Iron oxide was selected as the desulfurizing material, and its reactivity with H₂S-H₂-H₂O mixtures was studied by TGA. In particular, the effect of the partial pressure of the water-vapour was investigated.
When the partial pressure of water-vapour during the simultaneous reduction and sulfurization was less than 0.3 atm, the fractional sulfurization reached unity, giving rise to FeS_1.02±0.1 at 800 or 900°C. At these low water-vapour concentrations, the sulfurization rate increased as the partial pressure of water-vapour increased. As the temperature became lower than 700°C, however, the fractional sulfurization did not reach unity.
As the partial pressure of water-vapour became higher than 0.35 atm, the fractional sulfurization was decreased significantly even at 900°C. When hematite powder was prereduced to magnetite, the fractional sulfurization was also decreased, even at low watervapour concentration.
The weight change during simultaneous reaction in H₂S-H₂-H₂O mixtures at 800 or 900°C was found to be estimatable by simply adding the two curves obtained for the reduction and the sulfurization, respectively.

View full abstract

In addition to the conventional measurement of selective adsorptive capacity, heterogeneous density including a liquid mixture and an adsorbent was measured to obtain the amounts of each component adsorbed in micropores of molecular sieving adsorbents. Liquid adsorption equilibrium data are presented for mixtures of watermethanol, waterethanol, and methanolethanol on Molecular Sieve 4A at 25°C. The adsorbed liquid volumes, which are calculated from the amounts adsorbed, are linear with respect to the volume fraction of adsorbed phase within experimental error. For aqueous alcohol mixtures the adsorption characteristics are different from that of methanol and ethanol mixture, which may be attributed to the behavior of water molecules capable of entering into small β-cages of the adsorbent. Adsorptions of pure water and methanol were measured by use of 2-propanol as a solvent and the results were compared with the data of gas-phase adsorption. The agreement of the two kinds of data indicates that the liquid-phase adsorption results in this work are reliable.

View full abstract

Little has been known of the flow characteristics in the transition region of free shear flow. However, flow in this region must contribute very much to mixing in the turbulent region, since entrainment from the main flow to the mixing layer is large enough even in the transition region.
The following results are derived in this phenomenological investigation. The disturbance growth rate in the two-stream mixing layer is 1.7 times that of the one-stream mixing layer, since mean flow profiles of the two-stream layer are of a velocity defect type just after leaving the splitter plate edge and therefore include two of the inflexion points. The fundamental mode of the disturbance in a natural transition as in this work is coincident with the modified linear theory proposed by Miksad. Non-linear growth of disturbance continues until it gains maximum turbulent energy, which is equivalent to the velocity difference between main flows. Then it turns into roll-up motion, which can be confirmed by flow visualization. It is found that fluids in not only the high-velocity side but also the low-velocity side are entrained into two-dimensional vortical motion after rolling up. Therefore, this vortex is a double-rolled structure which gives rise to sub-harmonic modes in spectral measurement. It is also recognized that this vortical motion induces very much entrainment, at a rate equal to that in the fully developed turbulence region. Thus premixing is finished in this transition region.

View full abstract

Cellular thermal convection in a wedge-like vessel, heated from horizontal lower plate and cooled from inclined upper plate, is visculized, and effects of the inclination angle of the upper plate and of the aspect ratio on the onset of thermal convection and on the cellular pattern are investigated experimentally.
As the inclination angle is increased, cellular convection is induced at lower value of local Rayleigh number, independently of the aspect ratio.
Cellular patterns can be classified into two groups : longitudinal cellular patterns, which consists of roll cells with their axis parallel to the dirdction of inclination, and transversal cellular patterns, where roll cells are arranged perpendicularly to the direction of inclination.
These patterns are predominantly controlled by global conditions such as the aspect ratio and the angle of inclination.
The width of a roll cell is not dependent on the kind of cellular pattern but on local conditions such as the local depth of fluid layer.

