《发酵与生物工程手册》（英文版）12 Centrifugation.pdf_大学文库

Centrifugation 563 3. 1 Pilot Testing Preliminary data taken in the laboratory regarding the separation characteristics of a product can be beneficial when beginning the equipment selection process. If one is retrofitting an existing unit with an identical system, it will not be as complex and time consuming as designing a"grass oots" facility. Pland data can then be helpful. however, should the process be altered one should evaluate the effect on centrifugation Careful attention should be taken to ensure that the existing tank design, peripheral pumps, piping and agitators do not provide shear that will cause particle degradation. Capital dollars spent in this area on crystall tion studies, the selection of the correct pumps, etc, will directly impact the capacity of your equipment as particle degradation will significantly affect throughput and final residual moisture adversely, as it does filtration.( See Ch. 8 for a more detailed discussion If a process is existing and in-plant expertise is available to optimize the equipment, it would be advisable to do so before a purchase. Upon review of the existing design, sufficient improvements can often be made in an older piece of equipment thereby avoiding a more costly investment. Vendors usually offer this type of assistance at no charge from their office or at a daily rate in the field 3.2 Data Collection The first step collect pertinent information to the process including a process flow sheet, product information and completion of a typical questionnaire, as shown in Table 2 Knowledge of the most critical aspect of the process can guide the sometimes difficult selection process. For example, the requirement of a very dry product with strict impurity levels suggests a filtering centrifuge. A product with a feed rate of 150 gpm, without wash requirements, would lead us to a continuous sedimentation centrifuge A simple Buchner funnel test will indicate fast, medium or slow filtration. Slow filtering materials that have inordinate quantities of particles passing the filter paper will be submicron and difficult to capture in a filtering system. Therefore, a sedimentation centrifuge should be considered A phenomenon called"cake cracking" can occur and will be evident in this simple test. Not all materials exhibit this. It depends upon the surface tension of the product and its tendency to shrink as dewatering occurs Amorphous, thixotropic materials will exhibit this more than rigid solids

Centrifugation 563 3.1 Pilot Testing Preliminary data taken in the laboratory regarding the separation characteristics of a product can be beneficial when beginning the equipment selection process. If one is retrofitting an existing unit with an identical system, it will not be as complex and time consuming as designing a “grass roots” facility. Pland data can then be helpful. However, should the process be altered, one should evaluate the effect on centrifugation. Careful attention should be taken to ensure that the existing tank design, peripheral pumps, piping and agitators do not provide shear that will cause particle degradation. Capital dollars spent in this area on crystallization studies, the selection of the correct pumps, etc., will directly impact the capacity of your equipment as particle degradation will significantly affect throughput and final residual moisture adversely, as it does filtration. (See Ch. 8 for a more detailed discussion.) If a process is existing and in-plant expertise is available to optimize the equipment, it would be advisable to do so before a purchase. Upon review of the existing design, sufficient improvements can often be made in an older piece of equipment thereby avoiding a more costly investment. Vendors usually offer this type of assistance at no charge from their office or at a daily rate in the field. 3.2 Data Collection The first step is to collect pertinent information to the process, including a process flow sheet, product information and completion of a typical questionnaire, as shown in Table 2. Knowledge of the most critical aspect of the process can guide the sometimes difficult selection process. For example, the requirement ofa very dry product with strict impurity levels suggests a filtering Centrifuge. A product with a feed rate of 150 gpm, without wash requirements, would lead us to a continuous sedimentation centrifuge. A simple Buchner funnel test will indicate fast, medium or slow filtration. Slow filtering materials that have inordinate quantities of particles passing the filter paper will be submicron and difficult to capture in a filtering system. Therefore, a sedimentation centrifuge should be considered. A phenomenon called “cake cracking” can occur and will be evident in this simple test. Not all materials exhibit this. It depends upon the surface tension of the product and its tendency to shrink as dewatering occurs. Amorphous, thixotropic materials will exhibit this more than rigid solids

