the more detailed study.
Preliminary examination is followed by a quantitative study of the mineral content of the soil and the nature of contaminant association. Data are collected by scanning electron microscopy (SEM) and energy-dispersive X-ray analysis (EDX) on particle-size fractions produced by the sieving study. The data are used to determine if the contaminants are sufficiently unassociated to be separated by physical separation processes. In each mineral analysis, radionuclide content from radioanalysis is correlated with mineral type and concentration to help identify the contaminant soil material. Discrete minerals such as monazite, for example, contain specific radionuclides (thorium in this case). Alternately, radionuclides might be occluded in a process mineral or absorbed onto the surface of clay material. Cesium, for example, often interchanges with natural alkali metal ions to become an intimate part of the soil matrix. In addition to these tests, a variety of other characterization tests might also be necessary to fully delineate the radioactive components present in each of the size fractions contained in the contaminated soil. X-ray fluorescence (XRF) measurements can be performed to aid in the mineralogical identification. X-ray diffraction (XRD) might also be used to identify difficult-to-distinguish components in the various soil samples.
Sequential leaching by select chemical reagents can help identify the fixation mechanism of contaminants to the host soil matrix (2, 3). Reagents are chosen in order to selectively destroy specific binding minerals in the matrix. A soil sample is treated first with milder reagents followed by increasingly aggressive reagents that attack the host soil (2, 4, 5, 6, 7). Eventually, the matrix is completely solubilized. At each step, radionuclide analysis reveals which contaminant is released as the matrix is attacked, in turn, indicating which host-soil component is binding the contaminant.
In additional to sieving, other separation procedures are often important to characterization studies. Radioactive minerals typically have specific gravities greater than soil materials. Table I provides a list of the specific gravities of radioactive minerals that might be found in contaminated soils. For comparison, the specific gravity of common quartz sand is typically 2.65. With the exception of amorphous silica, the specific gravities of the radiominerals listed are all substantially higher than most of the materials found in the host soil. Therefore, radioactive minerals might be amenable to density-based separations. The only particles that might be lost using this process are those of the relatively small weight-fraction of noncontaminated heavy-minerals, which will also separate with the contaminated particles. The magnetic susceptibility of several radiominerals also suggests a physical separation option for volume reduction by contaminant removal.
The magnetic susceptibility of monazite and zircon (thorium bearing minerals) (8), for example, might permit their separation from host soil material. As part of the characterization study, density and magnetic separations are often performed on select particle-size fractions. The separated fractions are examined for mineral and radionuclide content in order to correlate contamination to a fraction or mineral having a density or magnetic susceptibility that might be exploited to achieve a separation from noncontaminated or less-contaminated host soil.
Table I Specific Gravities of Radioactive Minerals Found in
Contaminated Soils
Overall, characterization studies provide information on the nature of the contamination and the host material and the distribution and association of the contamination within the soil matrix. This information provides an indication of the feasibility of physical-based processes as treatment methods and indicates the specific technologies that might be employed for treatment. Treatability studies will thus be directed toward those methods that are applicable to radionuclide removal from contaminated soil matrices.
TREATABILITY STUDIES
Treatability studies are designed to select the most effective process(es) for separating contaminated materials from soil based on the properties of the soil components as determined in the characterization studies. With these properties, process strategies, drafted as flowsheet designs, are proposed using treatment units in combinations that can exploit material differences to produce a separation. The processes are limited to those methods that are specifically applicable to contaminated soil matrices at the site. Among the potential properties that might be exploited by treatment technologies are particle size, density, magnetic or electrostatic properties, surface (flotation) properties, and absorbance. Chemical treatment might be used to solubilize the contaminants and collect them by precipitation, coprecipitation or flocculation, and/or ion exchange.