View full abstract

Flow behavior around a cone rotating in a coaxial outer cylinder at rest has been observed by flow visualization technique. Under the rotational speed of the cone, flow is laminar circulation in one part of a working space, and a Taylor vortex is formed in the other part. Such stable coexistence of different flow regimes in a working space induces peculiar phenomena in flow behavior.
According to the flow behavior observed, seven distinct kinds of flow regimes have been identified and the corresponding regions are shown in the flow diagram, of which coordinates are Reynolds number defined by the angular velocity of cone and the radius of outer cylinder as the representative quantities, and the relative radius defined by the ratio of local radius of cone to that of outer cylinder.
The typical flow regimes are as follows : I. laminar circulation; II. laminar Taylor vortex with interface travelling towards the laminar region I in which a penetrating vortex is viscous dissipated; III. Taylor vortex with stationary interface and with hysteresis effect on the number of vortices in working space; IV. turbulent Taylor vortex with deformed interface travelling towards the most dissipative region; V. spiral turbulence rushing towards the fully turbulent region VII; VI. unstable flow of which the regime cannot definitely be characterized; VII. fully turbulent flow.
The criteria of Taylor transition in the case of a rotating cone are defined by the local value at the boundary of regions I and II, and coincident with the critical Taylor number to the rotating inner cylinder.
In the region II mentioned above, the shift of vortex interface is represented by the equation which is similar to a Lagrangian expression of effluent flow from a point source in a two-dimensional field, being independent of Re and cone geometry in the generalized coordinate. Furthermore, the phase shift of the right-and left-hand vortices has been measured and phenomenologically correlated with the specified values in the generalized coordinate. Based on this correlation, the size of vortex depending on location and Re has been estimated and compared with the measurements.
The torque of rotating cone and inner cylinders hasb een directly measured. Assuming the local cylinder approximation for a cone, the torque of the cone has been estimated from the correlation to the cylinder. The estimations agree with the measurements.

View full abstract

The lower and upper critical gas velocities determining the stable operating range and the liquid and gas hold-ups were measured in two columns, 50 and 150 mm in diameter, having relatively large hole diameters (7 to 13 mm) and free areas (0.2 to 0.61).
With increasing gas velocity, the three typical gas-liquid flow patterns appeared on the plate, i.e., spray, cellular foam and froth. Cellular foam approximated the flow regime for stable operation.
Column diameter, liquid viscosity and suspension of fine solid particles exerted almost no influence on either the critical velocities or hold-ups. With an increase in surface tension the former increased, whereas its effect on the latter was negligible.
By assuming an interaction between gas and liquid streams through holes, critical velocities based on the free area were correlated in a simple dimensionless expression. The correlation presented reproduces well the literature data covering a wide range of variables for sieve trays as well as turbo-grid trays.
The dimensionless clear liquid height based on the hole gas velocity was represented as a simple function of the ratio of liquid to gas mass flow rates. The gas to liquid hold-up ratio was also rearranged.

View full abstract

The effects of secondary flow as twin-fluid and forced vibration atomization on the disintegration of a kerosene jet flow issuing from a capillary nozzle in water were experimentally studied. The increase of secondary-flow velocity decreased the drop diameter and promoted the formation of uniformly sized drops within the laminar jet-flow region. The breakup length of the laminar jet as a function of the relative velocity of main flow to secondary flow was semi-empirically obtained by using the previous analysis of jet stability. When a forced vibration was applied to the jet, the breakup length was easily decreased and uniformly sized drops were formed in quantity. The experimental results revealed conditions under which drops of good mono-dispersity could be produced.

View full abstract

The mixing of two fluid streams by using single-and dual-jet injections in turbulent pipe flow was studied to clarify the optimum condition for rapid mixing. With four types of dual-jet injections, i.e. injection angles of 0 rad, π/6 rad, π/4 rad and sin^-1 (1-d/D) rad, characteristics of mixing were compared with each other to obtain the optimum injection angle and the optimum velocity ratio for each injection angle. An improvement obtainable in mixing by using dual-jet injections was clarified by comparing the optimum condition and the characteristics of mixing with single 0 rad and sin^-1 (1-d/D) rad jet injection. The optimum condition is discussed on the basis of one-dimensional analysis and an empirical trajectory of the deflected jet.
It was found that the optimum velocity ratio for dual 0 rad injections was the same as that of single 0 rad injection, and that the mixing distance of the former was about half of the latter because of isolated development of the two circular jets. A comparison of the increase in calculated flow rate of jet at its termination point predicts the optimum condition for 0 rad jet injection over a wide range of pipe diameter ratios. For injection angle of sin^-1 (1-d/D) rad due to an interaction of the two wall jets the optimum condition for dual-jet injections is smaller than that of the corresponding single-jet injection. On the other hand, the dual π/6 rad and π/4 rad jet injections have the same optimum velocity ratio as the 0 rad jet injection. The injection angle of π/6 rad is recommended as the best angle for rapid mixing by using dual-jet injections.

