Performance Testing - Science topic

Jani Rahkila I sent you a message with the raw data. Thank you for your help.

Compared to specimens with larger or non-equiaxed grains, ultrafine-equiaxed metal specimens offer the following advantages:

1. Improved Mechanical Properties:

A microstructure with ultrafine-equiaxed grains can have significantly improved mechanical properties, such as higher strength, ductility, and toughness. In equiaxed grains, the high density of grain boundaries results in reduced dislocation mobility and improved grain boundary strengthening.

2. Enhanced Wear Resistance:

In ultrafine-equiaxed grains, high grain boundary density can also contribute to enhanced wear resistance due to a higher resistance to crack propagation along the boundary.

3. Improved Fatigue Performance:

Having high-density ultrafine-equiaxed grains and their associated grain boundaries can improve fatigue performance by reducing crack initiation and propagation rates.

4. Better Formability:

In comparison to materials with larger grain sizes, ultrafine-equiaxed grains can exhibit improved formability because they can deform more uniformly and distribute strain more evenly throughout the material.

5. Reduced Component Weight:

With ultrafine-equiaxed grains, materials can be designed and manufactured to have the same strength and durability as larger-grained materials with better mechanical properties. Aerospace and automotive structures can benefit from significant weight reductions and fuel savings as a result.

As a result, metal specimens with ultrafine-equiaxed grains possess many advantages, especially in structures or load-bearing components, where high strength, ductility, and toughness are required.

Hello,

Not quite sure about characterization analyses but for CV, EIS, and other electrochemical techniques, you can check up this Gamry website: https://www.gamry.com/support-2/electrochemistry-courses-and-training/

Naima Ghemmit-Doulache thanks so much Dr for your answering.

But I want to compare Pt performance in cyclic voltammetry with TM-electrode, and I don't know which is better to use for this comparison Pt Nps electrode or Pt plate electrode which I have already. So I tried to find an answer to this question because I did not find until now a clear expansion for the reason to use Npe or plate.

It is suggested to carry out soil drying by air. Drying at 50°C is stronger and may affect, for example, some labile organic compound.

What was your mobile phase? What was its pH? Without knowing these, it is hard to imagine why you found gel formation. The column should have a pH range to work with. If you are within the range, then finding an answer will be difficult. At that point, you may wish to talk to the manufacturer of the column to learn more.

Dear Pia, thank you very much for your helpful recommendation!

To measure the FBP, an alternative method of determining the flat band potential is based on measurement of the net photocurrent as a function of applied potential. The flat band potential is predicted to be at the intercept of the square of the net photocurrent with the potential axis.

Hi

As there are various resources to be allocated to this requirement, please explain more about the amount of required computational power. Then, it is possible to suggest a solution

Regards

Ahmad

Dear Steve

following the prior comments. There is not definite answer to such a question - a fundamental question. Essential will be the objective of your work.

Do you want "just" find a catalyst that works well?

Do you want to understand how the reaction you are studying functions on a specific catalyst or group of catalysts?

Which feature of the catalyst are you interested in: atomic structure, micro-structure, etc.?

Knowing the answer to these questions may guide you to the methods you need for your work.

With respect to the two general approaches that you were raising: Theory vs. Experiments. I would like to handwavy point out some general issues.

Both approaches can be enormously time-consuming so there is no clear advantage of one over the other.

For theory, you got the essential issue, that you have to hypothesize some sort of structure beforehand. In your case, you could say: "I want to look at Ni-X(111) alloy (X = metal) surfaces." Then you may be able to predict which of the alloys has some advantage over the other. However, your synthesized catalyst will not be a pure Ni-X(111) surface, it will be a plethora of structures (metals, metal-oxide interfaces, nanostructures ...). So whatever you are predicting has the risk to be irrelevant for the catalysis. Same goes for the reaction mechanism, that you have to hypothesize beforehand. Yet, theory is extremely informative if it is used correctly, but this is not easy.

If you go down the experimental route, you face the same problem from another direction. Obviously your catalytic activity measurements will tell you which of the catalysts you are test is e.g. performing the best. However, your catalyst will be so complex and dynamic under reaction conditions that is hard to decipher what is the cause of the enhanced activity.

So there is no clear answer to this. However, in my opinion, essential to whatever you do is to read the literature in great detail. Look into work which was done decades ago as well as into the recent literature. This is by far the most time efficient way to understand what works and what not and what the scientific community is not understanding.

