Caption title. "EPA/600-S3-89-027." "Aug. 1989." At head of title: Project summary.

The full report is a summary of the information provided in Data Assessment Reports for the first two quarters of calendar year 1988. The reports were submitted to enforcement agencies by owners or opera- tors of Subpart Da Electric Steam Generating units in response to the requirements of 40 CFR Part 60 Appendix F. Appendix F establishes minimum quality assurance requirements for continuous emission monitoring systems when used as the New Source Performance Standard performance test method. Data on Relative Accuracy Test Audits (RATA) were obtained for 24 facilities. Fifty-three RATAs were done for SO2; 31 audits were done for NOX. Eight continuous emission monitoring systems at six facilities exceeded the criteria of 20 percent relative accuracy for acceptable performance. In three cases the audits were repeated and relative accuracy was demonstrated to be less than 20 percent. In the remaining cases the boilers were taken out-of-service for maintenance. No periods of invalid data were reported because of excessive Inaccuracy based on RATAs. A comparison of audit results with data from a prior study suggests an improved distribution of Relative Accuracy values for the Da monitoring systems. Data for approximately 190 Cylinder Gas Audits (CGAs) were obtained for SO2, NOX, CO2, and 02 analyzers at 19 facilities. With the exception of one audit, all analyzers were demonstrated to be operating within the specified plus or minus 15 percent accuracy. Approximately 60 percent of the results were within plus or minus three percent accuracy. No periods of invalid data were reported because of excessive inaccuracy based on CGAs. Almost 50 percent of the Data Assessment Reports (DAR) contained no information on Calibration Drift Assessment Nine DARs indicated no excessive calibration drift Fourteen reports included days on which one or more analyzers were found to exceed Appendix F criteria for excessive drift In all cases the criterion exceeded was for one day's operation. In almost all cases, periods of invalid data because of excessive drift were incorrectly reported in fractions of a day. Corrective actions for excessive drift were noted in the 14 DARs which included information on excessive drift. A majority of the corrective actions appear to be items that should be performed routinely in response to the facilities quality control plan.