I will be live blogging from the OSISoft Conference in a couple of weeks. It has been about 3 years since I have been able to make it to a conference and I am really looking forward to catching up on the latest developments.

We have been working on the Thermo Fisher Scientific Mercury Freedom installation for 3 weeks and have hit a series of snags. We are the first production installation (there is 1 R&D unit running) of the probe controller on the stack using the 82X configuration dubbed the “X Box”. Basically we did not want to run copper lines down the stack because here in Flordia we call that a lightning rod. We instead opted to install the probe controller on the stack and use fiber converters to connect the analyzer and the probe controller which normally sit in the rack together and use RS-485 over copper. The communication via fiber seems to work well and I must say that the analyzer and calibrator with their i-Series displays and menu system are very easy to use. The analyzer coordinates the actions of the calibrator and the probe controller. Thermo excels at this sort of thing and those components are impressive.

The 82X design basically takes the guts out of the standard Thermo blue box and puts them in a NEMA enslosure that is mounted on the stack. Fiber converters carry the RS-485 signals but other than that it is supposed to be identical to the regular probe controller.The long umbilical runs from the analyzer and calibrator up the stack and into the probe enclosure. The long umbilical going up the stack contains only tubing, thermocouples, and heat trace (the thermocouples and heat trace are terminated on the stack inside the X Box). Another short umbilical connects the probe controller box and the probe. This short umbilical contains tubing and a series of wires which allow the probe controller to read the thermocouples, power the heat traces, and control all of the probe components such as solenoid valves, the large ball valve, and heaters for the probe, stringer, and converter. This short umbilical has caused many of the problems. We discovered that the first set (1 for each unit we are installing) contained some bad wires meaning that there was no continuity from one end to the other. We were sent replacements which I am told have the same problem. For now we are using outside wires to carry these signals.

We have discovered some problems with the umbilical heater circuits. This has been a confusing issue from the beginning because Thermo states that you need a heater zone for each 200ft of umbilical. Our lines are in the 500+ ft range so I originally planned on 3 zones but the probe controller will only support one or two zones so we have 2 zones of about 250+ ft each. Right now we cannot get the probe controller to properly read the two zones and there are some problems with the control circuitry so we cannot maintain the proper temperature. We received upgraded controller boards since they have been tweaked since our unit shipped but this did not correct all of the problems. This could be a grounding or shielding issue and we have some work to do to get to the bottom of it. At this point we have shut the systems down with a purge to keep everything clean and we are working with Thermo to determine the next step.

It is expected to have some hiccups at this early stage. I am told that Thermo has about 100 systems shipped so far and I’m sure that most of those have not yet been installed. I’m pretty confident that we will get past these problems.

On another note, I received confirmation that your calibration gas hydrators, a firmware upgrade, and a lamp controller update for the analyzer were all shipped and should arrive next week.

I’ll post again next week with more details. I imagine that it will take a few weeks to get everything straight.

We are in the early stages of commissioning two new Thermo Fisher Scientific Mercury Freedom Systems at St Johns River Power Park in Jacksonville, FL. It is nice to have a customer who has decided to be an early adopter of this technology. There are still a lot of unanswered questions and uncertainty but I think that it is ultimately wise to jump in and get the system installed, get hands on experience, and develop your in house expertise as early as possible.

The mercury systems were added to an existing project to replace the NOx/SOx CEMS that were installed in 1994. We spent a considerable amount of time specifying the NOx/SOx systems but did not have much time to add the mercury systems to the specifications. After looking at the available options, the plant decided that the best option was to go with Thermo. I got up to speed on the mercury regulations and technologies as quickly as possible and added the equipment to the specification.  

We decided to use the Thermo supplied rack. In hind sight we should have chosen to purchase the components and install them in Hoffman racks to match the other systems. We specified the NOx/SOx dilution air train to be able to supply the 3cfm required for the mercury systems. I decided to let Thermo supply their air cleanup train so that if we had any problems with the system they could not point to the dilution air system as a cause. The plan was to commission the system and then switch to using the common dilution air train and retire the Thermo supplied components. We have hit a snag in that plan. We are going to add a nitrogen generator and use nitrogen as the dilution gas in the mercury system and this requires 6cfm so it looks like the dilution air train supplied with the NOx/SOx will not be able to supply both systems. I am looking for the easiest way to resolve the problem.