Maybe this helps, at least it helped me a lot.

Cheers, Arik

You can use both DSC and DMTA

Best regards !

You may enquire from Civil Engg./ Earth Sc. Depts. of various Indian Institutes of Technology (IITs) and National Institutes of Technology (NITs) by way of accessing their respective websites. Further, You may also contact through some private GT Labs. (e.g. Soiltech, Pune, India).

Hi,

Adult ADHD rating scale is available in open access.

To inject hho in a car engine, Amp reading helps too to have an idea of liter/min.

I would like to know how to use Tablet InhibitEx?

Dear Ahmed,

You can find several homogeneity tests in XLSTAT addinsoft tool (in EXCEL)

Zeljko Vujovic

I have replied you there ( https://www.researchgate.net/post/machine_learning_algorithms, as well as at https://www.researchgate.net/post/machine_learning_algorithms#view=5e800ad87068133e7e37252c )

Cheers,

Dr. Samer Sarsam

I raccomaded you an article: Multiple intelligences and minds for the future in a child's education - present on my reserchgate.net. I hope that it will be interesting for you. Zbigniew

Dear Vinajak,

it all depends on how much your material is bent. If the crurvature radius is so high that in the specimen can't be considered flat on your salmple holder, I'm afraid it is not.

If the curvature is negligible, XRD is perfect for surface phase analysis and, if your diffractometer can do it, surface residual stress measurement can be performed as well.

If you can cut out small pieces of material and make a single piece always centred by the X ray beam, XRD phase analysis is possible. In this case, however, residual stresses could not be possibly measured, as when you cut a stressed material you can release them.

Hi, Ge.

I have applied the DEAP dataset for EEG emotion classification before. The situation you mentioned also occurred in my research. From my perspective, here are some suggestions:

Firstly, Compared with the classification model you used, I think the input features are more significant. Maybe you could pay more attention about the feature extraction(Such as: PSD, based on frequency domain; HOC, on time domain; Discrete Wavelet Transform, on time-freq domain), selection and confusion parts, which was included the channel selection in terms of the different emotion categories.

Secondly, about overfitting issue, I thought it's kind of unsuitable to use the DNN model for this DEAP dataset unless you could increase the mount of data. Maybe you could give the data segmentation part(cutting the epochs) a shot.

Finally, in terms of accuracy problem, I suggest you double check which emotion estimation method was used in the specific paper. There are two aspects for EEG emotion classification: one is based on Valence-Arousal plane, which was learned from Speech emotion recognition; another is for the specific emotions(such as: angry, happy,sad,etc). So, be careful about the baseline you used to compare.

Regards

The purpose of reinforcement learning models is inferring some insights (quantitative and qualitative) from an empirical collection of various data to maximize some criteria of cumulative reward. These type of models do not use a causal explanation of any relationships between the variables (factors, features). In essence, this is a step by step empirical optimization in the reward loop.

In contrast, queuing based models (or better discrete event or systems dynamics simulation models) under the umbrella of operations research methodology offer the capability of testing multiple scenarios of behavior the complex variable systems based on mathematical / physical models of the systems. These type of models are based on understanding of the system to be modeled, i.e. they include some causal relationships and connections between the elements of the system.

The typical situation is when the validated simulation model first provides some 'what-if' scenarios outputs of the system performance. The optimization module then takes the output of the simulation model as the Objective function (OF), and, given a set of constraints, finds the set of the model's parameters (variables) that maximize or minimize the Objective function.

To summarize: reinforcement learning type models offer empirical data optimization without understanding causal relationship between the model elements.

In contrast, operations research methodology (queuing, discrete event of system dynamics simulation type models) include some causal relationships between the models's elements, i.e. includes understanding of basis of the model. This is a big advantage vs. a pure empirical learning.

@ Karl, please have a look of the attached file.

Paroma Deb Just do it in a sterile hood, I guess, use everything sterile. You can sterilize pretty much anything with UV.

1. Negative values have no issues but when you will transform into log it become unidentified or zero. So it will create problem as missing values.

2. time-invariant variable create problem as lag is taken in ARDL estimation and did not get lag values in time-invariant variable.

3. Depends on some tests.

if you are sure the data is approx. normal, you are good to go. I would suggest that you assign your participants to the intervention group and the control group at random. Best wishes, David Booth

Dear Manuele,

thanks for your kind reply. I know several methods for the evaluation of structure modification of proteins, the only thing I am concerned about is how to prepare a solution which can act as a "food simulant".

