Real-time artificial intelligence optimization systems have been shown to consistently reduce emission rates of fossil-fuel power plants. In this presentation, Comparison of State-of-the-Art Dynamic Machine Learning Methods for MPC of Coal-Fired Utility Generator Performance, ten different dynamic machine learning methods are analyzed and applied to a case study to determine the optimal model for enhanced coal-fired utility generator performance. Each method was compared on their ability to accurately and reliably predict NOx emission rates from a coal-fired power plant 60 timesteps into the future.

The most accurate and stable model for this application was the gated recurrent unit network (GRU), displaying an average RMSE of less than 5% across the complete time horizon, and a standard deviation of time horizon RMSE of less than 1%. 

See how the other models stacked up and watch a full presentation on how accurate each method was at predicting the NOx emission rate.

While operational support costs of using predictive machine learning models can be costly, the Griffin AI toolkit can be paired with the Power Plant Initiative, a maintenance program that reduces labor and maintenance costs. For a limited time save 80% on the Power Plant Initiative program with only a one year commitment or replace your existing system at a discounted installation rate. Try the Griffin AI Toolkit today or contact us to take advantage of the cost savings.

The post Comparison of State-of-the-Art Dynamic Machine Learning Methods for MPC of Coal-Fired Utility Generator Performance appeared first on Griffin Open Systems.

Numerous approaches to modeling and system configurations are discussed in the presentation, Long-Term Hybrid AI Expert Combustion Optimization System for Coal-Fired Electricity Generation NOx Reduction. Many of the approaches represent offline or only short-duration online performance studies and not integrated into a layer for real-time interaction with the control system. The hybrid AI-expert system aims to satisfy the need for a long-term evaluation and characterization of the performance and effects of a hybrid AI-expert combustion optimization system (COS) on NOx emission rates, as well as the secondary effects on the remainder of the system. The COS uses artificial neural networks to model the combustion process, particle swarm optimization (PSO) to optimize the 80+ manipulated variables, and expert logic to address adverse conditions in real-time using this adivarent control layer to leverage the existing digital control system assets for improved performance. 

Over a two-year operational period, the hybrid AI-expert system was able to aid in the reduction of the NOx emission rate by more than 22.5% compared to baseline, while improving unit temperature management and realizing high operator acceptance leading to a system service factor greater than 86%. 

Read more and watch a full presentation on the hybrid AI-expert system. 

The post Long-Term Hybrid AI Expert Combustion Optimization System for Coal-Fired Electricity Generation NOx Reduction appeared first on Griffin Open Systems.

Powdered activated carbon (PAC) injections for the removal of mercury (Hg) from combustion emissions can be costly and have adverse effects on other aspects of the combustion system – primarily exit gas stream opacity out of the stack.

It is preferable to inject only the minimum necessary amount of PAC to lower mercury below limits while not wasting PAC material and inflating opacity measurements. Common PAC injection control systems operate very slowly due to the nature of the adsorption process and the extremely low concentrations of mercury; however, the effects of injection can rapidly raise opacity, often beyond acceptable limits, requiring manual corrective actions. This often creates a cycle of control movements that culminate in the entire process performing inefficiently.

However, there exists an opportunity to intelligently balance priorities and optimize PAC injection rates to remove mercury while controlling opacity and minimizing PAC usage.

Advanced AI & No-Code Implementation for Rapid Results

The Griffin Open Systems’ Hg PAC Optimization application prioritizes multiple objectives in real-time to achieve optimal performance of your system while staying within all process limits. Through an advanced method of prioritization, long- and short-term objectives of multiple parameters are considered and respected, leading to overall improved performance.

The Griffin Open Systems’ Hg PAC Optimization application can be implemented in an agile manner. Requiring minimal I/O and leveraging its user-friendly no-code interface, operators and engineers don’t need to change code in any systems allowing the application to be implemented in as quickly as one week.

What to Expect from the Griffin Hg PAC Optimization Application

As the application can be implemented in an agile manner, engineers and operators can expect to see as high as a 15% – 20% reduction in PAC injection while lowering the amount of high opacity events and maintaining or improving mercury emission rates.