View full abstract

Flow characteristics of solids-liquid-gas mixtures in vertical upward flow were investigated experimentally. Water-suspended spherical glass beads (0.151.0 mm in diameter) and ion exchange resin (0.5 mm in diameter) were conveyed through transparent acrylic pipes (18.3 and 30.3 mm in inner diameter and 5.5 m in length) and air was blown into the slurry from the bottom of the vertical pipes by a compressor. Data were obtained at delivered volumetric concentration of solids in slurry up to 0.1, superficial slurry velocity up to 2 m/s, superficial gas velocity up to 0.25 m/s and void fraction up to 0.4. Holdup was in agreement with the Hughmark-type correlation using the mean density, ρ_sl, and the superficial velocity, u_slo of the slurry. The solid concentration was estimated by the experimental equation of Ohashi et al. when u_slo was used instead of mean slurry velocity of two-phase flow. The annular separation flow model was used to explain the pressure drops of threephase flow in the vertical conduits. In addition, the pressure drops also could be calculated by the application of the Lockhart-Martinelli type of correlation for two-phase flow, when the pressure drops of slurry with velocity u_slo were used instead of the pressure drops provided by superficial water velocity of two-phase flow.

View full abstract

The characteristics of heat transfer in a semitransparent laminar fluid flowing in an open channel, heated uniformly by a radiant heat source from the upper side, are investigated theoretically and experimentally.
Kinetic and energy equations are solved by using the idea of “Cold Semitransparent”, while the radiant properties are estimated from two models : the Multiband Model and the Planck Mean Model. Using an infrared lamp as a radiant heat source, and water and diethylphthalate as a working fluid, the temperature distribution within the fluid layer and the collection efficiency are measured under various experimental conditions.
As for collection efficiency, both the models explain similar tendencies, but concerning temperature distribution the result calculated from the Multiband Model shows better agreement with the experimental data than those from the Planck Mean Model, regardless of the optical thickness of the fluid layer.

View full abstract

Process fluids slowly penetrate into plastic materials of construction of chemical equipment during the running period, while additives and residual monomer of the construction materials are often extracted by these fluids.
Several models were studied from the viewpoint of comparable diffusion rate of solvent and solute, considering the experimental results of extraction by dilute nitric acid of lead contained in rigid PVC. A model based on the assumption of particulate dispersion of solute at the specimen surface is suitable for expressing instantaneous loss of solute and subsequent course of the early stage in extraction, but it does not fit the results of long-term experiments. An alternate model based upon internal particulate dispersion of the solute accompanied by mass transfer resistance around the particles is applicable to delayed extraction recognized in the abovementioned experiments.

View full abstract

The extraction of copper from aqueous solution by LIX65N-kerosene solutions containing Span80 as an emulsifying agent was carried out using an agitated vessel in which the oil phase was dispersed. The reaction between copper and LIX65N at the oil-water interface was found to be retarded by the addition of Span80, owing to the preferential adsorption of emulsifying agent rather than LIX65N at the interface. When the concentration of Span80 was greater than 1 wt % the apparent rate constant for the extraction with chemicalreaction was smaller than one-tenth that in the absence of Span80. For the LIX65N dispersion system in the agitated vessel, it was shown that the extraction process in the presence of the emulsifying agent is controlled by the chemical reaction at the interface, and therefore the mass transfer resistance is of no significance.

View full abstract

Hexane solutions of long-chain alkyl secondary amine were equilibrated with aqueous solutions of carboxylic acid. Chemical species formed were estimated and equilibrium constants for these reactions were evaluated.
It was found that these extraction equilibria could be interpreted by use of the following reaction mechanisms :
1) Mono-carboxylic acids (HA) react with amine (B) to form ion-pairs such as (BH) ⁺A^-.
2) The ion-pairs further bind to acid-forming (BH) ⁺A^- (HA) _n-1 type species. Values of n are 2, 3, 4 for propionic acid and acetic acid and 2, 3 for formic acid.
3) In the case of the reaction with formic acid, (BH) +A-B, (BH) 2⁺A₂^- and ((BH) ⁺A^- (HA)) ₂ type species are formed in addition to the above-mentioned ones.
4) Aqueous oxalic acid (H₂A) forms ion-pairs such as (BH) ₂+A₂^- reacting with amine.