I think you can quantify glycine with and without derivatization by HPLC.

You may also try Repeated Measured ANOVA or Survival Test (for toxicity, germination test).

you won' be able to present SIINFEKL peptide to T cell using a U937 or THP-1 cells due to mismatched MHCI. The SIINFEKL is a mouse restricted ( H2Kb) peptide for T cell activation and there are no known human HLA's presenting SIINFEKL peptide. you may see a response with ovalbumin if there is a HLA binding epitope.

Probably they are biased due to some reason, could be anything from animal scent, cues in the test room etc. Try by increasing the time with close novel arm.

gd luck

It is possible, but highly unlikely base on current regulatory guidelines. For a given dose value, the human can have a better bioavailability and exposure than the animal model used preclinically. However, the preclinical models are dosed at several multiples until toxicity or maximum acheivable dose is reached. Human Ph1 trials start at a fraction of the NOEL reported preclinically. At these starting doses, exposures are compared to the animal data so guidance is available for the escalating dose studies. So it is unlikely that you would ever be in a position that human clinical trials were achieving exposures that were in excess of any reported for preclinical studies. If preclinical studies are not designed correctly, then it is a possible scenario that clinical doses could provide unexpected exposures but in that case the company has wasted much time and money and should rethink their position as a drug company.

Although Venkata is right you will easily run into difficulties in getting all the possible pure compounds. On top of it, developing a gc method to completely separate all theoretically possible reaction products is not that easy.

You'll gain a lot of time if you start using GC-MS for this project. Most GC-MS implementations use Electron Impact as ionization source witch yields a 'fingerprint mass spectrum' witch is matched against a library, this will leave you with an estimation of chance that a certain peak could be the proposed compound . An extra confirmation is still needed but the amount of references needed is smaller

Hello Evans. The trouble with testing for normality of the residuals is that normality is most important when the sample size is small, and becomes less and less important as the sample size increases (see slide 9 of the file linked below, for example). Your test of normality, on the other hand, has very low power when n is small and too much power as n becomes large. So tests of normality are just not very helpful. And notice, by the way, that normality of the errors is a sufficient, but not a necessary condition. It is the sampling distributions of the regression parameters that must be (approximately) normal for valid inference in OLS models. HTH.

Another thought is that you should always include viability dye. There are many available, which can overcome your overlap with PI. Depending on the lasers available on your machine, you could use DAPI or 7-AAD. These work the same as PI. Another option would be the fixable live/dead amine-reactive dyes. These are available in a variety of colors from eBioscience, BioLegend and Invitrogen. Good luck.

I also agree with the updated gating above.

Dear Naeem,

there is no need for sampling or transforming your raster data into vector format, or using expensive ArcGIS software.

Sampling is not good because you want to test correlation of entire rasters, rather than sparse pixels.

Instead, I recommend you to use open-source QGIS 2.18 (https://www.qgis.org/en/site/) with plugin called MOLUSCE (https://plugins.qgis.org/plugins/molusce/).

You should load entire rasters into QGIS, and load them into MOLUSCE as spatial variables. The next step is Evaluating correlation, where you can calculate Pearson's or Cramer's correlation coefficient or Joint information uncertainty. This document may help you with detailed instructions (https://github.com/nextgis/molusce/blob/master/doc/en/QuickHelp.pdf).

If some of the layers with spatial variables have high correlation (over 0.7), you should exclude one of them from your final model.

The other software is GRASS GIS (https://grass.osgeo.org/).

It has function r.covar that calculates correlation matrix between input rasters (https://grass.osgeo.org/grass70/manuals/r.covar.html). The computation time is much faster than in QGIS, but be aware that it is not user-friendly software.

Dear Lee Min Kee ,

You have asked two different questions, Firstly;

How much diameter should i consider when punching a cathode slurry?

I wonder if it should be smaller or larger than the Li metal anode.

Ans: As you are going to use 2032 coin cell so definitely, your anode is not going to be larger than 18 mm ( as a safety, otherwise it may touch cell walls) and if you will buy Li metal disks for anode, customarily they come in the sizes of 12 mm. So, the cathode should always be either to the dia of anode or smaller to have better anode to cathode contact, and recommended size is 8 mm in dia for the cathode.

Secondly,

Is there any difference in performance test depending on thickness? If so, what effect does it have on thickness?

