![[logo] US EPA](../gif/logo_epaseal.gif){kind=logo}
Iowa
State Innovation Pilot Grant Program
ApplicantIowa Department of Natural Resources
EPA Region
Region 7
Project Title
Optimization of Power Plant Emissions: A Data Mining Approach
Project Location
Iowa City, Iowa
State Agency
Iowa Department of Natural Resources
Project Contact
Angela Chen, Ph.D., P.E.
Environmental Services Division
Iowa Department of Natural Resources
Wallace State Office Building
Des Moines, IA 50319
Phone: (515) 281-4736
Fax: (515) 281-6794
Email: angela.chen@dnr.state.ia.us
Funded by Any Other Federal Program
No
Regulatory Flexibility
None
Affirmation by Senior Official
This project is approved by Liz Christiansen,
Deputy Director of the Iowa Department of Natural Resources.
Project Duration
1 year
Budget Summary
[REDACTED BY US EPA]
Project Narrative
The goal of this project is to encourage boiler operators to modify boiler control settings to go beyond environmental regulation compliance.
To achieve this goal, the Department of Natural Resources (Department) will offer two incentives:
1) Allowing variances, without going through the PSD process, for test burns to verify emission reduction potentials.
2) Streamlining permitting process should the facilities decide to make permanent changes to their boilers to reduce emissions beyond compliance requirement.
To test the effectiveness of these incentives, the Department will partner with University of Iowa's Intelligent Systems Laboratory to develop and test Data Mining Algorithms at that university's power plant. The primary use of the knowledge and model developed in this project will be to minimize the environmental impact of emissions from the University of Iowa Power Plant. The trade-off between the boiler efficiency, energy production cost, and the volume of emission gases (NOX and SOX) will be analyzed. The proposed model can be easily generalized to other types of boilers and fuel. In fact, the University of Iowa Power Plant is in the process of introducing a mixture of a coal and an alternative fuel for firing boilers. In addition to analyzing the emissions from the coal, the project team will perform preliminary analysis of the combustion process of the new fuel mix. Boiler control settings leading to optimization of the environmental impact will be generated.
When fully developed, the availability of the model and incentives will be disseminated to boiler operators across the state. Success will be measured by the number of inquiries received and the number of boiler operators who take advantage of the model and the incentives to modify their boilers beyond compliance.
There are six boilers at the University of Iowa Power Plant. Across the state, there are more than 200 boilers that could potentially achieve emissions reductions and benefit from the results of this project. More significantly, it is estimated that there are tens of thousands of boilers nationwide that could use the same technique to identify emission reduction potentials and then use similar incentives to go beyond environmental compliance. The Data Mining Algorithms and Department's incentives can be easily transferred to other states and territories.
The Department will monitor the progress of the development and testing of the Data Mining Algorithms. Once this project verifies the potential emission reductions, the Department will follow up with the University of Iowa to implement changes needed. The Department will also disseminate the information to other power plants and work with them to reduce emissions at their power plants.
Project Schedule and Timeframe:
| Task | Completion Date |
| Receive funding from EPA | November, 2002 |
| Execute contract with University of Iowa | December, 2002 |
| Select data attributes | January, 2003 |
| Collect data at UI Power plant | March, 2003 |
| Mine the collected data | July, 2003 |
| Implement new control signatures | August, 2003 |
| Evaluate environmental impact | October, 2003 |
| Submit final report to EPA | November, 2003 |
Target Priority Environmental Areas: The proposed work addresses two priority environmental areas: reducing greenhouse gases and reducing smog.
Use of Incentives as a Tool: The State has been offering streamlined permitting processes for energy related projects. The state has also offered variances for test burns for such energy projects without going through the PSD process. The proposed method of reducing emissions has the potential to affect more than 200 boilers in the state. The goal is to encourage the operators of those boilers to go beyond compliance. In addition, since University of Iowa is a state university, the state will be leading by example to demonstrate that going beyond compliance is achievable and beneficial. In addition to reduced emissions, potential economic benefits from emission trading also will be investigated.