The treatability study is an iterative process applied to develop the most effective and economical remediation design and optimize its performance. The iterative study makes use of both computer simulation and bench-scale analysis of probable units and unit combinations in a complementary relationship to save time, money, and materials. The preferred design must be cost-effective, meet or exceed the stated cleanup criteria, and maximize the overall volume reduction of contaminated soil. To achieve these goals, it is important to eliminate unnecessary and costly tests by using computer-model simulations on the possible combinations of process units for the development of a flowsheet design. Standard partition-curve data provided by the manufacturers of each unit operation are used in the flowsheet development. Partition curves are used to convert characterization results into actual performance data that may be expected for a given separator. This approach ensures that the design flowsheet reflects "real-world" values for selecting and evaluating process units. The simulations also identify acceptable performance levels as well as problems for each process unit in the flowsheet. On identification of any significant deficiencies, simulation permits the development of alternative flowsheets and, in turn, identifies the need for additional characterization or bench-scale tests. Once an optimum flowsheet design is selected, appropriate bench-scale testing is performed on the unit operations at the parameter settings determined from the computer simulations. Using the bench-scale testing in this manner to fine tune the unit operations, instead of optimizing the full-scale technology, is less costly and time consuming.
Treatability studies begin by a review of the data provided in the characterization study. Any shortcomings in the data are identified, and additional tests are performed, or repeated. Data evaluation is followed by a review of the technical literature and databases related to radionuclide-contaminated soils and potential treatment technologies. A treatment plan is then formulated that includes: 1) identification and evaluation of those separation technologies with the potential to provide remedies for the site based on the initial characterization data; 2) verification of cleanup criteria for the site; 3) identification of any additional characterization data that might be necessary to complete the treatability study; 4) selection of criteria to be used to assess the potential performance of the treatment technology (these criteria include contaminant concentration and cleanup-level potential, volume reduction expectations, economic consideration, production of secondary waste, technology performance expectations, implementability of the technology process and plant and unit operational expectations); 5) formulation of an initial process flow diagram; and 6) assessment of the potential for treatment based on the characterization data and the information received from the equipment manufacturer and/or literature search. With this treatability plan, reasonable technologies are evaluated. Those that meet the selection criteria are assembled in an initial flowsheet design and reviewed before the optimization and testing of the technology begins.
To maximize the usefulness of the characterization data, a computer simulation program is developed using existing mathematical models to predict the performance of the conceptual flow diagram. The models are derived from the partition data available in the technical literature or provided by process technology vendors. Computer simulation programs require input of characterization data and partition data for each unit operation in the flowsheet. From these inputs, the computer simulation predicts the separation of each unit operation and provides the amount of material passed and retained, activities of both streams, "cut-size" of each component, amount of misplaced material, and overall separation efficiency (volume reduction and total activity) for the flowsheet design. After the computer simulation program is developed for the conceptual flow diagram, the optimization of that flow diagram, leading to a preferred process flowsheet, begins. The use of simulation programs in flowsheet design reduces the need for labor-intensive and costly laboratory and bench-scale testing. This process allows the minimization of cost and time without diminishing the effectiveness of the final design.
The first computer simulation is performed using a best-case scenario. It produces process results assuming that a perfect (100 percent) efficient separation is performed at each unit operation in the flowsheet. This simulation is very important because it identifies the maximum attainable performance for each unit operation as well as the overall flowsheet process. The best-case simulation results are evaluated using the site cleanup criteria to determine if it is technically feasible. If it is not technically feasible, then a redesign of the flowsheet is required. During the optimization process, it is known that each unit operation will not perform at a perfect (100 percent) efficient separation. However, this information allows variation of the operating parameters of each unit to predict the optimum separation of the flowsheet design. Although the simulations require significant input data, they can be accomplished in a short amount of time.
