Recipient Organization
Enertechnix, Inc.
PO Box 469
Maple Valley,WA 98038
Performing Department
Research and Development
Non Technical Summary
Combustion of biomass, especially in small-scale applications, produces high emissions of particulate matter (PM) that have been linked to adverse health effects and global climate change. Current industrial particulate control solutions are prohibitively expensive for use in small scale biomass burners. To overcome these challenges, inexpensive and effective combustion control solutions for small scale applications are needed. The proposed intelligent combustion control system for biofuel combustion in small scale applications can curb PM and gaseous emissions. If successful, the proposed pollution control system will be a disruptive technology and will be low cost enough to be implemented in thousands of small scale installations.
Animal Health Component
75%
Research Effort Categories
Basic
25%
Applied
75%
Developmental
(N/A)
Goals / Objectives
The proposed project will demonstrate the feasibility of developing an intelligent combustion control module to maximize the efficiency of biomass combustion while minimizing emissions of PM, CO, UHC and NOx. Our hypothesis is that by measuring the temperature, levels of main combustion product species, and particulate matter, a predictive model can be developed to intelligently control combustion of biomass fuels of varying composition, moisture content, and formats. This is possible due to a combination of recent advances in sensor development and implementation of novel modeling techniques, e.g. (i) real-time exhaust measurements using low cost sensors, (ii) real-time prediction of combustor conditions using CRN (chemical reactor network) modeling, and (iii) an intelligent combustion control algorithm for minimizing pollutant emissions and maximizing combustor efficiency. Development of an intelligent combustion control system relies on the ability of the CRN to model combustion processes in the critical pollution formation zones in the biomass combustor. A CRN model will be constructed based on the results of CFD simulations; the CRN model will establish the degree of modeling complexity (detailed chemistry and network arrangement) needed to predict the emissions. A combustion control algorithm for reduction of PM2.5 and NOx emission will be developed; this algorithm will be applicable to a wide variety of biofuels. The algorithm will (i) record sensor measurements, (ii) interpret inputs using a predictive CRN model, and (iii) adjust the wood burner controls to minimize pollution formation. This control strategy will be implemented in a commercially available biomass combustion system.Phase II will be focused on achieving the following technical objectives:1. CRN Optimization and Validation for a Variety of Fuels and Burn Scenarios2. Experimental Model Validation3. Sensor Package Ruggedization4. Automated Control Algorithm Development and System IntegrationExtensive Laboratory Testing using the GCE Platform
Project Methods
Task1 CRN Optimization using Computational Fluid Dynamic SimulationIn Phase II we will employ a CFD-CRN methodology in which detailed CFD simulations will be used to guide the development of a much more complete CRN model. Similar to previously developed CRNs for other biomass combustion the furnace will be subdivided into zones based on the characteristic chemical process prevailing in those regions, such as: main flame, recirculation zone, over-fire air zone, and a burnout zone. The overall approach involves first developing a CFD model of the combustor. The resulting flow field, temperature field, and reaction rate field are used to develop the CRN model that includes the detailed chemistry, including the trace pollutant species that are the target of the work.The CFD model focuses on the fluid dynamics. For this portion of the work, we will not develop a coupled wood devolatilization/volatile flame model due to the complexity associated with the two-way coupling between the hot flame gases and the wood (this complexity arises principally due to the complex reaction/transport processes inside the large pieces of wood). Instead, we will use measured devolatilization rates (via weight loss) to treat the wood surface as a boundary condition for the gas-phase CFD calculation. Since the principal output of the CFD work is the CRN model, this approximation is both appropriate and also more accurate than trying to include the very problematic solids-gas coupling.The successful outcome of this task would be development of a fast predictive CRN scheme that captures air recirculation mechanism and can be applied for flaming and smoldering combustion regimes. This task may be complicated by the effects of different wood composition and moisture content, wood shape and size, and different pyrolysis and tar oxidation chemistry.Task 2: Experimental Model ValidationIn this task, we will use state-of-the-art combustion diagnostic instruments at UW combustion facility to obtain additional information about the combustion process. The stack measurements from our low-cost sensor array will be compared to the boiler gas, particle and 3-D temperature profile