Demonstration of leapfrogging for implementing nonlinear model predictive control on a heat exchanger.

This work reveals the applicability of a relatively new optimization technique, Leapfrogging, for both nonlinear regression modeling and a methodology for nonlinear model-predictive control. Both are relatively simple, yet effective. The application on a nonlinear, pilot-scale, shell-and-tube heat exchanger reveals practicability of the techniques.

Cascaded process model based control: packed absorption column application.

Nonlinear, adaptive, process-model based control is demonstrated in a cascaded single-input-single-output mode for pressure drop control in a pilot-scale packed absorption column. The process is shown to be nonlinear. Control is demonstrated in both servo and regulatory modes, for no wind-up in a constrained situation, and for bumpless transfer. Model adaptation is demonstrated and shown to provide process insight. The application procedure is revealed as a design guide to aid others in implementing process-model based control.

Comparison of model-based and conventional controllers on a pilot-scale heat exchanger.

This pilot-scale heat exchanger demonstration compares two relatively simple nonlinear model-based control strategies to conventional proportional-integral (PI) control. The two nonlinear controllers, generic model control (GMC) and process-model based control (PMBC), use a first-principles model thereby providing characterization of the nonlinear process throughout the operating range. There are two approaches to GMC, one uses a dynamic model, the other a steady-state model. This work uses the steady-state model; accordingly, will use the term GMC-SS, which can be classified as output characterization for a PI controller, making it relatively simple to implement. PMBC uses a dynamic model and adapts to represent the process. These two nonlinear controllers were selected for this application evaluation because of their simplicity (they can be implemented in-house within many commercial control systems), diversity (steady-state and dynamic models), and demonstrated utility for control of nonlinear single-input-single-output processes. The application and results are presented and discussed. Summarizing the results: Within a small temperature operating range PI provides good control, but over the full operating range, the nonlinear and variable delay of the process lead to poor control with PI. GMC can handle the nonlinear issues, but using the convenient steady-state model; it also, provides poor control because of the variable delay associated with flow rate. PMBC was able to provide good control throughout the entire operating range. PMBC has a further advantage of only having one tuning coefficient, while PI and GMC-SS have two.

Assessing viscoelastic properties of chitosan scaffolds and validation with cyclical tests.

We evaluated and modeled the viscoelastic characteristics of chitosan and chitosan-gelatin scaffolds prepared using a freeze-drying technique. Chitosan and chitosan-gelatin solutions (0.5 and 2 wt.%) were frozen at -80°C and freeze-dried. Using the scaffolds, uniaxial tensile properties were evaluated under physiological conditions (hydrated in phosphate buffered saline at 37°C) at a cross-head speed of 0.17 mms(-1) (10 mm min(-1)). From the break strain, the limit of strain per ramp was calculated to be 5% and the samples were stretched at a strain rate of 2.5%s(-1). The ramp-and-hold type of stress-relaxation test was performed for five successive stages. Chitosan and chitosan-gelatin showed nearly 90% relaxation of stress after each stage. The relaxation behavior was independent of the concentration of chitosan and gelatin. Also, changes in the microstructure of the tested samples were evaluated using an inverted microscope. The micrographs acquired after relaxation experiments showed orientation of pores, suggesting the retention of the stretched state even after many hours of relaxation. Based on these observations, two models (i) containing a hyper-elastic spring (containing two parameters) and (ii) retaining pseudo-components (containing three parameters) were developed in Visual Basic Applications accessed through MS Excel. The models were used to fit the experimental stress-relaxation data and the parameters obtained from modeling were used to predict their respective cyclic behaviors, which were compared with cyclical experimental results. These results showed that the model could be used to predict the cyclical behavior under the tested strain rates. The model predictions were also tested using cyclic properties at a lower strain rate of 0.0867%s(-1) (5%min(-1)) for 0.5 wt.% scaffolds but the model could not predict cyclical behavior at a very slow rate. In summary, the pseudo-component modeling approach can be used to model the sequential strain-and-hold stage and predict cyclical properties for the same strain rate.