Centrifugation 565 If the cake cracks, a liquid level must be maintained on top of the cake before washing to prevent channeling of the wash liquors. a crude estimate of the wash ratio, gallons of wash per pound of dry cake, can also be made in the laboratory The filter cake developed on a buchner funnel will have certain characteristics; the product may appear to have a defined crystal structure or be more amorphous. Microscope studies will indicate the shape of particles, which will be helpful in trial runs. Needle crystals, for example may break easily at high filling speeds, and during discharge on a basket centrifuge with plough platelets tend to pack in compressible beds and may be better suited to sedimentation than filtration. A particle size distribution will help in this analysis and is also required for cloth selection. (See Ch. 6) Cake compressibility is the ability of a cake to reduce its volume, i.e porosity, when stress is applied. The resulting cake will display an increase hydraulic resistance. This is not necessarily caused by an average change in porosity, as a porosity gradient can occur by the redistribution of the solid material. Rigid granular particles tend to be incompressible and filter well even with thick cakes. Materials that are easily deformed such as amorphous or thixotropic materials will respond well to mechanical pressureor operation with thin cakes. ( See Ch. 6 on Cake Compressibility. Laboratory test tube centrifuges can determine if there is a sufficient density difference between the two phases to consider sedimentation as an alternative. If there is a sharp separation, one can anticipate the same in the field. One can also answer the following questions. Do the solids settle or float? Is the solid phase granular or amorphous? What is the moisture content? The characteristics of the solids indicate the solids discharge design required, i.e., scroll in decanters, or in disk centrifuges, flow-through nozzles or wall valves A laboratory centrifuge such as the Beaker design( Heinkel) can also simulate the operation of a filtering centrifuge and verify product character istics, filtration rates and wash requirements. various filter cloths can also be tried only one liter sampl With these data summarized, one can now discuss the application with vendors or consultants with expertise in centrifuge operations to help simplify the selection process. Pilot plant testing can be done with 10-25 gallons at the vendor's facility or with a rental unit for an in-plant trial. Ifsufficient material is available, semiwork tests are recommended as more data can be taken for scaleup. Equipment manufacturers should be questioned about how they are scaling up, whether it is based upon volume, filtercloth area, etc and what accuracy can be expected. Critical to a trials success is how representative

Centrifugation 565 If the cake cracks, a liquid level must be maintained on top of the cake before washing to prevent channeling of the wash liquors. A crude estimate of the wash ratio, gallons of wash per pound of dry cake, can also be made in the laboratory. The filter cake developed on a Buchner funnel will have certain characteristics; the product may appear to have a defined crystal structure or be more amorphous. Microscope studies will indicate the shape ofparticles, which will be helpful in trial runs. Needle crystals, for example, may break easily at high filling speeds, and during discharge on a basket centrifuge with plough platelets tend to pack in compressible beds and may be better suited to sedimentation than filtration. A particle size distribution will help in this analysis and is also required for cloth selection. (See Ch. 6.) Cake compressibility is the ability of a cake to reduce its volume, Le., porosity, when stress is applied. The resulting cake will display an increase in hydraulic resistance. This is not necessarily caused by an average change in porosity, as a porosity gradient can occur by the redistribution of the solid material. Rigid granular particles tend to be incompressible and filter well even with thick cakes. Materials that are easily deformed such as amorphous or thixotropic materials will respond well to mechanical pressure or operation with thin cakes. (See Ch. 6 on Cake Compressibility.) Laboratory test tube centrifiges can determine if there is a sufficient density difference between the two phases to consider sedimentation as an alternative. If there is a sharp separation, one can anticipate the same in the field. One can also answer the following questions. Do the solids settle or float? Is the solid phase granular or amorphous? What is the moisture content? The characteristics ofthe solids indicate the solids discharge design required, i.e., scroll in decanters, or in disk centrifuges, flow-through nozzles or wall valves. A laboratory centrifige such as the Beaker design (Heinkel) can also simulate the operation of a filtering centrifuge and verify product characteristics, filtration rates and wash requirements. Various filter cloths can also be tried using only one liter samples. With these data summarized, one can now discuss the application with vendors or consultants with expertise in centrifuge operations to help simplify the selection process. Pilot plant testing can be done with 10-25 gallons at the vendor’s facility or with a rental unit for an in-plant trial. If sufficient material is available, semiworks tests are recommended as more data can be taken for scaleup. Equipment manufacturers should be questioned about how they are scaling up, whether it is based upon volume, filtercloth area, etc., and what accuracy can be expected. Critical to a trial’s success is how representative

