Smokestack emissions from H-POWER are subject to regulation by the state Department of Health. In issuing a permit to H-POWER, the department set certain standards, therefore, and imposed upon H-POWER’s operators reporting requirements so that H-POWER’s performance can be measured against those standards.
Two basic types of reports are to be submitted. First, there are the reports, submitted quarterly, based on the continuous emissions monitoring equipment. That equipment measures the quantities of sulfur dioxide, nitrogen oxides, carbon dioxide and particulates that enter the atmosphere from H-POWER’s stacks. Four times a year, H-POWER’s operator, Honolulu Resource Recovery Venture (the partnership of Amfac and Combustion Engineering that runs H-POWER for the City and County of Honolulu) reports to the Department of Health the dates and times during which emissions of those pollutants exceeded permitted levels.
The second type of report is based on a more thoroughgoing series of tests for a wide range of airborne pollutants. These tests, conducted annually as a rule, look for the amounts of heavy metals (including arsenic and lead) and toxic chemicals (including dioxins and furans) that are typically found in H-POWER’s exhaust.
Environment Hawai’i was not able to review either type of report submitted for emissions in the 1992 calendar year. The annual test report had not yet been submitted to the DOH (testing was to have begun May 11), while the reports on exceedances recorded by continuous emissions monitoring equipment were not made available by the Department of Health.
Emission reports for 1991 were made available for inspection, however. They do not show an unblemished record of compliance.
Putting the Lead Out
In 1991, the annual, more exhaustive series of tests was conducted in late March by Entropy, a testing company under contract to HRRM. Results of those tests were summed up in a letter on May 21, 1991, from Ronald L. Davies to Paul Aki, head of the Department of Health’s clean air branch, accompanying the test results.
Davies wrote: “These results indicate compliance with permit limits by wide margins for all pollutants, with the possible exception of lead. Lead emissions, as recorded, were slightly above the facility permit limits. … Since the lead emissions are so close to the limits, we requested and received permission from your staff, at a meeting on May 17, to retest both boilers for metals and particulates.”
Actually four days earlier, on May 13, Aki had had a telephone conversation with Michael Hartman of Combustion Engineering (in Connecticut), in which Aki had indicated agreement with the idea of a second round of testing for metal. From notes of that conversation, it would appear that Hartman was wanting to keep the initial test results “preliminary” The “final results” would be based on “retest data.”
Whether lead emissions were “close to the limits” may be a matter of interpretation. H-POWER’s operation limit is one-tenth of a pound of lead per hour, per boiler (there are two boilers). In all six reported test repetitions for metal (three per boiler), those levels were exceeded. The lowest recorded level was 0.107 pound per hour; the highest was 0.205.
The second round of tests was conducted in late May by a different testing company (Clean Air Engineering). As stated in CAE’s report, “This retest is to remove any uncertainty about the lead emission rates and emission rates of particulates and other metals.” Indeed, the levels of lead were roughly an order of magnitude smaller. Instead of lead being measured in tenths of a pound, it was detected in most cases in the hundredths of a pound.
The level of recovery- to say nothing of its speed – is remarkable. That said, the circumstances under which the testing was done bear scrutiny.
Bad Rappers
On July 2, 1991, Davies, HRRV’s plant manager, wrote Paul Aki once more. In a letter covering the results of the second metals test, Davies offered an explanation for the poor performance of H-POWER during the first test run. (In what follows, ESP stands for electrostatic precipitator, one of the means by which material is removed from the stack gases. RDF stands for refuse-derived fuel, which is the dried and fluffed solid waste that is fed into H-POWER’s boilers.)
“Unusually wet fuel” and “non-optimum ESP settings” were cited by Davies as contributing factors to the excessively high lead levels. In any event, he wrote, “the amount of the average lead emissions over the permit limit was so small as to about equal the statistical uncertainty in the test methods for lead.”