The computer simulation programs can evaluate the flowsheet as a whole or for individual process streams. With this capability, problem areas within the flowsheet can be addressed independently. For example, a proposed process might be providing an overall product that meets the cleanup criteria, but a specific size-fraction alone could be above the cleanup criteria. The ability to use computer simulations to identify this possibility allows exploration of additional characteristics of the fraction that might be exploited to reduce the contamination and, in turn, increase the volume of remediated soil. All these evaluations can be made without the use of timely, labor-intensive, and costly laboratory and bench-scale testing. Once all the probable combinations of the selected technologies are evaluated in the flowsheet, a final or preferred flowsheet design is selected. The preferred flowsheet might actually be several flowsheet designs that produce comparable volume reductions. It is then necessary to evaluate each one to determine which is the most economical and feasible design for the site. If no plausible scheme can be identified, then it might be necessary to conduct additional characterization tests to identify one or more additional basic properties that might be exploited.
An example of a computer simulation for the flowsheet shown in Fig. 1 is presented in Tables II and III. The flowsheet design is based on characterization data provided in a report on radionuclide contamination at a remediation site under study (9). From Table II, the weight (mass) reduction of a perfect separation of the flowsheet is shown as 60.3 percent, with activity levels of 0.84 pCi/g U-238, 1.93 pCi/g Ra-226, and 1.10 pCi/g Th-232 in the cleaned product. (The densities of feed material and product streams are surprisingly similar since the soil particles are more loosely packed in product streams than they are in natural soil. Weight and volume percentages agree, therefore, within a several percentage points, and the term "volume reduction" is often used for weight reduction in the discipline. Caution in using these terms is necessary, since the difference can be important, especially in economic comparisons.) From Table III, the weight reduction of an optimum separation of the flowsheet is shown as 55.7 percent, with activity levels of 1.04 pCi/g U-238, 2.26 pCi/g Ra-226, and 1.81 pCi/g Th-232. This flowsheet is capable, with perfect separations at each unit operation, of achieving a 60.3 percent weight reduction below the cleanup criteria. However, based on operating parameters and misplaced material, the actual weight reduction of this flowsheet is 55.7 percent. The overall recovery of "clean" soil is 92.4 percent for this flowsheet design (actual weight/perfect weight reduction). These results represent a very efficient flowsheet design that would be a strong candidate for a treatment process. Economic consideration based on past experience and the literature (10), indicates that this level of weight reduction is extremely competitive with alternate remediation processes such as disposal alone.
Fig. 1. 5 TPH soil treatment process.
Table II Computer Model of a Perfect Separation - Overall Flowsheet
Table III Computer Model for an Optimum Separation - Overall
Flowsheet
After the final flowsheet designs are selected, necessary bench-scale tests are identified. Bench-scale tests are necessary to verify the computer simulations and to allow fine-tuning of the operating parameters of each unit in the flowsheet and demonstrate their reproducibility. Because of cost considerations, these tests are typically performed at a small scale (<500 lb/hr) using continuous or semi-continuous pilot-plant separators. If possible, these units should be similar to those that will be employed in a full-scale plant. This bench-scale test data provides the first opportunity to evaluate an actual operating circuit. The bench-scale tests are also often used to collect improvised partition data for various unit operations. Additional process simulations are often performed in conjunction with the bench-scale tests in the iterative design and testing process. Although bench-scale tests are often labor-intensive and costly, the test work can be minimized through use of the computer simulations. The laboratory bench-scale testing will determine the operating parameters for each selected separation technology and will demonstrate the ability to consistently reproduce the indicated results. Reproducibility of bench-scale test results is an important consideration in treatability studies, because it greatly affects the reliability of the overall process design.
Another useful function of the SEM method described above is the characterization it can provide of the treated products obtained from the various separation schemes during the treatability studies. The system can produce data on the types and relative amounts of each of the various components in the process streams. For example, soil treated by flotation can be examined to determine whether a contaminated mineral such as radiobarite is responding better to a density-based separation than its parent ore, carnotite. Measurements of this type might be performed on the products obtained from at least one of each of the separation processes in the proposed work. Additional measurements can be conducted as needed to resolve separation problems that might develop during the testing program.
After the bench-scale tests are completed, the results are applied to the flowsheet design(s). This fine-tuned flowsheet constitutes the final flowsheet design(s). Like the final flowsheet design from the computer simulations, it might consist of several designs that produce equivalent volume reductions. Each one will require an economic evaluation to determine which is the preferred conceptual process design.
