Glass facilities are facing new requirements from state rules and U.S. Environmental Protection Agency (EPA) enforcement action to install emissions monitoring equipment on their furnaces. States such as California, Pennsylvania and Wisconsin have passed regulations to limit the amount of NOx emissions from glass furnaces. The EPA has also targeted the glass industry for enforcement action. So far, this has resulted in one consent decree with a major glass manufacturer that requires emission and opacity monitors on all of the company’s furnaces.
Continuous emission monitoring systems (CEMS) present unique challenges for the glass industry. Emission monitors are a substantial investment for a facility, and they add significantly to a plant’s regulatory work load. It’s important to select the right equipment at the start.
A CEMS comprises three main elements: the sampling system, the analyzers, and the data acquisition and handling system (DAHS). The sampling system extracts and transports a representative sample of the stack gas to the analyzers. The analyzers determine the concentration of the pollutant in the sample, and the DAHS collects and reports the data (see Figure 1).
Gas Sample Extraction Methods
Flue gas is extracted from the furnace stack and transported to instruments where the concentrations of the components of interest (NOx, SO2, etc.) are measured. The method of extraction is one of the most significant design considerations in the CEMS system. It affects the type of monitors that will be used, the equipment that is needed, the amount of maintenance that will be required, and the types of problems that can be expected.
It is essential that the extraction method prevents the condensation of water. Condensed water in a sampling system can drastically affect the measurement accuracy and can create acid gases that will damage the analyzers.
There are three primary types of extraction and delivery methods. The selection of the proper method depends on what components are to be measured and their relative concentrations.
Hot/wet extraction measures the stack gases without removing the moisture. The entire system (sample transport line, analyzer and pumps) is heated to approximately 350°F to keep any moisture from condensing in the line, and the sample is measured on a hot/wet basis. The concentration is measured as ppm-wet, which means that it is the volume of the component in a wet sample. Some limits are in ppm-dry and require that the moisture be measured and the sample value corrected.
Dry extraction transports the stack gas in a heated line to the analyzers, where the moisture is removed in chillers prior to analysis. The gases then enter the analyzers and are measured in ppm-dry.
Dry extractive systems are common but have a significant drawback when used on glass furnaces. Glass furnace exhaust can contain SO3, which forms sulfuric acid mist in the chillers. This acid is difficult to remove and can cause maintenance problems, increase monitor downtime, and shorten the life of the monitors.
This approach adds a known quantity of clean, dry air to the extracted sample immediately after it is removed from the stack. The diluted mixture has a very low dew point, so the moisture doesn’t condense in the sample line or the analyzers. The sample is measured in the diluted state by the analyzers and the values are corrected for the dilution ratio. The concentrations are measured on a ppm-wet basis.
Of the three approaches described here, the dilution technique has proven to be the most effective sampling method for the unique characteristics of glass furnace flue gas.
Gas monitors can be single- or multi-component analyzers. Single-component analyzers are the most common and measure one parameter; a different analyzer is used for each component measured.
Multi-component monitors use one instrument to monitor a number of parameters. However, since the cost of the base instrument is high, these types of monitors don’t become cost effective until they are used to measure at least four or five components.
Stack flow is used to calculate the mass of emissions in units of lbs/hr, lbs/day, etc. Flow monitors determine the velocity of the stack gases in ft/second. The stack diameter and exhaust temperature are used to determine the flow rate in standard cubic feet per minute (SCFM).
Some sources calculate flow by measuring CO2. This practice, which is based on different fuels (e.g., natural gas, coal, fuel oil) creating a known volume of CO2 when they burn, works well in cases where all of the CO2 comes from the fuel being burned.
A significant amount of CO2 is formed in glass furnaces from the carbonates in the batch, however, which can lead to large errors in the flow measurement if the flow is not properly corrected. As a result, this method is not recommended for glass facilities.
Pitot tubes, coupled with electronic pressure transducers (manometers), can be used to measure the velocity in the stack. The pitot tube is usually positioned in a single location in the stack (see Figure 2).
While inexpensive, this method is prone to pitot tube fouling and may not be accurate if the stack flow profile changes. The flow has been seen to move around in the stack, making flow measurement tricky. Changes in pull rate, damper settings and educator fan speeds can cause this to happen.
Ultrasonic Flow Monitors
In ultrasonic flow monitors, an ultrasonic pulse is sent at an angle through the stack to measure flow. The flow causes the ultrasonic pulse to move faster going up the stack (with the flow) and slower going down the stack (against the flow). These time differences can be measured and used to calculate a stack velocity.
