United States Environmental Protection Agency Office of General Enforcement Washington DC 20460 EPA-340/1-80-011 April 1980 Stationary Source Enforcement Series Techniques to Detect Failure in Carbon Adsorption Systems ------- TECHNICAL REPORT DATA . (Please read Instructions on the reverse before completing) REPORT NO. 340/1-80-011 3. RECIPIEr . TITLE AND SUBTITLE Techniques to detect failure in carbon adsorption systems. 5. REPORT DATE April 1980 6. PERFORMING ORGANIZATION CODE . AUTHOR(S) Theodore B. Michaelis, P.E. 8. PERF PERFORMING ORGANIZATION NAME AND ADDRESS Engineering-Science 7903 Westpark Drive McLean, Virginia 22102 10. PROGf 11. CONTRACT/GRANT NO. EPA 68-01-4146 Task Order 70 2 SPONSORING AGENCY NAME AND ADDRESS U.S. E.P.A. D.S.S.E. 401 M Street S.W. Washington, D.C. .20460 13. TYPE OF REPORT AND PERIOD COVERED Final 14. SPONSORING AGENCY CODE 5. SUPPLEMENTARY NOTES Robert L. King, Task Manager 6. ABSTRACT The study was originated to determine if simple techniques, or equipment inspec- tion could determine that a given carbon adsorption system was operating at satis- factory adsorbtion efficiency. The study determined that no visual inspection would be adequate, but that two other techniques are feasible. The first technique, utilizes portable, concentration- sensitive, continuous-measuring equipment which can detect break-through. Break- through is indication that a carbon adsorbtion system is not operating properly. This technique is suitable for field inspections. Maintenance of overall solvent material balances, subject to inspection, is an alternate technique. A form for a material balance is presented. 17. KEY WORDS AND DOCUMENT ANALYSIS DESCRIPTORS b.IDENTIFIERS/OPEN ENDED TERMS COSATI Field/Group Air Pollution Adsorbtion Inspection Carbon Adsorbtion Volatile Organic Compounds 18. DISTRIBUTION STATEMENT Unlimited 19. SECURITY CLASS (ThisReport! Unclassified 21. NO. OF PAGES 20. SECURITY CLASS (Thispage) Unclassified 22. PRICE EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE ------- EPA-340/1 -80-011 Techniques to Detect Failure in Carbon Adsorption Systems by Theodore B. Michaelis, P. E. Engineering-Science 7903 Westpark Drive McLean, Virginia 22102 Contract No. 68-01-4146 Task No. 70 EPA Project Officer: Robert L. King Prepared for U.S. ENVIRONMENTAL PROTECTION AGENCY Office of General Enforcement Division of Stationary Source Enforcement Washington, DC 20460 April 1980 "' '" ' ~ '" "* 7"onm3n~tal Protection Agency , - L • -rary (5PL-16) 'J~ --•• - K. xi?orn Street, Room 167Q Chicago, IL 60604 ------- NOTICE Mention of trade names or commercial products does not constitute endorse- ment or recommendation for use by the D. S. Environmental Protection Agency. ------- TABLE OF CONTENTS Section I II III IV APPENDIX A APPENDIX B SUMMARY INTRODUCTION Monitoring Techniques Material Balance MONITORING TECHNIQUES Catalytic Oxidation Systems Chromatography Systems Diffusion Sensors Flame-Ionization Systems Nondispersive Infrared Systems Ultra-Violet Systems MATERIAL BALANCES ORGANIZATIONS CONTACTED OR VISITED MATERIAL BALANCE FOR DETERMINING OVERALL SOLVENT RECOVERY EFFICIENCY I"1 II~1 II- 3 H-5 III-l III-l III-2 III-2 III-2 III-2 III-2 IV-1 LIST OF ILLUSTRATIONS Figure Title H-1 Carbon-Bed Adsorption System Simplified Flow Diagram II-2 Typical Discharge Concentration Vs. Time for a Carbon Bed Vapor Adsorption System ------- SECTION I SUMMARY For many industries the Control Technology Guidelines, published by OAQPS, define overall efficiencies for the affected facility. The documents also define control-device efficiencies which may be expected to achieve these re- quired overall efficiencies. Enforcement of the guidelines will result in installation of many volatile organic compound (VOC) control devices in the near future. Carbon-adsorption is the most widely-used technique for controlling VOC discharges which can show a return on investment. Therefore, it is reason- able to expect that many new carbon-adsorption installations will join the many existing installations. Techniques exist for determining the exact con- centration of solvent in inlet and discharge gas streams. These techniques require high technology to perform, and are beyond the scope of what might be considered feasible for continued surveillance. The purpose of this study was to determine if simple techniques were available for assuring that large carbon-bed solvent-vapor adsorption sys- tems operated properly, and were maintaining the design adsorption effici- ency. Large systems were defined as multi-bed (two- or more-bed) systems equipped with automatic-cycle control. In the process of investigating these issues, ES made telephone contacts and field visits to the following: o Existing installations; o Organizations which design and construct adsorption systems; and o Trade associations. The study determined that no visual inspection