We were a little disappointed in the layout and design of the cabinet. Some of the dilution air components are installed on the outside of the cabinet. We have removed the beige panels from the front of the rack so that we will not have to unscrew them when we need to get into the cabinet.

The back of the cabinet is pretty cramped. In this picture the umbilical has not been terminated so it is just hanging in the cabinet.

Some of the air cleanup train is mounted to a pull out shelf in the middle of the cabinet.

The vacuum pump appears to be high quality. It probably should have been mounted on a pull out shelf as well to make it easier to get to.

We only have the analyzer and calibrator in the cabinet. The probe controller has been mounted on the stack. All power terminations are made on the stack and the analyzer communicates to the probe controller via a fiber optic line and RS-422 converter. This design variation is known as the 82X probe controller and is recommended for sites that are prone to lightning as this installation in Florida, the lightning capital of the US.

We are using the 83i probe for now. We have an agreement with Thermo that will allow us to switch to the GC probe if we desire. The 83i was delivered with the system and a GC was not available prior to commissioning so we went with what we had.

Here is the probe with the stinger and return tube installed. The stinger and “tea cup” at the end are hastelloy. The return tube is Teflon. The support tube is Restek coated stainless. I would have preferred hastelloy and I predict that we will need to replace this item in 2–3 years.

One of the things that we liked about the 83i probe is that it has more room and should be easier to work on than the GC. I’m not entirely sure that this is the case after seeing the GC in person and working on the i. The above picture shows the probe after mounting it to the stack. It looks nice and roomy but notice that the umbilical has not been landed.

With the umbilical landed things start to look more crowded. There is a design tweak that needs to happen here. If you look a the picture below you will see that there is one bulkhead available for three of the lines. Many of the lines however terminate deeper inside the probe and there are a lot of Teflon tubes and fittings in there. You have to be careful routing all of these lines to avoid melting them to the clam shell that heats the probe. This differs from the NOx/SOx probes which typically run stainless lines back to a bulkhead where all of the umbilical terminations are made in one location. This helps to reduce the opportunity for leaks, eliminates the possibility of melting tubing, and makes the probe easier to work on at 2am when you’ve been called out because of a problem.

The 82X probe controller is shown below. As you can see they pretty much took the probe controller guts out of the blue box and mounted them in a NEMA enclosure. This is also a little bit of a rough design. I would recommend specifying a bigger enclosure and scrutinizing the layout a little more. I don’t know if Thermo will be super responsive to your requests so you may have to talk nice to your integrator. We currently have one of these units powered up and have found that the temperature control for the heat trace is a little erratic. We found that there is an upgrade to the responsible component and we should have it on Monday.

We still have a lot of work ahead of us. We will have the base systems installed next week and should be performing daily calibrations. Thermo will be sending us a lamp upgrade for the analyzer as well as a calibration gas hydrator that are now part of the base systems. We have also ordered the mercuric chloride calibrator to perform the weekly system integrity checks and nitrogen generators. These will all need to be installed and tested.

Overall I would say that Thermo has done an excellent job in choosing shier approach to the problem of reliably measuring mercury. The things that I have complained about are minor tweaks and when it comes to the major design decisions that affect the overall performance of the system I don’t think that anyone else can beat them.

I will post more as time allows after we have the systems running. I will also post some details of the system integration with the CEMS PLC and the DAHS.

We have recently completed upgrading two NOx/SOx CEMS systems at St Johns River Power Park in Jacksonville, FL. This project was started during the summer of 2006 when I started a series of meetings with the plant staff to review their existing systems piece by piece and ultimately developed a very detailed specification for the upgraded hardware and software. This was an important step in successfully completing the project.

The plant has environmental personnel who are responsible for the reporting and I&C staff who are responsible for maintaining the hardware. During the course of several meetings we discussed many design issues that they wanted to correct with the new design. We also specified in detail the equipment that we wanted to use and the standards that we wanted to follow. The time that we invested at this early stage has paid dividends in the final product. As with all new systems, we have had to track down some annoying bugs and work out a few unexpected issues but the upgrade went about as smoothly as can be expected.