Ans: In my experience, thickness of the slurry does have effect on the performance of the cell and that's why we do optimization. But remember, a UNIFORM and SMOOTH coating is always desirable for a good cathode. If material has low conductivity then a thicker slurry cam s;sluggish the electronic and ionic transfer and could also have effect on the wetting process of cathode by the electrolyte.

Hope it will be useful for you in the preparation of LiSB's .

Best of luck.

Cheers

there are more model for stress cree recovery on viscoelastic materials such as maxwell , burger

Dr Desmond Bala Bisandu

I'm not familiar to R, since I work with python (keras,tensorflow,...) .

I can find some functions for calculating ROC curves but that's not the problem.

I can't seem to find in any literature or documentation how the performance is measured when one anomaly is present in different windows for sliding window algorithms.

No experience with this cell line but many others. Cells grown in 96-well plate in 0.1 mL medium/well. Include a 'no cell' control. Add 20 uL/well of 5 mg/mL MTT reagent [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] in PBS (stored in aliquots at -20 C) and incubate 1-4 hrs. Microscopic examination should show dye accumulated in cells. Aspirate off medium and add 100 uL/well DMSO, mix ~15 min on an orbital shaker and read plate at 570nm. Getting the cell density right is usually the biggest challenge. Since the medium plus reagent is removed prior to solubilization in DMSO, it's prudent to visually examine the cells micrscopically beforehand to determine if cell loss will occur with medium aspiration. If the cells are not well attached to the well, then I'd use the MTS [ 3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] protocol. Good luck!

If this is a bentonite-cement slurry then you will have considerable shrinkage and drying if testing in anything other than the fully saturated state. Similarly, unconfined compression will lead to false readings. You need to consider triaxial testing of your mixes.

Hi,

crevice corrosion can occur everytime, when you have crevices ;-) Some materials are more prone to it than others. Sometimes you have an initiation at the edges that leads to crevice corrosion when pitting is stabilized. So if you see pits on the whole surface and also crevice corrosion, your experiment can still be ok. Surface preparation is also crucial. Do you sand your specimen before your experiments (grit size?) or passivate it? If so, how and how long?

I personally dont like teflon holders. Have you tried the Gamry pot-holes or any other adhesive films to separate your working electrode? Maybe this is working well also at 37°C, we use this technique at room temperature succesfully. Another possivility is the use of an Avesta-cell. In this cell, the transition bewtween cell body and working electrode is flushed slowly with chloride-free solution, preventing crevice corrosion. This is of course more complex and not every sample is suitable for it.

I would try adhesive films at first :). Be sure to press the edges firmly so that the adhesive comes out a little bit.

Best regards,

Andreas

There is a reasonable method based on simple , rapid and complete extraction of acids from an oil test portion into reagent .....This technique is based on indirect acid-based using special reagent. The reagent is similar to the reagent for the pH - metric. The presence of triethanolamine ( B ) causes rapid ( 1 min ) and full extraction of free fatty acids from the test oil portion into the reagent.In that case , the standard titration technique with colour indicator is not applicable. You can refer to : CROATIA CHEMICA ACTA. CCACAA 78 ( 1) 99 - 103 ( 2005 )

Yes Rishbach, you can do it directly from paper provided that the excitation beam match the absorption band(s) of the rare earth ions present in the paper. Therefore, you should first find the absorption spectra of the ions and then adjust your experimental conditions based on these spectra.

Best regards, Jiri

@Jeffrey thank you for your suggestion and I am looking for something that can do the same in our plant. Yes scaling up issues would need to be tackled one at time.

@Abhishek we have glass lined reactors in our plant which we can use, but I am afraid that it might make a silane coating on the glass too which might interfere the future application of the vessel for other projects. We dont have graphite vessels and dont want to invest extra specifically for this project

@Neetesh you link is for silanization of enamel for teeth and has no mention of large scale activity

Alright, so you have two samples, did you try to combine thing into one sample with a flag variable that separates them? What I am saying is make both samples one variable, then create a second variable coded as 0 for the first sample and then 1 for the second sample and then add the two distributions combined into one variable and have them grouped by the flag variable. I don't have SPSS anymore so I would just do it for you. I have altered your excel file to show what I mean.

So for the explore option, you would put the populations variable (the combination of the two variables in the dependent list and then use the flag variable (coded as 0 or 1) in the factor list. Have you tried this?