We at Griffin Open Systems look forward to aiding you in achieving optimal performance of your cooling towers as well as all systems within your process. Please contact us to learn more about our many solutions today!

The post Hg PAC Optimization with the Griffin AI Toolkit appeared first on Griffin Open Systems.

The highly complex and sensitive nature of combustion processes makes it one of the most challenging systems to control in a coal-fired power plant. Behaving almost as its own living entity with each process having its own individual characteristics and behavior, managing, and controlling such systems is a difficult task.

Griffin’s Combustion Optimization System (COS) uses many advanced AI methods available within the Griffin AI Toolkit to develop highly specialized and accurate models of each unique system to reduce emissions, improve heat rate, and achieve other process objectives.

Custom Combustion Optimization

There is no “one-size-fits-all” method to gaining real and repeatable improvements from combustion optimization systems. Griffin’s COS application approach has been developed and refined over more than 50 years of experience providing combustion optimization solutions. Building upon the proven Griffin COS template of reliable and robust control, each COS application is tailor-made to address the unique characteristics of the subject system.

Advanced AI modeling and optimization methods are used to represent process complexities and take advantage of each and every existing opportunity; from subtle damper adjustments to shifting whole operating conditions within acceptable ranges, all prospective improvements are leveraged to provide maximum performance benefits.

Agile No-Code Implementation

Existing site-knowledge from those who know your system best – your engineers and your operators, often spanning decades of operation – is then formalized and rapidly implemented within the Griffin no-code platform to further complement the advanced AI components in place within the COS.

This process results in a highly customized and fully relevant optimization system capable of providing control benefits to the system in as little as 1 – 2 weeks. As the AI components continue to learn and adjust from observing real-time operation, observed benefits continue to improve automatically over time.

What to Expect From The Griffin Combustion Optimization System

The Griffin COS applications are flexible and purpose-built so that they can be configured to realize any number of performance objectives. Common benefits include:

• Reduction of NOx emissions
• Heat rate improvements
• Mitigation of unplanned outages
• Enhanced alarm condition management & resolution
• Consistent unit controls across all operating crews and conditions

A Griffin COS is a valuable addition to any combustion process with a proven record of improving multiple aspects of process performance. Regardless of system complexity or existing limitations, benefits can be obtained through this powerful application. Please contact us today to learn more about your options!

The post Combustion Optimization with the Griffin AI Toolkit appeared first on Griffin Open Systems.

After the ESPs were cleaned, Griffin Open Systems & Neundorfer proposed setting up an artificial intelligence (AI) combustion optimization system could not only bring their emissions within regulatory limits, but with minimal use of their SNCR to avoid further costly ammonium bisulfate build-up.

Using the plants existing DCS, I/Os, and sensors, Griffin Open Systems integrated the Griffin AI Toolkit to begin the modeling phase. Without having to purchase additional hardware or make significant investments in operations technology, the model was able to implemented in a short timeframe.

The AI Toolkit enabled the combustion optimizer to run multiple models in parallel. These models were able to continually retrain & iterate on themselves as new data was collected. Using the AI Toolkit’s sophisticated algorithms, new data was easily incorporated into models which enabled improved combinations of variables to optimize the combustion. This, combined with the Griffin AI Toolkit’s rapid prototyping capabilities which allowed existing operator & engineer knowledge to be incorporated into the models with ease, lead to the discovery & modeling phase being completed in an agile and efficient manner.

The models from the discovery phase showed that optimizations from the AI Toolkit could reduce NOx emissions so dramatically, that the plant’s SNCR would not need to operate to keep within emissions compliance. Using data from these models, the plant was able to appeal to government regulators to run without use of their SNCR.

The regulators approved this, and the plant has been running the Griffin & Neundorfer Combustion Optimizer in a closed-loop real-time control fashion since late 2019. The AI Toolkit was able to develop these models & implement real-time closed-loop control so quickly, that the cost savings from avoiding derates & costly ESP cleaning paid for the implementation of the combustion optimizer within a year of implementation.

The post Griffin AI Toolkit Helps Utility Avoid Regulatory De-Rates & Lower Operating Costs appeared first on Griffin Open Systems.