View full abstract

Interfacial tension between aqueous solutions of carboxylic acid and hexane solutions of long chain alkyl amine, Amberlite LA-2, in equilibrium state were measured for various concentrations at 303 K. The experimental results were analyzed on the assumption of adsorption of amine, acid and expected chemical species by the extraction equilibrium result onto the interface.
The adsorption equilibrium phenomena could be interpreted by the model based on Langmuir's adsorption and Gibbs' adsorption isotherm. Utilizing this model, interfacial adsorption equilibrium constants and interfacial area occupied by adsorbed species were evaluated.

View full abstract

The devolatilization of various coals with carbon content between 35 and 86 wt% (d.a.f.) was performed by means of a batch reaction furnace in the range of devolatilization temperatures of 300 to 1100°C. The effects of carbon content of original coals and devolatilization temperatures on the yields and compositions of devolatilization products were examined over a wide range of reaction conditions.
The composition of gaseous product is largely different between low-rank coals (lignite and brown coal) and high-rank coals (bituminous coal). The gas from the former contains a relatively large percentage of CO and CO₂, while the latter produces the gas of a comparably high content of hydrocarbons. The temperature effect on the H₂ production has a similar dependency for all the coals employed. The gross heating value of gaseous product from low-rank coals increases with increase in temperature and reaches a constant value (about 17, 000 kJ/Nm³) above about 800°C. On the other hand, for high-rank coals the gross heating value decreases exponentially with increase in temperature. The yields of both CO and CO₂ have a maximum at some carbon content of coals. For example, at 1000°C the maximum was observed at about 65 wt %. Above 800°C, the char-yield is scarcely affected by the devolatilization temperature, ranging between 45 and 60 wt% depending on type of coal. The gaseous sulfur compounds evolved include not only H₂ but also SO₂. The value of SO₂/ (SO₂+H₂S) increases with decrease in carbon content of original coals.

View full abstract

Using changes in light transmittance and electrical conductivity of solution with elapsed time, and also the size distribution of final product crystals, the impurity effect of several amino acids (L-phenylalanine, L-leucine, L-valine) on the formation and growth of L-glutamic acid nuclei by secondary nucleation in agitated solution was studied kinetically.
The following results were obtained : 1) Amino acids as impurity are easily incorporatedin the crystal in the following order at the same conditions : L-aspartic acid<L-valine< L-phenylalanine=L-leucine. 2) These impurities suppress the rate of formation and growth of nuclei and make γ_min larger, where γ_min is the radius of the smallest among crystals breeding secondary nuclei. 3) These impurity effects are stronger for the impurity which is incorporated with greater difficulty.

View full abstract

The mechanism of crystal growth of ADP was studied by the experimental results of linear growth rate (accumulating velocity) on (010) and (011) faces.
It was recognized that the growth rate of (010) faces fitted NaN theory, but for (011) faces the growth mechanism transited from BCF theory to NaN theory at relative supersaturation 0.02. In spite of this analytical result, phenomena that really take place on (011) faces could not be considered so dramatic a change and, therefore, it is suggested that : (1) two kinds of section, growth rates of which fit NaN or BCF theories, subsist simultaneously on (011) faces and (2) they interchange with each other the leading role of crystal growth with slight variation of growth environment.
The effect of impurity on the growth mechanism was also studied using AlCl₃, which has property of being a habit modifier.
The growth rate of (011) faces was not affected by the presence of AlCl₃, but that of (010) faces was influenced significantly. The critical concentration of AlCl₃ for growth rate was recognized. The decrease in growth rate seemed to be caused by the formation of macrosteps and not by interruption of the creation of two-dimensional nuclei.

View full abstract

Correlative equations among operational conditions of a continuous well-mixed model crystallizer are obtained from consideration of cumulative amount of product crystal, for systems whose linear growth rates can be expressed by Bransom equations and volumetric shape factors are a power function of crystal size. A new chart for considering operational conditions is devised from these equations.Correlation among nucleation rate, growth rate, size of product crystal, suspension density, production rate per unit volume of a crystallizer, power number of Bransom equations and power number of crystal size for volumetric shape factor is expressed in this chart. Additionally, a new simple method of decision for power number in Bransom equations is proposed.
Tests were carried out for crystallization of urea in urea-methyl alcohol solution using a laboratory-scale continuous mixed crystallizer with impeller. Growth rates were correlated with Bransom equations, and volumetric shape factors of urea crystal were expressed by a power function of crystal size. And data obtained were plotted on the proposed chart.

