Ensemble locally weighted partial least squares as a just‐in‐time modeling method

The predictive ability of soft sensors, which estimate values of an objective variable y online, decreases due to process changes in chemical plants. To reduce the decrease of predictive ability, adaptive soft sensors have been developed. We focused on just‐in‐time soft sensors, especially locally weighted partial least squares (LWPLS) regression. Since a set of hyperparameters in an LWPLS model has to be set beforehand and there is only onedataset, a traditional LWPLS model is difficult to accurately predict y‐values in multiple process states. In this study, we propose to combine LWPLS and ensemble learning, and predict y‐values with multiple LWPLS models, whose datasets and sets of hyperparameters are different. The weights of LWPLS models are determined based on Bayes’ theorem, considering their predictive ability. We confirmed that the proposed model has higher predictive accuracy than traditional models through numerical simulation data and two industrial data analyses. © 2015 American Institute of Chemical Engineers AIChE J, 62: 717–725, 2016     

Classification of the Degradation of Soft Sensor Models and Discussion on Adaptive Models

Soft sensors are used widely to estimate a process variable which is difficult to measure online. One of the crucial difficulties of soft sensors is that predictive accuracy drops due to changes of state of chemical plants. It is called as the degradation of soft sensor models. In this study, we attempted to classify this degradation of models in terms of changes in an explanatory variable and an objective variable, and the rapidity of the changes. Moreover, we discussed characteristics of adaptive soft sensor models, based on the classification results. By analyzing simulated data sets and a real industrial data set, we could obtain knowledge and information on appropriate adaptive models for each type of the degradation. © 2013 American Institute of Chemical Engineers AIChE J, 59: 2339–2347, 2013     

Development of a Compact-Sized Falling Needle Rheometer for Measurement of Flow Properties of Fresh Human Blood

A compact-sized falling needle rheometer with rapid operation and automatic flow analysis has been developed for viscometry of fresh human blood without anticoagulant. The volume of a fresh blood sample only needs to be 3 mL, and the measuring time is within 2 min after taking a blood sample from the human body. Measured flow properties of human blood are evaluated as a flow curve, that is, the relationship between the shear stress (τ) and shear rate (γ). Observed flow curves of fresh human blood show three typical fluid regions, that is, the Casson fluid region for a low shear rate range of 0 < γ > 140 s^?1, the transition region for a shear rate near 140 s^?1 < γ < 160 s^?1, and the Newtonian fluid region for a high shear rate range of 160 s^?1 < γ > 400 s^?1. Flow properties of human blood such as the yield stress (τ _y) in the Casson fluid region and the apparent viscosity (μ) in the Newtonian fluid region are measured, and they are compared between male and female blood. It is found that the range of human blood viscosity for males is (5.5 to 6.4) mPa · s, and for females is (4.5 to 5.3) mPa · s. The viscosities of male blood without anticoagulant show higher values than those of female blood. Human blood viscosities with anticoagulant show a lower value than that without anticoagulant. A linear relationship between the hematocrit value, that is, the volume percentage of red corpuscles in the human blood, and the apparent viscosity are observed for both male and female blood. This article is concerned with the flow analysis of fresh human blood viscosity without anticoagulant using a newly developed compact-sized falling needle rheometer.     

Numerical simulations of annular extrudate swell of polymer melts

Numerical viscoelastic simulations were carried out using a K-BKZ type of separable integral constitutive equation. Both reversible and irreversible models were tried for several types of damping functions to calculate the annular extrudate behavior of high-density polyethylene (HDPE). There are two aims in this study; first, to clarify the properties of these dumping functions, and second, to investigate the influence of rheological characteristics on annular extrudate swell. In these numerical simulations, relaxation spectrum and shear viscosity were fixed, and the other characteristics were varied. The reversional response of the damping function mainly has an effect on the magnitude of the area swell even if the die is straight. The irreversible model expresses the experimental results of annular extrudate swell better than the reversible model. The accurate fitting of N1 by the damping model is important for predicting it. The magnitude of N1 predicted from the Wagner exponential model is lower than that of the PSM model, and the area swell shows the same tendency as N1. A modified PSM model that allows the N1 curve to shift can fit the magnitude of area swell. The relationship between the diameter and thickness of the extrudate depends on N2/N1, and it was estimated by simple linear elasticity of solids. The time dependent viscosity varies with the type of damping function, and it influences the time-dependent swell.     

POROSITY AND RESIDUAL EQUILIBRIUM SATURATION OF SINTERED SPHERES OF GLASS BEADS

Sintered spheres(made of glass beads) were simply produced with a definite diameter, with a precision of ±0·5 mm in diameter, for a wide range of sintered angles (0-27°). The experimental values of the porosity of the sintered spheres were consistent with the calculated results for the coordination number N ≠ 8. The residual equilibrium saturation, as one of the basic physical properties of sintered glass beads was investigated experimentally and theoretically, resulting in the presence of a distribution in the angle of pendular ring.     