He continued: “In order to produce the high steam flow representative of maximum operating capacity (as required by… permit), we had to significantly increase the RDF firing rate, due to poor fuel quality, during the March tests. We failed to realize these conditions required a modification to the ESP setting. In addition, a failed rapper limit switch caused one set of rappers on the ESP to continually cycle, which reduced removal efficiency.” (In the electrostatic precipitator, electrically charged rods capture particles in the hot stack gases. These particles build up on the rods, which are then mechanically “rapped” from time to time, causing the particles to drop off and be disposed of as part of the fly ash.)
The report for the first test rounds appears to support Davies’ claim of high moisture content and relatively low heating value for the refuse-derived fuel used in those tests. Fuel firing rates averaged between 41 and 44 tons per hour. Data from the second round of tests show the fuel used then had a much higher heating value and almost half the moisture. Fuel firing rates averaged about 39 tons per hour in the second round of tests. (These figures compare poorly with the fuel characteristics reported in the 1990 emission test results: the heating value was much higher and the average feed rate 35 tons per hour – far lower than in either 1991 test.)
Testing Anomalies
In the 1990 emissions report (discussed in the [url=/members_archives/archives1990.php]October 1990 edition[/url] of Environment Hawai’i), logs indicated that several of the tests had been conducted under conditions that were not standard. The 1991 tests, both those conducted in March as well as those run in May had similar problems, according to the logs kept by operators.
The Clean Air Engineering report describes a couple of these “unusual events”:
“During Run 3 on Unit 2, a screen in the slurry feed line plugged. This caused significantly reduced slurry feed and higher stack temperatures for a 20-minute period, therefore Run 3 was repeated.
“An unusual event also occurred during Run 4, the replacement for Run 3. There was an explosion in the primary shredder… The explosion triggered a fire protection water spray on one of the two RDF feed conveyors to the boilers. RDF feed to Unit 2 was cut back for approximately two minutes and oil igniters were initiated in the unit, to avoid too rapid a reduction in steam flow.”
Two tests run in the March series “were aborted due to unacceptable post-test leak checks,” according to the Entropy report, necessitating replacement tests. A chain of custody record for another test was “misplaced.”
In March also, Entropy was supposed to test the continuous emission monitoring system of H-POWER. One CEMS test was “aborted per lost power to one of their analyzers,” a test log shows. On another occasion, according to the logs, “CEMS #2 went down during calibration-waited for calibration to end. When it didn’t, checked out system and found system off because of high water level in system. Brought system back on line at 11:55.
Other problems noted included a valve stuck open that caused “opacity to spike to 40% before problem corrected” – in other words, particulate emissions were out of control until the valve was repaired. (Opacity, or the relative darkness of the smoke, is the most common measure of particulate emissions.)
‘Continuous’ Monitoring
The problems noted with the continuous emission monitoring system during the March tests reappear in the quarterly excess emissions reports made by H-POWER’s operators to the Department of Health. This system is supposed to keep a record of the concentration in the stack gases of particulates, sulfur dioxide, carbon dioxide and nitrogen oxides. For the first quarter, CEMS “down time” (that is, time when the system was not working) was reported on 34 days; for the second quarter, CEMS downtime occurred on 26 days; third-quarter CEMS downtime occurred on 23 days; fourth-quarter CEMS downtime was reported on 27 days.
The average duration of downtime seems to be in the range of one to two hours. There are some noteworthy exceptions, however.
From June 11 to June 24, 1991 – nearly two weeks – an analyzer for carbon dioxide was not working. For more than a day in July all data from the monitoring systems on both stacks “were not reliable due to a power outage.” During boiler start-ups, sulfur dioxide limits are routinely exceeded, and by significant amounts. Equipment malfunction is another reason frequently given for exceeding permitted limits.
The report for the fourth quarter indicates that the nitrogen oxides monitor for Boiler No.1 was out of commission almost 7 percent of the time the boiler was operating. (It was actually out of commission a total of 192 hours, although 43 hours of that time the boiler itself was not in operation. “Equipment malfunctions” were to blame for 182 of those 192 hours, according to HRRV’s report.)
Volume 3, Number 1 July 1992