If this is still a problem you could use R if you are familiar. And if not I would be willing to run it for you.

Logan Netzer

Dear Amit,

I suggest you to see links and attached file in topic.

-Solving Rotated Multi-objective Optimization Problems Using ...

link.springer.com/content/.../10.1007%2F978-3-540-30549-1_74....

- An improved NSGA-III procedure for... (PDF Download Available)

https://www.researchgate.net/.../266656428_An_improved_NSGA...

- U-NSGA-III - MSU College of Engineering - Michigan State University

https://www.egr.msu.edu/~kdeb/papers/c2014022.pdf -Multi-Objective Test Problems, Linkages, and Evolutionary ...

Best regards

Hi Shadid,

Depending on sample size and normality, if you're thinking about a non-parametric test, then you might consider the Kruskal-Wallis method. Here's a descriptive link (see chapter 6):

have a great day!

--Adrian

Jones and Faure (1981 - Behavioural Processess): "Tonic immobility was induced in every test situation used in the first experiment, which underlines the robust nature of this phenomenon. However, although rigid standardisation of methodology may not be essential, we strongly recommend the use of the cradle and cloth for studies of immobility particularly when dorsal restraint is used. Not only was the reaction enhanced by this apparatus but the fact that tonic immobility was virtually always (19 out of 20) induced at the first attempt affords a high degree of consistency and reliability. The amount of handling necessary, with its possible stressful effects (Freeman, 1967; Candland et al., 1969), is thus reduced to a minimum. This high reliability may have been due to a purely mechanical effect in that the cradle provided greater support and restraint than the other substrates used."

Hope this helps.

Hi Kasper

Your Modification Indices is really the comparison you are looking for. They identify strain in the original model and tell you approx how much the overall model X² would decrease if the fixed or constrained parameter was freely estimated. Thus, MI's are analogous to the X² difference of nested models.

You seem to have followed a sound procedure. Check first if each item has a satisfactory factor loading (at least 0.4). Then check MI's for strain.

Instruments only tested with EFA are often reduced when they are evaluated with a CFA. as a CFA is a stronger analytical framwork.

See for example the procedures we followed to get the most parsimonous model in this article:

Best

jk

Hi Robert,

For Annexin V /PI staining it is recommended to measure the samples as soon as possible (we usually measure the samples within 30 minutes after staining). So for this staining fixation is not appropriate.

For other stainings we use 1-2% PFA in PBS to store the cells.

Please be aware that some fluorochromes are not stable and will not "survive" certain fixation methods.

Whenever we combine surface and intracellular stainings we use commercially available Fix+Perm kits.

Best

Wolfgang

good question I follow

Hi Adnan, this may help you..

De Winter, J.C., & Dodou, D.(2010). Five-point Likert items: t test versus Mann-Whitney -Wilcoxon. Practical Assessment, Research & Evaluation, 15(11),1-12.

parent. thank you

What you did was essentially correct. You said that "As I must test within each group, [...]", what actually means that the conditional distribution of the response should be (approximately!) normal. In a linear model the conditional distribution is just the distribution of the residuals. And in fact, statisticiand do a residual analysis to see if (how well) data and model go together.

With your subtractions you actually calculated the residuals of a full-factorial model. You did this quite manually, but that doesn't make it wrong.

I am only slightly concerned with your wish to "test" the normal assumption. A formal test won't answer your important question, namely whether the normal model is sufficiently good to model your uncertainty associated with the data. If the test is not significant, it may simply be "blind" w.r.t. relevant deviations from the normal model, and if it is significant it may only because of entirely irrelevant deviations from the normal model. Ideally, the normal assumption is justified on theoretical grounds (or prior experience/data). If you think the normal assumption should be ok, it doesn't harm to confirm this using a diagnostic plot of the residuals. In your case, a normal quantile-quantile plot would be good.

This paper should be very helpful for you. Use of Avicel will be better for plaquing flu.

There's many ways of measuring porosity, see this review:

If you have limited equipment, instead of the graduated cylinder method, you may get better results using a good analytical balance and using the Archimedes method. You weigh the sample dry, saturated, and submerged in a non-solvent liquid then you can calculate the volume of the open pores from the mass of wet scaffold minus dry and the bulk volume from the mass of the displaced liquid (saturated-submerged). The ratio will give the open porosity. You can calculate the total porosity by comparing the density of your material and the density of your scaffold (dry mass/bulk volume).