View full abstract

The inclusion mechanism of impurity and habit modification for o-chlorobenzoic acid (OCBA) and benzoic acid (BA) crystals precipitated from OCBA-BA-toluene system were investigated.
The relationship between crystal and solution compositions was obtained by differential crystallization. As the result of experiments, a large amount of BA was found to be included in OCBA crystals, although in BA crystals only a small amount of OCBA was included.
The washing effect of toluene on crystal impurities was examined by using benzyl alcoholas a tracer for mother liquor. It was clarified that BA is included in OCBA crystal structure and cannot be washed by toluene. However, OCBA in BA crystal may be due to adhesive mother liquor on the crystal surface.
Furthermore, powder X-ray diffration measurement was carried out for OCBA crystal obtained by differential crystallization and the diffraction pattern was compared with that of a mechanical mixture of OCBA and BA. As the result of comparison, it is suggested that BA is included in OCBA crystals by forming solid solution with OCBA. The shape of OCBA crystals (pillar-shaped) was deformed by inclusion of a large amount of BA. On the other hand, the shape of BA crystals (also pillar-shaped) was largely changed to scalelike form, even if only a small quantity of OCBA is present in the solution. This change may be due to the habit modification effect owing to the adsorption of OCBA on BA crystal surface.

View full abstract

Effects of turbulence on instability of a leading flame edge of a diffusion flame formed in a boundary layer over a flat plate with uniform fuel injection have been studied experimentally. Two different types of stability limit are identified. One is observed when the air flow approaching the leading flame edge is laminar. The stability limit of this type depends both on the approaching air flow velocity and fuel injection velocity. The other type is observed when the leading flame edge is strongly affected by the turbulence. In this case, the stability limit depends on neither approaching air flow velocity nor fuel injection velocity but only on turbulent intensity in the approaching air flow. When the leading flame edge becomes unstable due to turbulence, the turbulence appears to affect not the chemical reaction but mainly the diffusion process.

View full abstract

LaNi₅ powder having four mean particle diameters as cumulative particle size percentage 50%-3.4, 7.0, 18.0 and 40.5μ-was used for the anode of H₂-O₂ fuel cell. In the range of mean particle diameter 7 to 40.5 μ the slope of the square root of overvoltage and current density, √η/I, and the reciprocal of limiting current density, 1/I_L, were correlated with the mean particle diameter. Fine pores in the range 0.03 0.4 μ and relatively large pores in the ranges 0.5 1 μ and 4 5 μ were found to have an important effect upon the electrode characteristics.

View full abstract

Techniques for the removal of sulfur oxides (SO_x) and nitrogen oxides (NO_x) have progressed, especially the wet absorption method for SO_x and the dry catalytic reduction method for NO_x. However, for effective removal of SO_xand NO_x the development of a new idea and process to economize on energy cost and to advance the performance is urged. Thus, the purpose of this study is to establish a dry process based on the principal that NOx is re moved by catalytic reduction with NH₃ over metal sulfate, which is formed by sulfation of SO_x with metal oxide. Experiments were carried out by using a flow-type packed bed reactor and a thermobalance under atmospheric pressure at 623K. The inlet gas was 1, 000 ppm SO2-500 ppm NO-667 ppm NH3-5% O₂-10% H₂O-N₂. The various metal oxides (catalysts) were prepared by the coprecipitation of the different raw salts with alkaline aq. As the resutls, 40 mol% CuO-2O mol% Al₂O₃-40 mol% TiO₂ prepared from CuSO₄, Al₂ (SO₄) ₃, TiCl4 and NH3 aq. was found to be the most suitable metal oxides combination for dry simultaneous removal of SO_x and NO_x Additionally, the establishment of regeneration of catalyst employed is the most important. It was found that the catalyst could be regenerated by reduction (CuSO₄→Cu₂O) with NH₃ in the presence of water vapor, followed by mild oxidation (Cu₂O→CuO) with O₂-N₂ mixture.