technique could reliably predict or define loss of adsorption efficiency. However, it did identify two techniques which could determine the loss in system efficiency. These were the measurement of solvent concentration in the discharge, and the pre- paration of a material balance. Portable instruments are available for measuring low-level solvent con- centrations in a gas stream. Although the instruments may not be calibrated for the gas stream being measured, they do indicate the increase in solvent 1-1 ------- adsorption system is not functioning properly. This technique is recommended for inspection of existing systems. Those jurisdictions which encompass seve- ral carbon adsorption systems should be equipped with a suitable solvent-con- centration meter, and recorder, and have personnel trained in their use. Such training is not difficult. Alternatively, a consultant familiar with carbon- adsorption systems might be called in to make the tests when they are required. Preparation of an overall solvent material balance can determine solvent recovery efficiency for an entire facility. Inasmuch as the Industry Guide- lines define overall recovery as their primary goal, this might be the better technique. The overall material balance would not be difficult for a faci- lity to maintain, or for an inspector to review. Many existing facilities are equipped to maintain the necessary records, and many others could be so equipped by the addition of one or more totalizing flow meters to measure re- covered solvent flow. Existing systems should be inspected periodically by either of the above- mentioned techniques. New systems should be equipped with instrumentation to continuously record discharge solvent concentrations, as well as with equipment to measure the quantity of solvents recovered. EPA may also require that each facility prepare monthly reports similar to those prepared for opacity-monitor- ing installations. 1-2 ------- SECTION II INTRODUCTION This project was originated.to determine if visual inspection could as- sure that large carbon-bed solvent vapor recovery systems retained the recov- ery efficiency for which they were originally designed. A large system is defined as a two- (or more) bed system which is regenerated by automatic con- trol. A list of typical failure modes was drawn up following discussions, and visits with trade organizations, design and construction corporations, and operating corporations. These are listed in Appendix A. The failure modes identified were: o Steam valves leak; o Air discharge valves leak; o Insufficient steam flow during regeneration; o Insufficient time for steamout; o Collapse of carbon bed; o Loss of carbon activity; o Bed blockage due to buildup of lint, dust, polymers, or other materi- als; o Uneven bed distribution; and o Improper setting of system controls. Based on inspections at six plants, it was determined that an expert might detect several of these failures in some systems, but it would be impossible to detect most of the failure modes by physical inspection. Two alternate techni- ques were then identified for determining the effectiveness of operating carbon- adsorption systems. These were: o Monitoring the discharge to determine if breakthrough occurs; and o Preparation of a material balance to determine overall system effici- ency. A discussion of the operation of a carbon-bed adsorption system may pro- vide an understanding of how these techniques may be applied. Figure II-4--is a simple operating flow diagram of a typical carbon-adsorption unit. Bed A is "on-stream". Solvent flows through the blower, up through carbon Bed A. II-l ------- FIGURE II-l STEAM CARBON BED ADSORPTION SYSTEM SIMPLIFIED FLOW DIAGRAM VESSEL A I VESSEL B SOLVENT LADEN AIR COOLING WATER TO ATMOSPHERE CONDENSER RECOVERED SOLVENT II-2 ENGINEERING-SCIENCE ------- Solvent is adsorbed by the carbon, and the purified air flows to atmosphere. Bed B is being regenerated. Steam flows down through the bed, heating the carbon, driving off previously-adsorbed solvents, and carrying them to the condenser. Condensed water and solvent flow from the condenser to whatever treatment is required to make the solvent reusable. Steam flow rate, and steam-out time are adjusted to provide adequate regeneration of Bed B before Bed A is loaded. Usually, Bed B is placed on stream when regeneration is com- pleted, and Bed A is regenerated. In a properly operating system, the bed being regenerated is placed onstreaca before the other bed is fully loaded with solvent. MONITORING TECHNIQUES Continuous discharge monitoring can demonstrate that breakthrough does, or does not ccur. Although not an exact test, it is probable that a system which goes through several cycles of full-load operation, without reaching breakthrough, is operating at adsorption efficiencies approaching the original adsorption effi- ciency for