A power law approach to orifice flow rate calibration.

Although standards for orifice flow meter design, installation, and calibration are supported herein, noncompliant devices exist in many pilot-, lab-scale, and on-board applications. For these, a common calibration practice is to preserve the ideal square root relation and determine a device specific discharge coefficient value. This work provides theoretical and empirical analyses to support relaxing the square root relation between orifice pressure drop and flow rate for noncompliant devices. The resulting power law relation is shown to improve accuracy, precision, and rangeability. Whether a device specific square root or power law model is used, it requires off-line or in-line calibration data. As such, a power law calibration model may only be useful for on-board and small-scale applications.

Evaluation of a statistically based controller override on a pilot-scale flow loop.

This work investigates the use of a statistically based override on an automatic controller. The objective is to temper controller action, and to obtain a better balance of variance reduction in both controlled and manipulated variables than is currently obtained with conventional filtering. Experimental results on a pilot-scale flow loop demonstrate the benefit. Results are supported by both simulation and theoretical analysis.

A statistically based filter.

A simple procedure is developed that uses concepts from statistical process control to filter noise from a process variable. Code is presented for the algorithm, and it is experimentally demonstrated on a pilot-scale process for which both noise and level change significantly and frequently.

Grouped-neural network modeling for model predictive control.

A group of feed-forward neural networks (NNs), each providing the prediction of an individual process output at a future step, is used as the dynamic prediction model for the model-based predictive control (MPC) scheme in the proposed work. These NNs are parallel (independent) rather than cascaded--they are trained and implemented in parallel. Therefore, the complexity and effort in the training stage is decreased and compounded error propagation is eliminated from the prediction. A new strategy of compensating for the process-model mismatch under this grouped-NN model structure is also developed. Effectiveness of the scheme as a general nonlinear MPC is demonstrated by simulation results.

Treatment of no-reflow and impaired flow with the nitric oxide donor nitroprusside following percutaneous coronary interventions: initial human clinical experience.

OBJECTIVES: The objective of this study was to test the hypothesis that the intracoronary administration of a direct donor of nitric oxide is a safe and effective method to treat impaired blood flow (no-reflow phenomenon) that occurs during percutaneous transluminal coronary interventions (PTCI). BACKGROUND: The absence of blood flow or decreased blood flow in a coronary artery following PTCI despite the presence of a patent epicardial vessel or graft is designated "no-reflow" or "impaired flow." This alteration in blood flow is a serious complication of percutaneous revascularization strategies that results in an increased incidence of morbidity, myocardial infarction and mortality. METHODS: Nineteen consecutive patients undergoing standard percutaneous revascularization procedures complicated by either no-reflow or impaired flow that received intracoronary nitroprusside treatment were studied. One patient had two procedures performed on two separate grafts on two successive days. Interventions were performed on either saphenous vein grafts or native vessels and utilized angioplasty, stent deployment or rotational atherectomy strategies. Following interventions that were associated with impaired flow, varying total doses (of nitroprusside 50 to 1,000 microg) were administered into the coronary artery or saphenous vein graft. The angiographic archives before and after intracoronary administration of nitroprusside were analyzed for TIMI grade flow and a frame count method was used to quantitate blood flow velocity. RESULTS: Following a PTCI that resulted in either no-reflow or impaired flow, nitroprusside (median dose 200 microg) was found to lead to a highly significant and rapid improvement in both angiographic flow (p < 0.01 compared with pretreatment angiogram) and blood flow velocity (p < 0.01 compared with pretreatment angiogram). No significant hypotension or other adverse clinical events were associated with nitroprusside administration. CONCLUSIONS: The direct nitric oxide donor nitroprusside is an effective, safe treatment of impaired blood flow and no-reflow associated with PTCI. The use of nitroprusside to treat syndromes secondary to microvascular dysfunction may provide a novel therapeutic strategy for treating no-reflow or impaired blood flow following percutaneous interventions.