In fact, the post hoc of Ficher is the test that seems to have less influence in relation to the number of evaluated subjects. Thus, in my experience, it does not seem to require changes in statistics. But, of course, this is a superficial view. I

t is necessary to understand the totality of the methodological mechanisms of its research in order to have a more careful evaluation of the statistical analysis.

Hi!

You obtained some interesting results. First of all, your data for intact group is closer to well distribution. Data of experiment group is divided into at least two clasters (low (0.4-06) and high (0.8-1.2). Some of eggs gave no immune response. these are primary data. But! You need in calibration curve to evaluate real concentration of antibodies. This approach have to give you more clear results.

Regards,

Tatiana

Dear Neetika, 

You can try to write your own code for the problem using the very famous soft such as Maple.

Horak FB, in a study Objective biomarkers of balance and gait for Parkinson's disease using body‐worn sensors.He developed a metric for mobility assessment in Parkinson's disease, which includes: postural stability in stance, postural responses, gait initiation, gait (temporal-spatial lower and upper body coordination and dynamic equilibrium), postural transitions, and freezing of gait.  

Horak FB, Mancini M, Objective biomarkers of balance and gait for Parkinson's disease using body‐worn sensors. Mov Disord. 2013. 28: 1544–1551. pmid:24132842

Dear Sarah,

I understand that you have data with a mean that meets expectations ;-) however s is quite high!

The first thing you should always do is to look at the data: You plotted a histogram - good!

For selecting the right number of classes for the histogram, I'd recommend to chose an uneven number around SQRT(n). Does this look about symmetrical, monomodal (one highest value class), no surprises, i.e. no classes being higher populated than a population should be that is expected to be normal? No values that lie far outside from the apparent "curve"?

If so, fine, you can calculate the mean and s. If there are "outliers" you should investigate the reason for those and improve your measurement technique to replace them by good data. If there are no apparent outliers and everything is symmetrical, try to lower variability by standardizing your method. Imagine whic factors do have an impact on the result and estimate who big this impact is. Sort the factors by impact and address from the highest.

Remember: Confidence interval of the mean = mean plusminus t*s/SQRT(n).

So s=7 IS high, which hints to outliers or, as said, unacceptably high variability.

Hope this helps, good luck!

Holger

Vienna software as I found is for audio processing, not a network emulator.

The test I'm doing is for computer networks, to check the speed and the throughput of the network components.

Water vapour permeability test is essential for fresh produce packaging material which decides its transpiration and firmness. Also burst strength, puncture test to be carried out for packaging material for its durability.

Does the concept of "false positives" make sense at all?

Benjamini.Hochberg can control the FDR (false discovery rate - that's not a FWER!). It is surely a good option (much better than Bonferroni control of the FWER), but to my personal opinion I think that a meaningful application of hypothesis tests always requires to state a relevant effect size that is tested for (which I bet is not given and likely impossible for your task). Note that the top ranked comparisons will remain the top-ranked, no matter what correction you apply. I suggest you decide how many will be worth for further investigation and then simply select this number of top ranking candidates.

Superoxide dismutase (SOD) is an important antioxidative enzyme. It does catalyze the dismutation of the superoxide anion into hydrogen peroxide and molecular oxygen. Activity of SOD can be determined by a number of methods such as Pulse radiolytic methods (Rigo et al. 1975), Beauchamp and Fridovich (1971); Misra and Fridovich (1972); Tyler (1975). SOD can also be measured by inhibiting the enzyme of an O2- dependent reaction. A number of kits are also now available for measuring SOD.

However, in relation to ROS as well as its relation to microbial cells you can get some knowledge from my recent publications with CAB International, Oxfordshire, UK. There some interesting findings have been illustrated on ROS, antioxidants in relation to microbial cells, specific to AM fungi. I am also forwarding the published paper as it may help you.

Generally speaking for low-viscosity fluids you want to keep the test time very short and the test power relatively low - else you may introduce convection. I do most of my liquid testing with MTPS as opposed to TPS - but principles are similar. For liquids, MTPS keeps the test time <1s and the heater power very low. With liquids of similar viscosity to water, once you go above ~2-3s you start getting some degree of convection and after 7s convection starts in earnest, so a short test time is critical - as is making sure the liquid is fully stagnant prior to onset of testing.