The two heads of the flow monitor must be offset to create the necessary time differences, which may require the installation of a second platform on the stack (adding to the costs) or create maintenance issues (see Figure 3). These monitors also have temperature limitations and cannot be used on some of the high-temperature stacks found in the glass industry.
Optical Scintillation Flow Monitors
Temperature causes turbulence in gases, which affects light transmission. This effect is called optical scintillation (OS). Common examples include the twinkling of stars, which is caused by the atmosphere, or heat distortion from hot pavement.
OS can be measured and used to determine air velocity. This technology, which has been used for some time to measure wind shear at airports, has recently become available for stack monitors.
The advantages of OS are that the two heads are mounted straight across from each other on the stack, so no extra platforms are required (see Figure 4). The OS technology also works at high temperatures, including those on the hottest glass stacks. While this is a new technology compared to other types of flow measurement, it has been proven in over a dozen glass furnace stack installations.
Data Acquisition and Handling System
As the main human/machine interface, the DAHS is the heart of the CEMS system. It is what the operator uses to determine compliance and to identify and address problems. The DAHS, which consists of a computer and may include a data logger or PLC, typically performs the following functions:
Desirable DAHS features include:
The DAHS is a software system that is written and created by the programmer, and it is customized for each application. Various systems feature significant differences in their capabilities, ease of use, and look and feel. Special attention should be given to evaluating and selecting the DAHS to ensure that it fits the plant’s specific needs.
CEMS Best Practices
The goal of any monitoring program is to provide maximum data capture (system uptime) and the tools that the operator needs to stay in compliance—and to do so with a minimal workload to the plant operating personnel. The CEMS should be designed so that all of the daily checks are done at the DAHS in the control room. The facility’s quality assurance/quality control (QA/QC) plan should list the daily checks required to ensure that the system is operating properly. These checks may include sample flow rates, system pressures, vacuum pressures, calibrations and calibration gas pressures, and system alarms.
Many systems require the operator to leave their workstation and perform these checks on the instruments themselves. The CEMS may be located some distance from the control room, which means the operator must stop monitoring the furnaces to do the checks.
A better approach is to place sensors on all of the critical parameters and tie them into the DAHS. This practice allows the required daily checks to be done in the control room without the operator leaving their station.
Alarms need to be clearly displayed and not subject to interpretation, and should tell the operator if action is needed immediately. Alarm responses should also be very simple. Furnace operators are trained to operate furnaces, not CEMS, and should be expected to respond to emissions problems by managing the furnace.
An outside service should be available for managing CEMS hardware or software problems. The service should include a 24/7 service number, an email alert system and an Internet connection to the system. The Internet connection allows the service provider to immediately connect to the system and start troubleshooting.
If the system has been designed with adequate connectivity, the service provider can perform a detailed diagnosis and often fix the problem remotely. If the problem can’t be corrected online, then the service provider can determine if the problem can wait until the next scheduled maintenance visit or if a technician needs to be sent to the site immediately. In all cases, decisions about the CEMS should be made with the help of a trained CEMS professional.
Periodic preventive maintenance (PM) should be required no more than quarterly. In addition to PM, some CEMS systems require monthly or weekly checks that need to be done by plant personnel. This requires the plant personnel to be trained to perform and document the PM checks. Failure to perform or document a PM could result in a reportable deviation.
Instead, the system should be designed so that PM is performed once a quarter. Most sites have a CEMS service vendor on-site once each quarter to take care of required cylinder gas audits (CGAs) and relative accuracy test audits (RATAs). While on site, they can also perform the required equipment PM, ensuring that the maintenance is done by trained technicians and minimizing the plant’s workload.
The plant must have spare parts available to repair the CEMS. The equipment manufacturer determines the required spares and quantities, and these should be listed in the site’s QA/QC plan. Most plants buy a set of spares when they purchase the CEMS and keep them in their general storeroom. The spares can be expensive, sometimes costing $20,000 or more, and many of them are delicate electronic components (boards, power supplies, sensors, etc.) that can be damaged if not handled and stored properly.
An alternative is for the site to keep only consumable spares (those that are replaced during the quarterly PMs) on-site, and to have the CEMS vendor stock the other spares. The vendor must agree to 24-hr delivery of the parts to prevent excessive monitor downtime. In this arrangement, the vendor ensures that the parts are properly stored and maintained, and the plant is only charged for the parts when they are used.
For additional information, contact Air Tox Environmental Co., Inc. at 479 Tolland Turnpike, Willington, CT 06279; call (860) 487-5606; fax (860) 487-5607; or visitwww.airtoxenviro.com.