CONCEPTUAL PROCESS DESIGN
If the treatability study provides a soil treatment strategy that is competitive with alternate remediation proposals, the remediation study continues with a conceptual process design study. This step provides a final conceptual engineering design for the treatment plant. The preferred design must be cost-effective, meet or exceed the stated cleanup criteria, and maximize the overall volume reduction of contaminated soil. In addition, it must be satisfactory for on-site treatment operations, meet operational health and safety standards, and produce acceptable waste stream products. To achieve these goals, comprehensive engineering analysis and a feasibility study is conducted to identify the preferred process design that can be readily implemented for soil treatment and remediation. This study includes process flow diagrams, sizing of process equipment, a complete cost analysis, preliminary material balances, a conceptual process flowsheet, and the final drawings for the plant.
In general, process design proceeds by the following steps: 1) All characterization and treatability test data are reviewed. These data are used to conduct preliminary engineering analyses and to formulate viable processing alternatives capable of meeting the project objectives. 2) The technology optimization and flowsheet design are evaluated in light of the collected data. At this point, the analysis focuses primarily on technical issues related to process performance relative to the degree of contaminant removal and percentage reduction in soil volume. Computer simulation techniques developed in the mineral processing industry are again used to evaluate each processing alternative. This approach improves both the speed and accuracy of data analysis procedures and significantly reduces the requirements for bench-scale and pilot-scale testing with their attendant cost. 3) Following the completion of computer simulations, bench-scale tests are used to resolve any uncertainties associated with the configuration and the operational behavior of the proposed circuitry and to refine the performance of the process units. 4) The conceptual design is then reevaluated, with emphasis on both technical and economic factors. 5) Based on these evaluations new processing alternatives might be introduced, and the engineering analysis procedures and technical/economic evaluations repeated. This iterative process eventually leads to development of a preferred processing strategy that will serve as the basis for the final conceptual design. Once identified, the preferred conceptual design is subjected to a detailed cost analysis so that the cost-benefit relationship for the remediation scheme can be established with a reasonable degree of certainty.
To properly size process equipment for a given flowsheet, the site characterization data is used to define the volume and location(s) of the contaminated soil. (There might be several locations.) The volume of contaminated soil is important to sizing the process equipment, since it is a key factor in determining the capacity or throughput of the conceptual process design. The capacity must be sufficient to accommodate the volume of soil to be processed in a reasonable and economical amount of time. Another important consideration in sizing process equipment is the constraints of the process site where the contaminated soil is located. The configuration of the treatment process must be designed to accommodate the location of operation. With these criteria in mind, the principle of "economy of scale" must be considered. In many cases, small-scale production equipment is more costly (on a per ton basis) than large-scale units. Therefore, most mineral processing plants are constructed to process high tonnages so that the unit cost per ton can be kept low. A large soil washing plant is particularly attractive for a site with a large volume of soil requiring treatment or if the soil from multiple sites can be combined for treatment.
Once the configuration and throughput of the treatment process are selected, the sizes of the processing units are determined. Their size must accommodate not only the throughput but also the configuration of the treatment process. For each separation technology unit, the volume of feed, product, and reject soil can be determined at equilibrium by a simple material balance around the unit (soil in = soil out). The individual process units are thus sized to accommodate the overall process design capacity and to handle the output of the preceding unit in the flowsheet. Because sizing usually requires a change from development-size units (laboratory and bench-scale units) to full-scale operational units, sizing the units is the stage in the design study where scaleup factors must be addressed. Scaleup is a very critical phase in the design process, since unit and plant operating efficiencies are often sacrificed when size is increased. To insure optimal results, the iterative design method described in this paper is usually applied to the scaleup operation, using both bench-scale and pilot-scale testing complemented by computer model simulation. In some situations, problems encountered during scaleup will necessitate use of an alternative technology to accomplish a process and simultaneously accommodate the larger scale. For example, soil particle-size separation might be performed by screening during laboratory or bench-scale testing but by wet classification in the full-scale treatment plant. Personnel with considerable processing experience at both levels of design and operation are needed to make these scaleup decisions. Technology vendor engineers can be very helpful with these decisions. At this time, the plant design and operation must also consider governmental regulations, such as those for noise control and land-usage when the processing units are sized.