If you have worked in Oil Field Chemicals/production chemicals, we never needed to exchange so many msgs. Despite of bad English (!!) I believe I was quite clear with my issue in first place, if you have answer I appreciate otherwise lets not waste time on understanding who wants what!! ..... 

Your question seems to be a combination of two questions:

1) Given two parameter sets, which one is better with respect to the KS test statistic? The first one gives a value of 0.0796, the second one 0.0921 (this is then reflected in the higher p-values for the first one (0.891) compared to the second (0.765) using the KS distribution, but this is not required in this case).

In this sense the first parameter set is the better one. This looks a bit puzzling at first, as the second set is the ML estimator, but keep in mind that different optimization criteria will lead to different parameter estimations.

2) How to do a KS test, if the parameters are fitted to the distribution? The question addressed is whether the data is coming from a Weibull distribution. If one uses fitted parameter values, the standard p-values of the KS distribution do not apply and one needs to use simulations. The result will depend in principle on the parameter estimation procedure used (MLE with bias correction in the paper mentioned above).

Using LcKS (Lilliefors corrected KS) from the "KScorrect" package in R, one gets a p-value of 0.35, so the hypothesis that the data is coming from a Weibull distribution is not rejected.

A final comment: Given the uncertainties of the estimated parameters I am not sure, whether the difference between the two is of any relevance.

Dear Mohamed Said,

First of all thank you to your time and detail writing with a needful attachment.

Please be informed that I am not trying to compare different model of gas turbine and find a single number of NHR. Instead, I would like to calculate the expected NHR and Net power output for n numbers of machines (installed in different location). I need this matrix of NHR and net power O/P for every plant to forecast the expected energy production and energy consumption. As the site performance test recent ISO values would be more conservative in terms of efficiency, I may under estimate the total energy consumption. Accordingly I am trying to correct the performance test recent data @ ISO conditions to an average temperature like 30DegC to see the expected net power generation and energy consumed as well. The intention is to do this without using individual plant machine type correction curves.

I welcome  any new ideas to do this as long as we doesn’t require all GT model correction curves and from the above specified sets of available data .

 Regards,

Dear Elsharif,

The National Instruments (NI) issued the attached file which gives a comparison between many types of the thermocouples including the maximum errors over

a full temperature range and typical errors at room temperature. 

Please have a look at the attached PDF file. 

Best Regards

Hello Rohit,

I think your problem already imposes the demand for simulations to be carried out at 1 atm pressure, so NPT would be required (assuming its periodic). To achieve the required pressure and temperature for long enough time such that the system pressure and temperature oscillations are very small around the desired values. During tensile deformation you can either use fix deform (with flag erate) to achieve a constant strain rate. The transverse directions can be relaxed by NPT @ 1 atm.

Caution: The simulation cell needs to be big enough that you don't have periodic boundary effect during deformation.

Apparently, not. The best advice is to use the open source progarm R, which has developed the test, or Eviews8, which incorporates directly this routine in its libraries. 

Allometric scaling provides a method for examining the structural and functional consequences of changes in size and scale among organisms.

Your question is best answered by Richard Harris's (1995, pp. 52-55 and 158) book, entitled Cointegration Analysis in Econometric Modelling. As Harris points out on p. 53, you should include a constant and/or trend either in the cointegrating regression or in the ADF regression, but NOT in both. On p. 158, he provides MacKinnon's (1991) table of critical values. On p. 54, he cites Bruce Hansen (J. of Econometrics, 1992, v. 53, pp. 87-121) to argue against including a time trend in the ADF regression, irrespectively of whether the residuals from the cointegrating regression contain a trend or not, because the inclusion of a time trend leads to loss of power to a test that is already of low power. So, if I were you, I would include a trend in the coin. regression, but not a constant, and in the ADF regression, I would include only a constant (no trend). Of course, if the time trend in the coin. regression seems to be insiginifiant, drop it. Note that n stands for the total number of variables in the cointegrating regression, including the left-hand side variable, but excluding constant and trend, so the case of n=1 on the top parts of MacKinnon's Table given on p. 158 of Harris's book, as well as the case of N=1 in Table 2 on p. 157 of the Engle and Yoo paper, 1987, are simply reprints of the top part of Fuller's (1976, p. 373) Table 8.5.2 for testing a variable for a unit root.

If you used the generalized translog cost function, you should also test it against more restrictive technological forms.  Attached is an example I've done with a co-author with the referred tests illustrated in Table IV.  I hope that this helps.