A soft sensor method based on values predicted from multiple intervals of time difference for improvement and estimation of prediction accuracy

Soft sensors are widely used to estimate process variables that are difficult to measure online. However, their predictive accuracy gradually decreases with changes in the state of the plants. We have been constructing soft sensor models based on the time difference of an objective variable, y, and that of explanatory variables (time difference models) for reducing the effects of deterioration with age such as the drift without model reconstruction. In this paper, we have attempted to improve and estimate the prediction accuracy of time difference models, and proposed to handle multiple y-values predicted from multiple intervals of time difference. A weighted average is a final predicted value and the standard deviation is an index of its prediction accuracy. This method was applied to real industrial data and then, could predict more number of data with higher predictive accuracy and estimate the prediction errors more accurately than traditional ones.     

Complete amino acid sequence of chitinase-A from leaves of pokeweed (Phytolacca americana)

The complete amino acid sequence of pokeweed leaf chitinase-A was determined. First all 11 tryptic peptides from the reduced and S-carboxymethylated form of the enzyme were sequenced. Then the same form of the enzyme was cleaved with cyanogen bromide, giving three fragments. The fragments were digested with chymotrypsin or Staphylococcus aureus V8 protease. Last, the 11 tryptic peptides were put in order. Of seven cysteine residues, six were linked by disulfide bonds (between Cys25 and Cys74, Cys89 and Cys98, and Cys195 and Cys208); Cys176 was free. The enzyme consisted of 208 amino acid residues and had a molecular weight of 22,391. It consisted of only one polypeptide chain without a chitin-binding domain. The length of the chain was almost the same as that of the catalytic domains of class IL chitinases. These findings suggested that this enzyme is a new kind of class IIL chitinase, although its sequence resembles that of catalytic domains of class IL chitinases more than that of the class IIL chitinases reported so far. Discussion on the involvement of specific tryptophan residue in the active site of PLC-A is also given based on the sequence similarity with rye seed chitinase-c.     

Brefeldin A protects ricin-induced cytotoxicity in human cancer KB cell line, but not in its resistant counterpart with altered Golgi structures

Brefeldin A (BFA), an isoprenoid fungal metabolite, dramatically disrupts intracellular protein transport and protein secretion. BFA protects cells from the cytotoxicity of a plant toxin, ricin or pseudomonas toxin, but not that of diphtheria toxin (Yoshida et al., 1991. Expt. Cell Res., 192: 389-395.). In this study, we examined whether BFA could differentially change the cytotoxicity of ricin between BFA-sensitive cells and BFA-resistant cells. As a BFA-resistant cell line, we used a resistant cell line, KB/BF2-2, derived from BFA-sensitive human cancer KB cells. BFA treatment caused the disappearance of typical Golgi cisternae and the concomitant appearance of dilated vesicles in the cytoplasm in KB cells. By contrast, KB/BF2-2 cells had already altered Golgi structures with poor development of cisternae and also many vesicles in the absence of BFA, and BFA treatment did not further induce the morphological changes. Although a plasma membrane-specific marker protein, alpha-adaptin, was localized similarly in KB/BF2-2 as KB, Golgi specific markers such as beta-cop and gamma-adaptin were distributed in the cytoplasmic small vesicles as well as Golgi compartments in KB/BF2-2 cells in the absence of BFA, and the mutant cells showed no apparent changes in the distribution even when exposed to BFA. Ricin inhibited protein synthesis in KB and KB/BF2-2 to similar levels while pretreatment of KB cells with BFA at 0.1 microgram/ml almost completely reversed the inhibitory effect of ricin. By contrast, the pre-exposure of KB/BF2-2 cells to 1.0 microgram/ml BFA only partially rescued the ricin-induced inhibition of protein synthesis. Exposure to BFA at 30 min before ricin addition or at 0 min with ricin rescued the protein synthesis inhibition, but no rescue occurred when BFA was added 30 min after ricin addition. BFA could not rescue the protein synthesis inhibition by another toxin, diphtheria toxin. Our results suggest that BFA-resistant mutation causes a specific change in the endocytic membrane traffic of ricin in human cells, and also that cytotoxicity of diphtheria toxin does not share a common pathway of the intracellular transport with that of ricin.     

An Automatic Test-Generation System for Large Digital Circuits

A new test-generation system (FUTURE) for large digital circuits (more than 10K gates) is based on a nine-valued FAN algorithm. Fault simulation adopts a concurrent simulation adopts a concurrent simulation technique. The system consists of four major modules: fault modeling, random pattern generation, algorithmic pattern generation, and fault simulation. The system can be a powerful CAD tool and effectively generate test patterns for large sequential circuits with Scan Path.