Costimulation of CD8alphabeta T cells by NKG2D via engagement by MIC induced on virus-infected cells.

NKG2D is an activating receptor that stimulates innate immune responses by natural killer cells upon engagement by MIC ligands, which are induced by cellular stress. Because NKG2D is also present on most CD8alphabeta T cells, it may modulate antigen-specific T cell responses, depending on whether MIC molecules--distant homologs of major histocompatibility complex (MHC) class I with no function in antigen presentation--are induced on the surface of pathogen-infected cells. We found that infection by cytomegalovirus (CMV) resulted in substantial increases in MIC on cultured fibroblast and endothelial cells and was associated with induced MIC expression in interstitial pneumonia. MIC engagement of NKG2D potently augmented T cell antigen receptor (TCR)-dependent cytolytic and cytokine responses by CMV-specific CD28- CD8alphabeta T cells. This function overcame viral interference with MHC class I antigen presentation. Combined triggering of TCR-CD3 complexes and NKG2D induced interleukin 2 production and T cell proliferation. Thus NKG2D functioned as a costimulatory receptor that can substitute for CD28.

Dynamic optimization of hybridoma growth in a fed-batch bioreactor.

This study addressed the problem of maximizing cell mass and monoclonal antibody production from a fed-batch hybridoma cell culture. We hypothesized that inaccuracies in the process model limited the mathematical optimization. On the basis of shaker flask data, we established a simple phenomenological model with cell mass and lactate production as the controlled variables. We then formulated an optimal control algorithm, which calculated the process-model mismatch at each sampling time, updated the model parameters, and re-optimized the substrate concentrations dynamically throughout the time course of the batch. Manipulated variables were feed rates of glucose and glutamine. Dynamic parameter adjustment was done using a fuzzy logic technique, while a heuristic random optimizer (HRO) optimized the feed rates. The parameters selected for updating were specific growth rate and the yield coefficient of lactate from glucose. These were chosen by a sensitivity analysis. The cell mass produced using dynamic optimization was compared to the cell mass produced for an unoptimized case, and for a one-time optimization at the beginning of the batch. Substantial improvements in reactor productivity resulted from dynamic re-optimization and parameter adjustment. We demonstrated first that a single offline optimization of substrate concentration at the start of the batch significantly increased the yield of cell mass by 27% over an unoptimized fermentation. Periodic optimization online increased yield of cell mass per batch by 44% over the single offline optimization. Concomitantly, the yield of monoclonal antibody increased by 31% over the off-line optimization case. For batch and fed-batch processes, this appears to be a suitable arrangement to account for inaccuracies in process models. This suggests that implementation of advanced yet inexpensive techniques can improve performance of fed-batch reactors employed in hybridoma cell culture.

Broad tumor-associated expression and recognition by tumor-derived gamma delta T cells of MICA and MICB.

Human MHC class I-related molecules, MICA and MICB, are stress-induced antigens that are recognized by a subset of gamma delta T cells expressing the variable region Vdelta1. This functional association has been found to be limited to intestinal epithelium, where these T cells are prevalent and where MICA and, presumably, MICB are mainly expressed. However, increased frequencies of Vdelta1 gamma delta T cells have been observed in various epithelial tumors; moreover, MICA/B are expressed on diverse cultured epithelial tumor cells. With freshly isolated tumor specimens, expression of MICA/B was documented in many, but not all, carcinomas of the lung, breast, kidney, ovary, prostate, and colon. In tumors that were positive for MICA/B, the frequencies of Vdelta1 gamma delta T cells were significantly higher than in those that were negative. Vdelta1 gamma delta T cell lines and clones derived from different tumors recognized MICA/B on autologous and heterologous tumor cells. In accord with previous evidence, no constraints were observed in these interactions, such as those imposed by specific peptide ligands. Thus, MICA/B are tumor-associated antigens that can be recognized, in an apparently unconditional manner, by a subset of tumor-infiltrating gamma delta T cells. These results raise the possibility that an induced expression of MICA/B, by conditions that may be related to tumor homeostasis and growth, could play a role in immune responses against tumors.