Few things are as intimidating as ripping out a vital power plant system that has been running for about 13 years and committing to installing its replacement. We had a ten day outage for the installation of Unit 1. The goal was to use the lessons learned during the first installation to allow us to upgrade Unit 2 in 3 days. It was not easy but we were able to accomplish this goal.

The guys from Williams Industrial Services stand over the carcass of their recently defeated foe, the old CEMS cabinets. These things were built to last forever and did not give in easily.

Buzzards started to circle the stack about mid way through the outage. No one can remember that happening before. Despite the bad omen we pressed on and were ultimately successful.

A separate crew installed the new cabinets on the duct.

The new cabinets were staged outside the shelter so that they could be installed as soon as the old cabinets were removed.

The new cabinets were put in place.

We swapped the duct and stack analyzer racks after the factory acceptance test so that the new layout matched the old cabinet layout. This ended up requiring a great deal of re-wiring and re-tubing on site. For Unit 1 each cabinet was brought in separately and they were bolted together on site. This required the factory to disconnect all of the wiring between the cabinets. It took much longer than expected to reconnect everything. For Unit 2 we brought the air cleanup train, the duct analyzer cabinet, the stack analyzer cabinet, and the PLC cabinet into the shelter as a complete unit, rolling the cabinets along a few pieces of conduit. This proved to be a much better plan.

This picture shows the new Unit 1 cabinets on the right and the old Unit 2 cabinets on the left. 14 or so years of development makes a huge difference and the I&C techs love the new analyzers. They previously had pre-C series Thermo analyzers. I like to joke that the old analyzers actually had a little bird with a chemistry set who would then tweak a potentiometer to adjust the 4–20mA output like an old Fred Flintstone cartoon. The new i Series analyzers really pay off when you are prepping for linearities. The techs look so much more relaxed as they confidently flow the gas knowing that the analyzers will read as they should. They definitely don’t miss the drama of the old system.

One of our many design improvements was to mount the Teledyne Ultraflow 150 TIE components in a larger NEMA 4 box. This allows more room to work and the cables fold up neatly in the bottom. We don’t expect to have much trouble with the instrument but when they do it will definitely be easier to work on. Another tweak was to add a slow instrument air purge to the cabinet so that it always has clean dry air inside instead of stack gas.

The calibration gas cabinet to the new system has a couple of new tweaks. The CGA bottles have two solenoids in series to insure that a leak in a single solenoid will not cause them to fail a test.

One thing that we did miss in our spec and didn’t catch during the FAT was that the solenoid arrangement in the cabinet does not allow the solenoids to be easily removed. This was one of the few items in the new system that we were not completely happy with. I’m pretty sure that as soon as we wrap up the project the techs will be in here reworking this cabinet.

s
We upgraded the calibration gas racks. The old system had the regulators mounted to the bottle with Teflon tubing going to the cal gas solenoids. We installed a bulkhead and mounted all of the regulators in fixed positions with SS tubing to the solenoids. A braided SS hose attaches the regulator to the bottle. This makes everything easier to read and the bottles easier to change out. One problem during installation is that we did not have an experienced tube bender available so it ended up looking like an Escher drawing of impossible angles. This still needs to be cleaned up and is a remaining punch list item.

One major improvement in the design of this system was to install accumulator tanks near the probes for blowback. The old system performed blowback through a 1/4” line with the solenoid actuated down in the shelter. The new system uses the accumulator tank mounted at the probe and 3/8” lines. The accumulator pressure is monitored and the solenoid is actuated small PLC’s mounted on the stack and duct. The blowback is much more forceful than the old system’s and the time to perform a blowback was cut to a third of the old time. This means that the blowbacks are more effective and disrupt the system for less time.

We call this the multiplexer box and we have one on the stack and one on the duct. The top left corner has a little AB Micrologix PLC which acts as a MODBUS slave to the PLC in the shelter. The PLC controls the blowbacks and multiplexes other measurements onto a single fiber cable going down to the shelter. The PLC also measures the temperature of the cabinet so that we know how hot the equipment gets during the day. The PLC will easily allow expansion of the system if we need to add anything to the duct or stack in the future and was relatively inexpensive. We also used a larger box to allow for expansion. The box is also purged with instrument air. The old system used a fixed multiplexer that allowed a couple of 4–20mA and a couple of digital signals to be passed. The two blue boxes on the top right corner control the probe and stinger heaters on the new M&C probe.