After the throughput and overall configuration of the treatment process is selected and the equipment is properly sized, an evaluation is performed to determine the most economical size and layout of each unit operation within the overall process design. For example, it might be more economical to divide the feed stream into two separate streams and feed two smaller units operated in parallel. While the technical results may be the same, the capital operating costs may differ greatly.
Since the conceptual process design might produce more than one flowsheet with different combinations of separation technologies, an overall economic evaluation of each flowsheet (cost per cubic yard of remediated soil) is performed to determine the final conceptual design for the site. After the economic factors are established, any conceptual process design that is technically feasible, achieves maximum volume reduction, and meets the site cleanup criteria is subjected to an overall economic evaluation. Several combinations of separation technologies might achieve cleanup goals, however, only one will be the preferred process design (the semiworks plant). This design is selected by performing a thorough comparative economic evaluation including each technically feasible design. The economic evaluation considers capital investment, direct costs, indirect costs, overhead costs, operating costs, treatment cost versus disposal cost, and net savings over the life of the remediation project. The lifetime of the project is calculated from the capacity or throughput of the treatment process, volume of soil to be treated, hours of operation per month, and availability of the treatment process.
The initial capital investment is the cost to develop the site, build, transport, and assemble the treatment process on site, and connect utilities. Each proposed treatment process is evaluated based on its capital investment. Because of the capital investment, a treatment process might be uneconomical and be a less-likely candidate for the preferred process design. Operating costs include direct, indirect, and overhead costs. The direct costs include, among others, the labor to operate the treatment, utility costs, and other equipment and supplies needed to operate the treatment process. The indirect are items such as site supervision, security, maintenance of the process, and expendable supplies. The overhead costs include, among others, administrative costs, field office, site security, clerical labor and materials, and laboratory support. These costs are usually similar with each conceptual process design and, therefore, are not a significant volume factor in selecting the preferred process design.
From the direct, indirect, and overhead costs, an operating cost is calculated. The operating cost is calculated as a function of the total life of the project and as a function of the unit cost (treatment cost per cubic yard of soil). Each treatment process is evaluated based on the operating cost. From the economic comparisons for each treatment process, a preferred conceptual process design will be selected based on the total capital and operating cost for the site. This process design is the most technically-sound and economical for the site.
Once the initial capital investment and operating costs have been calculated for the selected process-design, the net savings of using the treatment process to remediate the site versus an alternate remediation proposal (including complete removal, transportation, and disposal in an approved landfill) is evaluated. The total cost of the remediation using the conceptual process design is calculated using three factors: (1) percentage of volume reduction, (2) total operating cost, and (3) and initial capital investment. By comparing the total cost of the remediation using the treatment process with the total disposal cost for the site, the cost effectiveness of the treatment process is determined. If net savings are significant, then the conceptual process design should be viable. It is important to recognize that a reasonable savings does not ensure that treatment should be undertaken. For example, even a good sampling program might provide only a 95 percent confidence that the samples are actually representative of the site material to be processed. Therefore, the potential savings must always be compared to the potential risk for failure. Since the risk is very difficult to estimate, the final decision to apply a treatment process should be left to very experienced personnel.