the VOC inlet loads during the test. Instruments suitable for these tests are discussed in Section III. Figure II-2 presents typical discharge sol- vent concentration plotted against time. The diagram is a representative example example of the general shape of such a curve. Actual time, and discharge concen- tration will vary from system to system. Initially, discharge concentration will be high because the bed remains hot from regeneration, and because the vessel is full of steam and solvent vapor. The bed cools rapidly, due to the evaporation of water which condenses on the carbon, and the discharge solvent concentration falls rapidly (1 to 2 minutes) to low levels (50 to 250 ppm). Discharge concen- tration then rises slowly. The beds are usually switched at the point marked "A" on graph. If a bed is kept on stream, the discharge concentration continues to rise slowly until it reaches point "B", where it begins to rise rapidly. This point is known as breakthrough. It can occur because bed changeover is delayed, because bed capacity is reduced, or because steamout is incomplete so that the time to breakthrough is reduced to less than the cycle time. A measure of total solvent capacity can be determined by interrupting the normal operating cycle and delaying regeneration until breakthrough has occurred. This can be express as: Time to Breakthrough x Solvent Loading at Time of Test - Actual Normal On-Stream Time Capacity II-3 ------- FIGURE 11-2 800 700 ^ 600- i500 Ul o o 400 (9 oc «c 300- 200- 100- 50- TYPICAL DISCHARGE CONCENTRATION VS, TIME FOR A CARBON-BED VAPOR ADSORPTION SYSTEM i i i i r r 12345 25 30 35 40 50 ON-STREAM TIME (minutes) NOTE: This curve is an example. Actual time will vary for each application. n-4 ENGINEERING-SCIENCE ------- This is not an exact measure of adsorption system capacity, but it can indicate significant deterioration. MATERIAL BALANCE An overall system material balance is the best technique for determining the overall system efficiency addressed in the OAQPS Guideline Series. Data to prepare a material balance could be assembled in many existing facilities. Other facilities could be modified simply by addition of totalizing flow me- ters to the recovered solvent lines. This is discussed in more detail in Section IV. II-5 ------- SECTION III MONITORING TECHNIQUES Determination of time to breakthrough requires that solvent concentrations in the 50-500 ppm range be continuously measured. Several types of instruments are available for measuring solvents at these low levels. o Catalytic oxidation; o Chromatography; o Diffusion sensor; o Flame ionization; o Instantaneous sampling; o Nondispersive infrared; and o Ultra-violet ionization. These systems each have some advantages and disadvantages. They require cali- bration for the solvent being measured if exact measurements are required, but can also be used without calibration to indicate relative concentration levels. If the output of a suitable detector is connected to a recorder, the adsorber- discharge concentration could be monitored over several cycles to demonstrate the absence of breakthrough. For a facility operating at normal production rates, the absence of breakthrough may be taken as evidence that the adsorp- tion system is operating satisfactorily. A brief description of each of the available instruments follows. Costs for VOC instruments are contained in a report "Summary of Available Portable VOC Detection Instruments" prepared by PEDCO Environmental, Inc. under Task No. 120 of EPA Contract 68-01-4147. It would not be difficult to train inspectors in use of the instrumentation, and in their application to carbon-adsorption systems. CATALYTIC OXIDATION SYSTEMS Catalytic oxidation systems measure the heat evolved when the organic va- pors in a gas stream are oxidized by a platinum-black catalyst. These units are light, and easy to use. System response varies with the heat of combus- tion of the organic vapor, but most hydrocarbons, including most chlorinated hydrocarbons can be detected. At low vapor concentrations, catalytic oxida- tion units must be zeroed every few minutes. They are not suitable for con- tinuous low-level measurement. III-l ------- CHROMATOGRAPHY SYSTEMS Chrornatography systems require considerable operator training if reliable results are desired. They are not suitable for this service. DIFFUSION SENSORS Diffusion sensors have been used^ for several years in carbon adsorption sys- tems. According to the manufacturer, detection of gases is accomplished by em- bedding a heater and collector in a metal oxide silica type material. At opera- ting temperature, a very high resistance exists between the sensor elements. Upon sensing gas, this resistance changes significantly producing several volts output in relation to the gas concentration. These instruments, available in portable sizes, can be set up to indicate