The new M&C probe on the stack and duct replaced EPM out of stack probes. The M&C has a superior design for the coarse filter that allows it to be easily changed. The coarse filter also has a much larger surface area than the M&C so we expect fewer problems. The probe sits at an angle because the welder who originally installed this flange 20 years ago did not pay attention to the bolt pattern when he attached it to the stack.

This picture shows the Teledyne UF 150 TIE enclosure on the left, the M&C probe in the middle, and the blowback accumulator box on the right.

It seems that as the original Part 75 CEMS systems start to reach the 12–13 year old mark it makes sense to replace them. Old PLC’s, analyzers, and other instruments are much harder to maintain than the newest generation of equipment. Spare parts are becoming harder to locate. Availability suffers and your technicians spend too much time baby sitting temperamental equipment. It ultimately becomes justifiable to replace everything. Many utilities are also realizing as they prepare for new Mercury CEMS that it is a good time to revisit their NOx/SOx CEMS as well. Your techs will have their hands full with these complex new system so upgrading the old systems to free up their time is a good strategy. It is important to do the preliminary legwork and specifiy a system that will fix any problems or design flaws in your old system and insure that your new system will perform reliably.

Everyone struggles to comply with the assortment of environmental restraints that are the result of federal, state, local, and permit requirements. It is important to present information to your operators in an unambiguous manner that lets them know immediately if they are on target. I recently put together a little solution for a customer who was having problems. They could produce reports from their DAHS but it was cumbersome and the unit operators do not have time to login to the DAHS, select the report and set the dates, and then print it up. Furthermore the DAHS reports did not give them an adequate presentation of their progress toward the limit.

They use the GE NetDAHS software. Most people do not know it but there is a Minute Data ODBC driver that will allow you to access any of the data channels configured on the DAHS. I was able to build a web page that queries the minute data and calculates the hourly averages along with a running total of their emissions and present the status as shown in the screen shot below. As you can see it shows them how much they have emitted that day, the projected total based on their current operation and tells them whether they should be concerned. 

I was also able to add a trend of their emission rate and a graph of their accumulating total compared to the emission limit. This allows the operators to quickly glance at the page and get a feel for where they stand.

If you are having problems tracking your compliance, you should evaluate the way that you report and present the data to your operators. You may be able to eliminate compliance lapses by organizing and presenting your data in real time allowing your operators to quickly see where they stand and make more informed decisions.

From the geek standpoint I pulled a few tricks out of my sleeve to make sure that the application would run smoothly. I used ASP.NET and C# to code the web pages. I made use of the ASP.NET caching features to cache the dataset when a query is made. If another user requests the page within one minute, the cached dataset is used. This insures that no more than one request is made to the ODBC driver per minute so that the NetDAHS server will not get bogged down if a bunch of users start requesting the page all at once. It also means that once the dataset has been queried, additional requests are returned to the user much more quickly.

I met Troy Sonnier of Milivox while at the EUEC. His company digitally recorded all of the presentations given at the conference. His people had video cameras and laptops and were directly connected to the presentation computer so that they could directly record the PowerPoint presentation with the presenter audio. The camera captured the presenter and questions at the end of each session. It looks like a first class operation.

The recordings are available on DVD here http://www.millivox.net/public/EUEC.aspx. I purchased the package so that I can review some of the better presentations with my clients and view presentations that I could not attend. If you missed the conference, it might be well worth it to purchase the DVD’s of sessions that you wish you could have seen.

I’ll post a review of the package once I receive it in February.

I’m a little overwhelmed as I look through my 26 pages of notes that I have taken during this conference. I still have a lot to digest and a few new things to research.

Overall, here are my impressions regarding Mercury Measurement.

Time is wasting and the vendors are still working hard to improve their equipment. None of them have systems that would be considered “utility quality” in terms of reliability, repeatability, and serviceability. Many of the probe problems have been addressed and plugging is not a much of an issue as it was a year ago. Right now the problem seems to be in reliable calibration. They are addressing issues in obtaining a solid zero call and in drift between wet and dry measurement techniques on the span side. They are also exploring lower level measurements. There is more to learn here.

All of the vendors are still developing their oxidized mercury generators to test their converter efficiency as required. Richard McRanie of RMB says that if you are only calibrating on elemental mercury, you have no idea what you are measuring. You are effectively filling up your database with numbers but you don’t know if they mean anything.