View full abstract

In a 1 m-square gas-fluidized bed of O.184-mm silica sand particles experimental data were obtained on jet size and initial bubble diameter above the gas distributor, along with radial and vertical distributions of bubble characteristics such as the size, frequency and rising velocity. Jet height, initial bubble diameter and bubble size were compared with existing experimental and theoretical equations, more of which are based on data from smaller-scale fluidized beds.

View full abstract

The simple method of measuring water droplet sizes already proposed was applied to the size measurement of organic liquid particles as fuel mist.
This method requires that the liquids have good electrostatic conduct and non-polarity. To apply this method for the measurement of organic liquid particle sizes, which have high electric resistivity and polarity, it is necessary to correct the influence of such qualities on the size measurement. These influence can be predicted theoretically.
As the result of experiments on size calibration varying electric resistivity and dielectric constant of organic liquids, the validity of the correction method based on theory was con-firmed.

View full abstract

The principal component analysis method was applied to the classification of particle shape of Mullite beads. The parameters measured from particle silhouette were shape indices of elongation, circularity, etc. and Fourier coefficients of angular bend function. Some of the shape indices were correlated among them and four clusters of them were obtained on the correlation space. However, the Fourier coefficients were mutually independent.
Eight parameters which can be defined for non-holomorphic shapes were selected and analyzed similarly. The first principal component was found to relate to the elongation and the second to the roughness of profile. The contribution of the two components to total variance was 84%. Two groups of Millite bead separated in shape by an inclined rotating disk were tested. The space of the two principal components was effective for shape classification.

View full abstract

Deposited dust layer on fabric filter was dislodged by normal stress induced by impaction acceleration. Polyester woven fabric and polyester felt fabric pressed at high temperature were used as filters. Kanto loam powder (JIS-8) was used as test powder. Filtration velocity was kept constant at 4.5 cm/s and impaction acceleration was controlled from 3.4 × 9.8 to 220 × 9.8 g m/s² by dropping height. The dislodging mechanism was classified into the three states of initial dislodging, progressive dislodging and final state of dislodging and was evaluated by the collecting condition (deposited dust load) and the dislodging condition (impaction acceleration).
It was found from these results that the dislodging phenomena are not evaluated by the dislodging force expressed by production of deposited dust load and acceleration alone. These results and observations on the dislodging phemomenon show that not only the dislodging force but also the external work performed on the deposited dust layer play in important role in the dislodging mechanism.

View full abstract

To make clear the flow around agitated turbine impellers having varying blade width, tomographic observations and measurements were carried out with a rotating camera and a plane-light source.
The following results were obtained. 1) The flows (helical flow etc.) found in the fanturbine were also observed in all the turnines used here. 2) The shear rate and angular momentum in the region of discharged helical flow were greater than in other region. 3) The power consumed in the impeller region was 1636% of power input. But the rate of energy dissipation per unit volume in the impeller region was 1020 times as much as the rate in vessel. The rates in impellers of B/D= 1/4, 1/5 (B : width, D : diameter of impeller) were greater than in other impellers. 4) A model of flow profile in the helical flow was proposed.

View full abstract

Aerated mixing vessels are often used in chemical and biological industries and the thermal control of liquid sometimes produces significant effects.
The present authors have proposed the method of estimating the capacity coefficient of gas absorption in aerated mixing vessels using the two-region model, where it is assumed that the region of bubbling-controlling cndition and that of agitation-controlling condition coexist in the aerated mixing vessel.
In this paper the two-region model is applied to estimate the heat transfer coefficients to non-Newtonian liquid in the aerated mixing vessel. The calculated heat transfer coefficients show good agreement with the experimental values.

View full abstract

Many problems in process systems engineering can be reduced to that of solving a set of non-linear algebraic equations. This system of algebraic equations is usually given as a model for presenting the performance of an unit process or the basic equations for designing a chemical process system. Moreover, most optimization problems can also be reduced to that of sloving a system of algebraic equations. Often, obtaining their solutions is the most difficult and important part of the problem. The most difficult problem is how to find the independent variables to which arbitrary values can be assigned. The next is how to most effectively obtain all the solutions of other variables depending on them.
In this paper, a new method is presented to resolve these problems by introducing new concepts such as “the normalization operator”, “a local coordinate system”, and so on. Also, by giving a system of algebraic equations their directed graphical representation, a structural analysis of the equations is performed. In the latter half of this paper, the method is applied to a model for presenting the performance of a distillation column.

View full abstract