concentration as low as 50 ppm full scale, or up to several thousand ppm full scale. One type of instrument could, therefore, be used to monitor both inlet and discharge solvent concentrations. An instrument of this type would be an excellent tool to detect breakthrough in working carbon-adsorption systems. FLAME-IONIZATION SYSTEMS Flame-ionization units are among the oldest of the low-level solvent-de- tection systems. A sample of the gas flows continuously through a pure-hydro- gen flame, and the resulting ions induce current flow to an electrode placed adjacent to the flame. Current flow is proportional to solvent concentration. Essentially, all hydrocarbons may be detected. These units are suitable for the application. INSTANTANEOUS SAMPLING There are several techniques which might be used to withdraw samples from the gas stream, and analyze them. The analysis may be rapid enough to be considered instantaneous, but the time between samples is usually several minutes. These techniques may be suitable for the application, but are usually not as convenient to use as continuous measuring techniques, and they cannot provide a continuous record. III-2 ------- NONDISPERSIVE INFRARED SYSTEMS Nondispersive infrared systems require that the system be set for each hydrocarbon to be measured and that the measured sample be free from water. Operation of these systems is relatively complex. The maximum absorption fre- quency for each measured hydrocarbon must be known, and the instrument must be set for that frequency. Less complex systems are available to detect break- through, so the system is not considered desirable for this service. ULTRA-VIOLET IONIZATION SYSTEMS Ultra-villet ionization is a relatively new technique for detecting the presence of organic-carbon molecules. High-energy ultra-violet light ionizes the molecules, which induce current flow between two elecrodes. System re- sponse varies with light-source energy, and with sensitivity of the organic compound, but most commercial hydrocarbons, including chlorinated hydrocarbons, can be detected. Commercially-available units are light, safe, easy to use, and suitable for this application. III-3 ------- SECTION IV MATERIAL BALANCES Material balances, taken over a reasonable time period (usually more than one week) can provide demonstration that the entire facility is maintaining the collection efficiency required by the facilities' operating permit. To be ef- fective, the- operating plant must maintain accurate records of solvent-contain- ing materials entering the process, and of the recovered solvent. Recovered solvent returned to the process is considered as part of the solvent entering the system. The solvent concentration of all materials entering the process must be known. A .form for a material balance is presented in Appendix B. Most of the information can be assembled by the source from shipping da- ta, and from component data available from raw-material suppliers. The only data which must be generated directly by the source is the quantity of recov- ered solvent. This can be determined by use of totalizing flow meters measur- ing solvent from the recovery systems. These meters are similar to residen- tial water meters. Many installations have such flow meters installed. It would not be difficult to install them in existing systems. All new systems should be so equipped. IV-1 ------- APPENDIX A ORGANIZATIONS CONTACTED OR VISITED ------- TRADE ASSOCIATIONS Flexible Packaging Association Gravure Research Institute Graphic Arts Technical Foundation OPERATING FACILITIES Meredith-Burda Corporation - Rotogravure Technicolor Laboratories, Inc. - Film Processing Adhesives Research, Inc. - Pressure Sensitive Tape Kecoughtan Laundry - Central Dry Cleaning SCM Corporation - Pressure Sensitive Tape Wabash Tape Corporation - Magnetic Tape Alco-Gravure Corporation - Rotogravure ADSORPTION SYSTEM MANUFACTURING ORGANIZATIONS NuCon Corporation Oxy-Catalyst Division of Research-Cottrell Corporation Vara International, Inc. SENSOR MANUFACTURING CORPORATIONS HNU Systems, Inc. International Sensor Technology, Inc. Bacharach Instrument Company Mine Safety Appliance Corporation Analytical Instrument Development Company A-l ------- APPENDIX B MATERIAL BALANCE FOR DETERMINING OVERALL SOLVENT RECOVERY EFFICIENCY ------- United Stales Environmental Protection Agency Office of General Enforcement Division of Stationary Source Enforcement Series Washington DC 20460 Official Business Penalty for Private Use $300 Publication No EPA-340/1-80-011 Postage and Fees Paid Environmental Protection Agency EPA 335 m 'TS 3 3 I C+1 1—' CD i O> O o o B cm CD If your address is incorrect, please change on the above label. tear off. and relurn to the above address If you do not desire to continue receiving this technical report series. CHECK HERE D tear off label, and 'eturn it to the above address ------- |