Time is wasting and NIST and EPA still do not have traceability requirements established or protocols published for dealing with calibration against the standard. They are further along establishing requirements for the elemental generators. Most of the vendors use the same approach so developing a standard and testing criteria has been difficult but not as difficult as the oxidized generators. The vendors are using widely varying approaches for the oxidized generators so NIST is just now starting to develop evaluation criteria. These are non-trivial issues and will require time to develop. Meanwhile 1/1/2009 looms and we have no way to certify a system.

EPA reference methods other than Ontario Hydro are moving along. The instrument reference method (IRM) has been simplified and should be published soon but there is still a lot of contention with some of the elements. Richard McRanie of RMB has pointed out that testers are not moving on this due to the high cost and high uncertainty. Once an IRM is finalized they will not have much time to purchase equipment, test it, and train people. There are a lot of questions about the traversal requirements to measure stratification. You have to remember that the Hg concentration is prone to 10–50:1 excursions so you don’t know if you are measuring stratification or temporal changes in concentration. Jeff Ryan has cobbled together equipment and seems to be doing some excellent work on developing equipment specifically designed for the IRM that will be lighter and easier to use.

A sorbent based reference method is also on the way. Dennis Laudal who co-developed the Ontario Hydro is working on it. EPA is pushing the instrument reference method while many EPRI members are pushing the sorbent based method. I believe that a conceptual sorbent-based RM will be published in the next couple of months but the majority of the limited resources that the EPA has to expend on this will probably be spent on the IRM.

Vendors seem to be finding new problems at each of the early installation sites. There are still a lot of plant configurations, fuel sources, etc. that have not been tried and even among similar plants there are different problems. There is no way to predict what will happen when you install your CMMS, you will have to try and see.

Everyone is measuring x today and will need to be at 0.1x in the near future as they implement the required cuts in emissions. There is no telling if the system that you install today to measure x will work as well at 0.1x. This is uncharted territory. Thermo has started to test nitrogen as a dilution gas and has seen a 5x improvement in detectability. This seems to be an easy retrofit but we don’t know if there will end up being other ways to deal with this or if nitrogen will solve all of the problems. You may be in for major redesigns in the coming years.

Overall there are not many people out there with maintenance and troubleshooting experience with these systems. Also, experience gained at other sites will be valuable but there may be much to learn about your particular site. If you wait until the last minute to install, you may have trouble developing the expertise in house in time. This is a self serving statement on my part but it is important to remember that there is not going to be a trial period like part 75 CEMS where you have your certified system running but are not yet trading allowances. Mistakes and problems are going to cost $$$ from the start.

If you are waiting for all of this to wash out so that you can make a totally safe choice, it probably isn’t going to happen. If you are planning to go Appendix K then you should have a realistic plan of action and gauge the costs. A CMMS is going to cost between $275,000 – $480,00 for the instrument rack, calibrators, cleanup system, and umbilical. It will be expensive to maintain and there will be a lot of unknowns. There are risks associated with Appendix K. Make sure that you are realistic in your appraisal.

Measurement is going to be one vital piece of the pie but there are also a lot of other issues. You want to be able to make reliable and accurate measurements so that you can evaluate Mercury abatement strategies. There are tons of unknowns here and there are a lot of implications to different strategies. You have to realize that sorbents will end up in your ash and other waste products. You don’t want to make your ash unusable or create another waste problem. Recent court rulings also dictate that you can’t increase your particulate emissions in order to reduce your mercury emissions. In general you can’t increase the emission of one pollutant to reduce the emission of another.

To top all of this off, there are new studies that are undermining the assumptions that were used to justify the “slow” implementation of CAMR. There is going to be a big push by environmental and health groups to demand higher cuts in emissions on an even faster basis. The current legislation is way ahead of the technology now so I’m not sure how much faster it could happen but there will definitely be pressure and I would imagine new congressional bills will eventually be on the way. There is no telling how this will pan out. One thing for sure is that it is not going to be a boring choice of careers.

The information that you see on my site will most likely be the nuts and bolts details of implementing Continuous Mercury Monitoring systems. I have found a good source of mercury policy information with details of legislation and court challenges to CAMR. You can find it at the Mercury Emissions Blog.

enjoy.

