OCR Text |
Show The sensor validation program and the emission prediction model are combined into one large model referred to as a Software CEM. This application runs in a batch queue on a supervisory computer in the control room. Other auxiliary application tools are provided on the supervisory computer that allow the operating system to continually check to assure that the application continues to m n , and to alarm and immediately restart the application in the event of failure. Data is continually extracted from the Distributed Control System (DCS) and fed to the sensor validation section of the application. The raw input data is validated and fed to the emissions model. The emissions prediction is placed into the D C S for display to the operator. A n historical record of emission rates is maintained in the plant database system. Finally, the state regulatory body requires that the application pass a relative accuracy test audit to receive certification. A R A T A test was performed by Ramcon Environmental to achieve final certification. Complete turn-key consulting services are available, including development, installation, and maintenance of the Software CEM . Savings to industry through the broad application of this technique will be very large. Benefits of The SOFTWARE CEM Hardware CEMs provide only one benefit: reliable measurement of emissions. In contrast, Pavilion Technologies' Software CEMs offer a host of additional benefits. These additional benefits include lower initial capital cost, lower maintenance cost, a better understanding of key variables affecting emissions rates, and the potential for closed loop control to reduce emissions while meeting operational constraints. Process Insights, the software package used to construct Software CEMs, can be used by the plant technical staff to model and optimize plant unit operations, resulting in dramatic operational savings. The combination of use of the software for regulatory purposes and creation of savings through selection of a second project for process modeling and optimization, provides a path to make compliance a profit source, rather than a cost center. Savings generated from the process modeling and optimization project can far exceed the cost of the compliance project. References: 1. Denmark, Farren and Hammack, "Turning Production Data Into Value: A History and Analysis of Texas Eastman's Experience with Adaptive Technologies in Process Modeling", Texas Eastman Management Engineering Services 2. Keeler, James D., "Vision of Neural Networks & Fuzzy Logic for Prediction and Optimization of Manufacturing Processes", SPIE Conference Proceedings, Vol 1709, pg. 447-456, 1992, SPEE Publishing, Bellingham, Washington. 3. Ferguson, Bruce, "Chemical Process Optimization Utilizing Neural Network Systems", SICHEM '92 Invited key Note Address, Seoul, Korea 4. Kane, Les "How combined technologies aid model-based control", H P In Control Engineering Editors section, Hydrocarbon processing, M a y 1993 Pg.4 III-15 |