ENGINEERING DESIGN
When the economic evaluation is complete and the preferred conceptual process design has been selected, the semiworks treatment plant is constructed, tested, and optimized; and the final treatment plant is then constructed. Both are capable of meeting the desired cleanup criteria, maximizing the volume reduction, while remaining economically feasible for the site. Activities that are completed during this phase include: 1) preparation of a narrative report that completely and concisely describes the processing systems incorporated within the semiworks plant; 2) preparation of a complete summary of all test data, literature information; assumptions, engineering estimates, and design criteria used in the development of the semiworks plant design; 3) development of complete material balances that specify the expected solid and liquid flow rates based on particle-size analyses, float-sink data, and other characterization and bench-scale and pilot-scale test data; 4) preparation of a detailed listing of required unit operations, including equipment type, unit size, throughput capacity, reagent/chemical requirements, power requirements, air/water requirements, operating limitations, vendor cut-sheets; 5) construction of a detailed process flowsheet that clearly illustrates flow rates, solid contents, particle size distributions, contamination levels for all primary and recycled process streams; 6) construction of a detailed plant layout diagram that specifies the physical arrangement of the primary operations, ancillary processing units, and all connecting streams; 7) preparation of a list of deficiencies associated with the design of the semiworks plant that can be overcome or minimized through additional laboratory, bench-scale, or pilot-scale testing; 8) construction, testing, and optimizing the semiworks plant; and 9) construction of the final treatment plant to be used on the site.
The preliminary flowsheet is formulated solely on the basis of the available technical data, focusing on the design of processing circuits that offer the best overall performance in terms of separation efficiency. In the second stage of the flowsheet design process, a revised flowsheet is again developed to address economic factors. Unitoperations and/or processing strategies that are not cost effective are redesigned and/or replaced by more economical alternatives that might be, however, less efficient in terms of technical performance. Finally, the revised flowsheet design is reevaluated in light of secondary considerations important to the successful implementation of the proposed soil treatment facility. At a minimum, these include those related to maintenance, operation, control, long-term performance, public acceptance of each of the selected processing circuits. If necessary, additional modifications to the flowsheet are implemented to accommodate these secondary considerations. The impact of these final modifications on plant performance and cost are evaluated and, if acceptable, are incorporated into the final design of the semiworks conceptual flowsheet. In addition, the conceptual flowsheets incorporate features and/or provisions that meet or exceed all applicable federal, state, and local regulations to ensure operational safety and minimize potential hazards to the environment. This iterative process is a combination of characterization studies, treatability studies, and process design. The process describes the best approach to achieve the most technically and economically feasible process design.
Those designs considered to be economically viable are subjected to flowsheet simulation to evaluate the technical performance of proposed conceptual designs. The simulation input values include 1) specification of the characteristics of the feed soil in terms of particle size, particle density and contaminant distribution and 2) specification of the relevant design parameters or operating conditions for each unit operation, known as the partition data, 3) detailed information related to the separation performance of each unit operation, and 4) the characteristics of each flow stream. The output values include a summary of clean and contaminated products leaving the soil treatment facility.
The simulator offers many potential advantages. First, the simulator is allowed a wide variety of processing alternatives and flowsheet configurations. Simulation allows the user to compare the performance of different plant configurations and to establish the impact of different operating modes on plant performance. The simulator provides consistent results with a rapid turnaround, resulting in substantial savings in analysis time, manpower, and cost. Second, the simulator allows the "cleanability" of other soil samples to be evaluated with a minimum of experimental testing. In essence, the simulator provides a vehicle for extrapolating the findings of this study into other soil remediation projects. Finally, the simulator allows potential problems associated with the buildup of particulate material or contaminants within recycled slurry streams or process water to be identified and resolved.
Next, the semiworks plant is constructed and subjected to detailed testing and optimization. By definition, the semiworks plant represents the first full-scale production facility for the proposed remediation scheme. The semiworks plant might be slightly different from future copies of the full-scale plant since there will be additional lessons learned during the construction process. Also, as with any prototype, the semiworks plant might not have all the refinements that will eventually be incorporated into a final treatment facility. However, the semiworks plant should be capable of achieving the target cleanup levels at the rated throughput capacity.
The last step in the development process is construction of the final treatment plant (or plants, if there is more than one treatment site). Since the semiworks plant represents the first full-scale production model for the remediation site, it is possible that, with appropriate modifications and adjustments to provide the design changes developed form study of the semiworks plant, the semiworks plant could be converted into the final treatment plant. With this step a successful remediation study is complete and cleanup can begin.