Dynamic reoptimization of a fed-batch fermentor.

Traditionally, fed-batch biochemical process optimization and control use complicated models and off-line optimizers with no on-line model adaptation and re-optimization. This work demonstrates the applicability, effectiveness, and economic potential, of a simple phenomenological model for modeling, and a novel optimizer for on-line re-optimization and control of an aerobic fed-batch fermentor.

A method to determine the required number of neural-network training repetitions.

Conventional neural-network training algorithms often get stuck in local minima. To find the global optimum, training is conventionally repeated with ten, or so, random starting values for the weights. Here we develop an analytical procedure to determine how many times a neural network needs to be trained, with random starting weights, to ensure that the best of those is within a desirable lower percentile of all possible trainings, with a certain level of confidence. The theoretical developments are validated by experimental results. While applied to neural-network training, the method is generally applicable to nonlinear optimization.

Modulation of Cm/T, G/A, and G/T triplex stability by conjugate groups in the presence and absence of KCl.

Apparent equilibrium association constants were determined by gel mobility shift analysis for triple strand formation between a duplex target containing a 21 base long A-rich homopurine run and several end-modified C(m)/T (pyrimidine motif; C(m) = 5-methylcytosine), G/A (purine motif), and G/T (purine-pyrimidine motif) triplex-forming oligonucleotides (TFOs). Incubations were carried out for 24 h at 37 degrees C in 20 mM HEPES, pH 7.2, 10 mM MgCl2, and 1 mM spermine. The purine motif triplex was the most stable (Ka = 6.2 x 10(8) M-1) even though the TFO self-associated as a linear duplex. Conjugation of a terminal hexanol or cholesterol group to the G/A-containing TFO reduced triplex stability by 1.6- or 13-fold, whereas an aminohexyl group or intercalating agent (acridine or psoralen) increased triplex stability by 1.3- or 13-fold. These end groups produced similar effects in C(m)/T and G/T triplexes, although the magnitude of the effect sometimes differed. Addition of 140 mM KCl to mimic physiological conditions decreased stability of the G/A triplex by 1900-fold, making it less stable than the C(m)/T triplex. The inhibitory effect of KCl on G/A triplex formation could be partially compensated for by conjugating the TFO to an intercalating agent (30-350-fold stabilization) or by adding the triplex selective intercalator coralyne (1000-fold stabilization). Although the G/T triplex responded similarly to these agents, the stability of the C(m)/T triplex was unaffected by the presence of coralyne and was only enhanced 1.4-2.8-fold when the TFO was linked to an intercalating agent. In physiological buffer supplemented with 40 microM coralyne, the G/A triplex (Ka = 3.0 x 10(8) M-1) was more stable than the C(m)/T and G/T triplexes by factors of 300 and 12, respectively.

Factors influencing the extent and selectivity of alkylation within triplexes by reactive G/A motif oligonucleotides.

G/A motif triplex-forming oligonucleotides (TFOs) complementary to a 21 base pair homopurine/homopyrimidine run were conjugated at one or both ends to chlorambucil. These TFOs were incubated with several synthetic duplexes containing the targeted homopurine run flanked by different sequences. The extent of mono and interstrand cross-linking was compared with the level of binding at equilibrium. Covalent modification took place within a triple-stranded complex and usually occurred at guanine residues in the flanking double-stranded DNA. The efficiency of alkylation was dependent upon the sequence of the flanking duplex, the solution conditions, and the rate of triplex formation relative to the rate of chlorambucil reaction. Self-association of the TFOs as parallel duplexes was demonstrated and this did not interfere with triple strand formation. With an optimal target, cross-linking of the triplex was very efficient when incubation was carried in a physiological buffer supplemented with the triplex selective intercalator coralyne.