scans. A GCE biomass boiler will be installed at the UW Combustion lab. The boiler will be modified in order to perform probe measurement and visual inspection of the boiler cavity. Optical access to the combustion cavity will provide flow field visualization and visual identification of burn stages: ignition, smoldering or flaming combustion, char oxidation stage.A gas sample probe will be inserted into the boiler though an array of ports; the position of the probe will be varied to cover the entire 3-dimensional space of the combustion cavity and cover all of the combustion zones (pyrolysis layer, main flame, burn out zone and exhaust stack). Suction pyrometry will be used for gas measurements.Temperature measurements will be performed using several thermocouples: the wall temperature will be measured by thermocouples embedded into the refractory lining. Shielded thermocouples will be used for gas temperature measurements.The measurements obtained in the biomass combustor will be used to validate the CFD simulations and guide the CRN development. A successful outcome for this task will be to obtain measurements of particulate and CO emissions for several burn scenarios, capturing high-PM and CO-emission events in-situ.Task 3 Sensor Package Optimization and RuggedizationThe existing field prototype of the system relies on low cost measurements of CO and particulate matter concentration in the combustion exhaust stack and temperature in the relevant combustion zones. A sample-dilution system was built in order to operate gas and PM sensors in their linear regimes. The sample and dilution air flows are metered by small stainless steel tubing with suction for both lines provided by a small diaphragm pump. One of the most important tasks is to optimize this dilution system; this task will include extensive testing of small diameter flow metering tubes in high particle loading environments. Several different flow metering schemes such as venturi pumps and limiting orifices will be tested at a variety of flow rates providing dilution ratios in the 1:10 to 1:200 range. Several exhaust dilution system prototypes will be built and their fouling propensity in the particle laden exhaust gas will be evaluated. Repeatability and fouling propensity of the sensing module will also be evaluated.The performance of the ruggedized sensor arrays will be tested in a combustion environment to provide information about sensor-to-sensor measurement repeatability, sensor longevity in the particle laden environment, and the need to use filters and water traps.A successful outcome for this task will be low (preferably no)-maintenance operation of the dilution and sensing systems for periods consistent with end-user needs.Task 4 Automated Control Algorithm Development and System IntegrationIn the Phase 1 project we have manually operated the GCE biomass burner air-staging to influence the CO and PM emissions. It was found that algorithms based solely on temperature are limited in predicting and controlling combustion in a complex-burn situation; in particular, we have identified two specific scenarios that lead to very high emissions and low combustion efficiencies due to the incomplete pyrolysis and combustion in the primary flame zone.Optimization algorithm: Spatial information needed to optimize the combustion cannot be obtained from global temperature measurements; however it is available as an output of the CRN model. The complete species, temperature and combustion rate information from the wood pyrolysis, flame, and burnout zones will be analyzed by the algorithm and optimized to control air staging (controlled by three independent solenoid valves) and overall flow rate (controlled by the fan speed). The input to the model will be based on the real-time flow conditions and sensors (CO, PM and Temperature) input. The primary output of the algorithm will be correction to the fan speed and the opening or closing of the solenoid valves for introduction of primary and secondary air resulting in adjustment of the combustion air staging.The algorithm will be integrated with the existing temperature-based control algorithm used in the GCE burner. GCE will assist with integration of the newly developed software and their air-staging and fan speed controls.A successful outcome for this task will be implementation of a sensor-based control algorithm in the GCE biomass burner to control combustion air staging.Task 5 Extensive Laboratory Testing using GCE PlatformThe integrated algorithm will be extensively tested with variety of fuels, fuel form factors and fuel moisture levels. The testing will be performed at UW facility using an instrumented GCE biomass test unit (see task 2). At a minimum, the test will include the following wood species and formats: (1) Oak crib (2) Oak logs (3) Douglas fir crib 4) Douglas fir Logs. These two wood species are significantly different with respect to their composition and represent two different geographic regions of the US.A successful outcome for this task will be (i) maintenance-free operation of the unit for extended periods and (ii) the ability of the algorithm to optimize combustion of multiple wood species, form factors and moisture contents.