CONCLUSIONS
There are only a few basic prerequisites for a meaningful and successful remediation study on soils contaminated with radionuclides, but they are critical to the study. They begin with an overall understanding of the study process itself, each individual step, their interrelationship and importance to the process, and the contribution that the experiences of the mineral processing industry makes to a successful study. In the past, characterization studies have often been mistaken for complete treatability and remediation studies, and the study process has been left incomplete, prematurely terminated because of this mistake or a limited understanding of a thorough remediation study. In addition, those individuals participating in the study and the stakeholders themselves must have a reasonable knowledge of the physical, chemical, and mineralogical attributes of radionuclide-contaminated soils. A complete and judicious selection of the best remediation solution requires this knowledge to make a proper treatment selection. It also requires a commitment to the study process itself: to see the protocol to its proper end and to use the results in the remedy selection.
An understanding and appreciation of the cost of the remediation study at its beginning is very important. The cost is not absolutely fixed but is dependent on the results as they are generated during the study itself. Some studies will be completed sooner than others, because a treatment process is found to apply to the contaminated soil early in the characterization study. Others will last longer and proceed to a cleanup process only if the proposed treatment technology successfully competes with alternate remediation proposals after proceeding through the characterization, treatability, and design studies. The indeterminate nature of remediation studies does not mean that the cost cannot be reasonably estimated. A range of costs can be prepared at the outset of the study, but everyone involved needs to be aware of its indeterminate nature. These costs should, however, be considered within the broad overview of the overall cleanup cost of the site. Ultimately, regardless of the remediation process finally selected, a treatability study will be a small part of the total cost. It is reasonable to invest in a proper, thorough remediation study at the outset of a site cleanup with the potential for significant cost savings as the result of the study rather than select a more costly and possibly less suitable solution because of an incomplete evaluation of the remediation alternatives.
Complete site characterization with an overview of radionuclide concentration and mineral nature of the native soil as a three-dimensional function of the site matrix is essential. This information provides a realistic estimate of the magnitude and extent of contamination. It is imperative that the cleanup criteria are established prior to a characterization study. Absence of well-defined cleanup criteria will be detrimental to the entire remediation study and lead to very costly errors in evaluation, planning, and remediation attempts.
The tiered approach to remediation studies is especially important to their successful completion. The approach provides an evaluation of the applicability of treatment technology at appropriate times in the study, after sufficient data has been collected to make an informed decision. A tiered approach applies a protocol that permits the study to proceed in an orderly manner, collecting data in steps that will reveal the next phase of the study needed, eliminating unnecessary tests, and permits the study to follow promising results while avoiding expenditures on obviously unproductive tests. The approach does not, however, imply compromising the study by bypassing or attenuating tests needed to obtain the data required to make the best decisions about the study protocol. Neither does the approach suggest that one more complex treatment method might not be successful just because another, simpler method is not productive. Treatment decisions are greatly affected by the success of the separation processes and their ability to compete with alternate remediation methods, but unless complete, reliable data are available to make these decisions, treatment processes might be prematurely eliminated from consideration or, alternately, given too much emphasis as a promising remedy. In addition, treatment processes may consist of multiple unit processes based on more than one physical separation principle. Errors produced by this lack of understanding are potentially very costly, time consuming, and negative to the development and use of treatment technologies that in other circumstances might make important contributions to remediation efforts at other sites.
Successful remediation studies rely on application of separation technologies developed for the mineral processing industry and on the iterative nature of the studies used in the industry. These studies should be supported by computer simulations complemented by laboratory, bench-scale, and pilot-scale studies. This iterative process is applied to develop the most effective and economical treatment/remediation design and optimize its performance to save time, money, and materials. The preferred design must be cost-effective, meet or exceed the stated cleanup criteria, and maximize the overall volume reduction of contaminated soil. To achieve these goals, it is important to eliminate unnecessary and costly tests with computer-model simulations on the probable combinations of process units selected from study.
REFERENCES