We are just starting the second day of the EUEC. Yesterday offered a gem or two during the Mercury sessions.

Dieter Kita of Thermo Fisher gave an update on their systems. For the most part he covered things that I already knew. There were a few new items. The first is that they are using a humidifier to add a little moisture to the calibration gas before it goes up the stack. He says that this is giving them more consistent calibrations.

The HgCl2 generator to test the converter efficiency has been coming along. He showed a few trends showing the performance improvement that they have achieved so far. The generator now stabilises in about 5 minutes. Earlier versions took closer to 30 minutes to stabilize. This improvement was vital for an operational CMMS.

He also elaborated on the chlorine source that they are using. When I initially heard about the generator they were speculating that they would use a solid chlorine source. Later they said that they were going to use a gas cylinder. There has been concern that this would cause safety and hazard concerns. Dieter said that they are going to use a small cylinder with a 250ppm Cl gas. This should alley some of the safety concerns 

Another concern that has been raised in many of the presentations is the trouble monitoring low levels of Hg (<-0.5 micrograms/m^3). Dieter talked about a new development for their low source systems. They are using nitrogen gas for dilution. This gives them a 5x improvement in detection limit and eliminates much of the noise in the system. After looking at the trends that he presented, I suspect that we may see nitrogen dilution systems become the norm as monitoring progresses. It will add a little cost and complexity to the system but as removal rates progress, the tradeoff will definitely be worth it. The key to good measurement in low level situations is to get a solid and repeatable zero. The nitrogen dilution seems to help in the situation.

Tom Robertson of EQM gave a little status report of Hg monitoring. He said that to the best of his knowledge only about 50 CMMS have been installed nationwide of the 1566 coal plants that will require monitoring. The utilities that he talked to who had installed systems indicated that the maintenance costs and demand of technician hours were far greater than they are used to with part 75 systems.

There is growing concern that equipment vendors will not be able to keep up with the manufacturing demand. There will also be a shortage of experienced technical personnel and test crews. His said that the from his conversations with the EPA, they are not buying the argument that there will not be enough stack testers to certify all of these systems. The bet that the EPA will allow a delay in implementation might not be a good one, especially with the political pressure to implement the reductions even faster than CAMR requires.

I asked about the strategy of using Appendix K monitoring while waiting for the bugs in CMMS to be hammered out. He indicated that the costs associated with this would likely be close to the costs of implementing a CMMS.

Stephen Niksa of Niksa Energy presented the progress of his company’s efforts to model and predict Hg emissions using their MercuRator software. He appears to be using very complex first principle models to predict Hg removal and emissions from coal samples. It might be worth your time to take a look just to have an idea of what is out there. If he is successful at prediction it could save a lot of money in extended testing.

Today I am looking forward to Bobby Chen of Western Kentucky’s presentation regarding concerns about the proposed instrument reference method. I caught his presentation at the Thermo Super Group meeting in Chicago a few months ago. I will post any new insights.

An item of humor. As I sat through the train of PowerPoint presentations, I was reminded of a site that I saw about 6 months ago. Lincoln’s Gettysburg presentation given in PowerPoint. Check it out, it is well worth it.

I have just finished the early Hg CEMS sessions at the EUEC. Sid Nelson Jr. of Sorbent Technologies, Inc. gave an interesting presentation discussing recent Hg deposition and health effects studies.

The CAMR was based on a couple of fundamental assumptions, one that the deposition of Mercury was global, i.e. there are no hot spots and two that there are no adverse health effects at measured Hg levels in the US population. These two assumptions make it reasonable to implement the Hg regulations over the next 10 years as CAMR does. Recent studies suggest that neither of these may be true. I will attempt to obtain the papers that he discussed and post them. It is his assertion that the CAMR will not end up being the final rule and that there will be a push to accelerate the timetable fore reductions. The recent changes in congress will most likely lend themselves to that push.

Steve Norfleet of RMB Consulting gave a presentation discussing the various state rules. Illinois and Pennsylvania have proposed Hg regulations that are far more stringent than the CAMR and accelerate the attainment of targets. Sid suggests that these rules will be used as the model for future regulations at both the federal and state levels. It might be well worth the effort to browse through these rules and gauge the possible implications to your operation. It may give you a worst case scenario to consider when developing future plans.