Uniqueness: Traditionally, it takes between eight to twelve months for a facility to receive a PSD permit. The streamlined process would significantly shorten the period to about 4 months.
The data mining approach is also unique in that it extracts novel knowledge from databases that can be used for better decision-making in industrial and business applications. For example, based on the prior data, the level of emissions from a power generator can be minimized. An important property of data mining algorithms is that the extracted knowledge can be explicit and used for informed decision-making. Using the data mining approach, a subset of all 'reliable' data (objects) for learning is determined. The properties learned on the reduced data set are used to make accurate predictions (decision-making) for a large percentage (hopefully 100%) of cases with unknown outcomes, e.g., values of control parameters minimizing emissions from a boiler burner.
Build on Lessons Learned: In the past, many emission sources would take the approach "if it is not broken, do not fix it". This traditional approach eliminates the opportunity for further improvements and the opportunity for further emission reductions beyond compliance. The proposed project takes the initiative of examining an existing power plant that meets the current permit requirements and attempts to further reduce pollutant emissions and at the same time increase efficiency of the steam production. The potential emission reductions from this project and its economic impact through emissions trading could lead others into action by examining their emission sources at reduced operating cost.
Develop and Apply the Innovative Tool to Effectively Demonstrate Success: The Department will work with the two major Iowa utility companies (Alliant Energy and Mid-American), the power plants at two state universities (Iowa State and University of Iowa), and other stakeholders to further develop the tool and demonstrate its success.
Transferring Innovation
Potential for significant environmental improvement: The proposed tools, allowance for variances during test burns and streamlined permitting process, will be attractive to most parties involved with energy related projects in Iowa. With the intention of making those facilities going beyond environmental compliance, the emissions that would otherwise be permitted at higher levels will be further reduced.
The proposed Data Mining Algorithms will derive a new model showing relationships between the level of emissions, efficiency, and operational cost of a boiler. This model will be easy to interpret by power plant operators, engineers, and managers. Using this model the level of emissions can be minimized, subject to cost and other operational constraints.
Potential for widespread use: Both the incentives and the analytical tool can be applied to all power plants that use coal, gas, or alternative fuels. Depending on the level of power plant automation, the proposed solution can be implemented in the form of a new software, a control chart, or control guidelines and therefore it will work with a control system of any type.
Promote changes or develop a culture: The application of this innovation will demonstrate that a meaningful change comes from integration of research results from vastly diverse areas, in this case a combustion process and a modern computational theory of data mining. Close collaboration between educational institutions, government, and industry is paramount to the success of these high impact innovations.
In addition, the application of this innovation will help to promote and encourage businesses and organizations to go beyond compliance without outside forces. When the economic and environmental benefits are shown clearly, a culture of innovative environmental problem solving as a way of doing business will be developed in Iowa and beyond.
Guaranteeing Measures and Accountability
Goals for the innovation: Reduce emissions from power plants by providing permitting incentives and developing a data mining tool for rational control of boilers.
Indicators to measure progress: The effectiveness of the incentives can be measured by how many facilities use them to reduce emissions and go beyond compliance. The progress of this project can be measured by the percentage drop in the emissions (NOX and SOX) that can be attained subject to the current operational constraints. In the future, the environmental impact can be assessed under modified constraints.
Objectives: Data collection, model development, model evaluation, implementation of the results at University of Iowa Power Plant, assessment of the results.
Performance indicators: Degree of compliance with the project tasks and schedule.
Information dissemination: Presentations of the proposed data mining approach and the results at meetings sponsored by the Iowa Energy Center, national energy related meetings, and publishing the results in technical journals.
Timeframe:1 year
Plans for measuring and evaluating:The subcontractor will meet bi-weekly with the University of Iowa Power Plant experts to update project partners on progress and concerns. Monthly report from the subcontractor will be required to keep the Department informed. The Department will send quarterly progress reports to EPA.
Achievable short term (2-3 years) results:Use the incentives and the analytical tool to reduce emissions at the University of Iowa Power Plant.
Achievable long term results: Development of the necessary regulations and implementation of the proposed system across the nation.
Proposal Budget: [REDACTED BY US EPA]