566 Fermentation and Biochemical Engineering Handbook a slurry sample is. This is especially important with fermentation processes that change over timeand in-plant trial of reasonable scale may bemandatory There can be clear advantages to using a centrifuge over a filter, such as a drier product or a more effective separation. This will be dependent upon the application. However, there can be applications where a nutsche or even several types of centrifuges appear, from small scale testing, to yield similar product quality results. Ultimately, the decision will depend upon"operabil ity, that is installation, maintenance, and day-to-day operation. For these purposes, it is strongly recommended that the plant invest the time and relatively small capital costs for a rental unit to run an in-plant test on the product. Production scale rental units can be operated by taking a side stream from an existing process. This facilitates comparisons of the new equipment to existing plant operations. Rental cost for one month will be approximately three percent of the purchase price of the test equipment, credit for part of the rental is usually offered against the purchase of a production unit. Rental periods may be limited, however lease options are available 3.3 Materials of Construction Various materials are available as with all process equipment, ranging from carbon steel coated with rubber, Halar or Kynar, to stainless steel 304 316 and higher grades, or more expensive alloys such as titanium, Hastelloy C-22, C276, or C4. These grades of Hastelloy will, however, double the capital outlays Coatings should be avoided if possible as product"A "can diffuse into the surface and potentially reverse its path. It therefore has the potential to contaminate product"B. Being permeable, they are also subject to peeling Coatings can be used most effectively on stationary parts dedicated to liquid use that are therefore, not exposed to maintenance tools, etc Working with an existing process will provide the most reliable data for choosing the most economical material for the service required. a new process may require a corrosion study by an in-plant metallurgist, if questionable. Test coupons are a relatively inexpensive way to conduct testing on a small scale and can be obtained from equipment vendors or from the materials manufacturers themselves. Companies such as Haynes, Allegheny, etc, will also have descriptions of the suitability of thei different alloys for various processes. An overview of corrosion testing and materials is presented in Perry's Chemical Engineer's Handbook, Sixth Edition. Sec. 23

566 Fermentation and Biochemical Engineering Handbook a slurry sample is. This is especially important with fermentation processes that change over time and in-plant trial of reasonable scale may be mandatory. There can be clear advantages to using a centrifbge over a filter, such as a drier product or amore effective separation. This will be dependent upon the application. However, there can be applications where a nutsche or even several types of centrifbges appear, from small scale testing, to yield similar product quality results. Ultimately, the decision will depend upon “operability,” that is installation, maintenance, and day-today operation. For these purposes, it is strongly recommended that the plant invest the time and relatively small capital costs for a rental unit to run an in-plant test on the product. Production scale rental units can be operated by taking a side stream from an existing process. This facilitates comparisons ofthe new equipment to existing plant operations. Rental cost for one month will be approximately three percent of the purchase price of the test equipment, credit for part of the rental is usually offered against the purchase of a production unit. Rental periods may be limited, however lease options are available. 3.3 Materials of Construction Various materials are available as with all process equipment, ranging from carbon steel coated with rubber, Halar or Kynar, to stainless steel 304, 3 16 and higher grades, or more expensive alloys such as titanium, Hastelloy C-22, C276, or C4. These grades of Hastelloy will, however, double the capital outlays. Coatings should be avoided if possible as product “A” can diffuse into the surface and potentially reverse its path. It therefore has the potential to contaminate product “B.” Being permeable, they are also subject to peeling. Coatings can be used most effectively on stationary parts dedicated to liquid use that are, therefore, not exposed to maintenance tools, etc. Working with an existing process will provide the most reliable data for choosing the most economical material for the service required. A new process may require a corrosion study by an in-plant metallurgist, if questionable. Test coupons are a relatively inexpensive way to conduct testing on a small scale and can be obtained from equipment vendors or from the materials manufacturers themselves. Companies such as Haynes, Allegheny, etc., will also have descriptions of the suitability oftheir different alloys for various processes. An overview of corrosion testing and materials is presented in Perry S Chemical Engineer s Handbook, Sixth Edition, Sec. 23