Abstract Increasing environmental and economic awareness are driving the development of combustion technologies to efficient biomass use and clean burning. To accomplish these goals, quantitative information about combustion variables is needed. However, for small-scale combustion units the existing monitoring methods are often expensive or complex. This study aimed to quantify correlations between flue gas temperatures and combustion variables, namely typical emission components, heat output, and efficiency. For this, data acquired from four small-scale combustion units and a large circulating fluidised bed boiler was studied. The fuel range varied from wood logs, wood chips, and wood pellets to biomass residue. Original signals and a defined set of their mathematical transformations were applied to data analysis. In order to evaluate the strength of the correlations, a multivariate distance measure based on information theory was derived. The analysis further assessed time-varying signal correlations and relative time delays. Ranking of the analysis results was based on the distance measure. The uniformity of the correlations in the different data sets was studied by comparing the 10-quantiles of the measured signal. The method was validated with two benchmark data sets. The flue gas temperatures and the combustion variables measured carried similar information. The strongest correlations were mainly linear with the transformed signal combinations and explicable by the combustion theory. Remarkably, the results showed uniformity of the correlations across the data sets with several signal transformations. This was also indicated by simulations using a linear model with constant structure to monitor carbon dioxide in flue gas. Acceptable performance was observed according to three validation criteria used to quantify modelling error in each data set. In general, the findings demonstrate that the presented signal transformations enable real-time approximation of the studied combustion variables. The potentiality of flue gas temperatures to monitor the quality and efficiency of combustion allows development toward cost effective control systems. Moreover, the uniformity of the presented signal correlations could enable straightforward copies of such systems. This would cumulatively impact the reduction of emissions and fuel consumption in small-scale biomass combustion. / Tiivistelmä Kasvava ympäristö- ja kustannustietoisuus ohjaa polttoteknologioiden kehitystä yhä tehokkaampaan biomassan hyödyntämiseen ja puhtaampaan palamiseen. Näiden tavoitteiden saavuttamiseen tarvitaan mittaustietoa palamismuuttujista. Nykyiset palamisen seurantaan tarkoitetut ratkaisut ovat kuitenkin pienpolttolaitteita ajatellen usein kalliita tai monimutkaisia. Tässä työssä tutkittiin mitattujen savukaasun lämpötilojen riippuvuussuhdetta tyypillisiin kaasukomponentteihin, lämpötehoon ja tehokkuuteen. Tätä varten analysoitiin mittausaineistot neljästä erityyppisestä pienpolttolaitteesta ja suuresta kiertoleijupeti-kattilasta. Puupolttoaineina olivat klapi, hake, pelletti ja hakkuujäte. Analyysi tehtiin alkuperäisillä mittaussignaaleilla ja niistä matemaattisesti muunnetuilla signaaleilla. Riippuvuussuhteiden selvittämiseksi johdettiin informaatioteoriaan perustuva monimuuttuja-etäisyysmitta, jonka lukuarvolla mitataan signaalien samankaltaisuutta. Esitetty analyysimenetelmä sisälsi myös riippuvuuksien ajallisen muutoksen ja suhteellisten aikaviiveiden arvioinnin. Tulosten arvojärjestys perustui etäisyysmitan arvoon. Riippuvuussuhteiden samankaltaisuutta mittausaineistojen välillä vertailtiin 10-kvantiileilla. Analyysimenetelmän toimivuus vahvistettiin kahdella tunnetulla koeaineistolla. Savukaasun lämpötilojen ja palamismuuttujien mittaussignaaleissa oli samankaltainen informaatiosisältö. Vahvimmat riippuvuudet olivat muunnettujen signaalien yhdistelmillä pääosin lineaarisia ja palamisteorian mukaisia. Merkittävää oli, että tietyillä signaalimuunnos- ja palamismuuttujapareilla oli sama riippuvuussuhde kaikissa mittausaineistossa. Tämä todettiin myös simuloinneilla arvioitaessa savukaasujen hiilidioksidipitoisuutta lineaarisella, kiinteällä mallirakenteella. Mallin tarkkuus oli riittävä kolmella erityyppisellä kriteerillä jokaisessa mittausaineistossa. Tulosten perusteella signaalimuunnoksilla voidaan arvioida palamismuuttujia reaaliaikaisesti. Savukaasujen lämpötilojen potentiaali palamisen laadun ja tehokkuuden seurannassa mahdollistaa kustannustehokkaiden säätöratkaisujen kehityksen. Löydettyjä yleistettäviä riippuvuussuhteita hyödyntämällä niiden käyttöönotto lukuisissa polttolaitteissa helpottuisi. Pienpolton päästöjen ja polttoaineen kulutuksen vähentyminen olisi tällöin kumulatiivista.

An evaluation of several commercial adsorbents for use as contacting media in a process for combustion gas desulfurization was performed. Linde Synthetic Molecular Sieves types 13X and AW500 and Davison Silica Gel were the materials studied. The motivation for this investigation was based on the premise that a suitable method for removing sulfur dioxide from combustion gas streams has not been realized. Sulfur dioxide adsorption capacities were obtained for equilibrium (non-flow) and dynamic flow conditions. Sulfur dioxide adsorption on the molecular sieves could be described by a modified Langmuir expression of the form q/q_m = kc^1/n/ 1 + kc^1/n Calculated isosteric heats of adsorption were found to agree with measured data. Small scale dynamic studies consisted of contacting the adsorbents in a packed column with a simulated combustion gas. Evaluation of the adsorbent materials consisted of monitoring the effluent gas concentration after exposure to a step change in sulfur dioxide concentration. Sulfur dioxide loadings were greatest for the type 13X molecular sieve followed by AW500 molecular sieve. The effect of gas throughput was minimal which suggests that mass transfer was adsorbent side controlling. As gas temperature increased, sulfur dioxide adsorption decreased linearly for 100 percent sulfur dioxide concentration and non-linearly for low concentrations (0.003 percent). The effect of water vapor on sulfur dioxide adsorption capacity was determined by monitoring the effluent gas composition for specified sulfur dioxide-water vapor mixtures. Breakthrough time for sulfur dioxide was found to be an inverse function of the inlet water vapor concentration. For a typical combustion gas stream, (8 percent water vapor) the breakthrough time is roughly 10 percent of the water vapor free value. Based on the results obtained, a shallow bed (0.15m, 0.5 ft) of either type 13X or AW500 molecular sieve removed 5 to 3 times that of activated charcoal for a gas temperature of 57.2°C (135°F) and low gas pressure drop 4.6 cm H₂0 (1.85 in. H₂0). Adsorption degradation studies were performed to determine the loss in sulfur dioxide adsorption capacity after adsorbent regeneration. Both the 13X and AW500 molecular sieve could be regenerated, but the loss in adsorption capacity depended on the gas contacting conditions. X-ray spectroscopy was used to determine the homogeneity of the sulfur distribution within the adsorbents. The sulfur dioxide adsorption capacity for subsequent regeneration cycles was found to be a function of the Al₂O₃/SiO₂ ratio of the molecular sieve. Application of the molecular sieve adsorbents in a simulated combustion gas for sulfur dioxide removal was found to be superior to several adsorbents for the temperature range 21-76°C (70-170°F). However, as gas temperature increases, sulfur dioxide adsorption decreases. No adsorption of sulfur dioxide above 148°C (300°F) could be measured. / Ph. D.