Algorithm for Authomatic Generation of Calibration History and Analysis

Dear Author, I would like to know if this software can be applied to any type of instrument, especially the instruments that can measure several magnitudes. In addition, if the certificates have different measurements for the same function and range, how the algorithm organize the data? Regards

Dear Author, Thank you very much for the presentation. My question is about the certificates from wich the algorithm takes the data. All the certificates have to have the same format or the algorithm searchs in every certificate and detects the tables in order to take the values? Thank you

Dear Marcos Yes, it can be applied to any type of instrument. You have to train it, to obtain the expected values.

Dear Lucas, It does not need to have the same formats, for the example find the values from 2012 to now, and where there are certificates with different formats and different ways to report them.

>>MAURICIO JAVIER ALBERINI PLEASE READ THIS COMMENT << We are expressing our doubts in this platform. The idea is to have topics to talk about and questions to answer during the "presential/virtual session" and answer them during the Workshop using Microsoft Team (14th, 15th and 16th of July 2020). As moderator, I will "interview you" using the questions generated during this week by participants in what we named: The Exchange in plenary. We will have 1 hour of exchange for Automation of laboratory processes topic (it means that in 1 hour we need to cover those 6 videos with questions and answers)

>> IMPORTANT MAURICIO JAVIER ALBERINI PLEASE READ THIS COMMENT << Due to great success of the questions and answers section during the current week. And because we are having just a few minutes during the plenary to cover all comments or questions. We decided to change the rules a little bit. We will be pleased if you could continue answering questions through the comment section (this include CENAM's platform or YouTube's comments section). Excellent job! My thanks to you and all in helping get as far as we have during this event! Regards, Angel Ochoa

Mauricio, thanks for your presentation. In my understanding, you have now experience of handling the problems raised by the fact that measurement results and data are documented in different ways and formats. Are you in contact and exchange with the colleagues that are developing the Digital Calibration Certificate as they are standardizing the format to store and document measurement data?

Dear Alexis, I had no contact with them. The problem I am solving is for cases of non-standard formats, to make it easier to build a history for an instrument. With a standardized format this problem would not exist.

Structured and standardized database development will enhance digitization of metrology.

Yes having a well-designed database is essential. As well the format. For all that I have been testing a JSON structured database it is ideal for numerical analysis.

I find this work very interesting. In my case I have all the history done by hand, which is why I am thinking more about the future. A good idea might be to standardize or generate guidelines for how a report should be put together to make it easier to retrieve information for history. Do you see this possible? On the other hand, I would like to ask you if this system could be applied to certificates from other NMIs, mainly written in English. Thank you very much!

Hello Alejandro, thank you very much! It is not necessary to standardize the reports, so that it works, and if it would also work for certificates in English, you just have to train it with that language.

Muy interesante trabajo, especialmente para los laboratorios bajo acreditación ISO 17025.

En cuanto a la “biblioteca” de datos, para ser consultada, ¿qué requisitos de confidencialidad fueron tenidos en cuenta, al momento de trabajar con diferentes marcas de equipamiento?.

¿Este software puede ser aplicado a instrumentos de control de masa o de temperatura?

Great, we are both of us working almost in the same area. Once I needed to create something similar but looking for something into PDF files. Did you have any problem with encrypted files? Is the search process focused on text, structure or images? The database is great! I am currently working with something similar to that concept. Are you considering to share those results with other laboratories in INTI using a filter to show them those that are theirs or by ID? (This could be useful to add that data into their software to avoid introducing values by hand or to check if the certificates are up-to-date). Temperature could be something that your coworkers could easily introduce into their code (I amm sure I could thank you if I were developing a software in LATU and you could share those results to let me use them directly into my software) Are any limitation or is there, generally speaking, no other classification option for other instruments? Are the plots presented in your video generated directly by C#? Have you tested times of software consumption by minimizing your app while the search process is in progress? Do you need to show all the Text box fields? Are you using the processor or a video card with compute capability? Do you use parallel algorithm? (just in case the answer is yes: did you tried using CUDA kernel to create and synchronize iterative processes directly on the GPUs? Talking about the other projects that you mentioned (6:10) in your video: I would like to know a little bit more about the way you are generating those reports. What do you think could be the essential idea to successfully complete your project and be able to generate those reports (as part of your experience in INTI)?

Hi Mauricio, thanks for the talk. I was wondering, usually we produce calibration history for our own instruments, and in the case of external ones (users) we do so in a non systematic way. This is usually work for the owner of the instrument. But from what you are doing, the plan is to get comparisons between instruments of users and get batch information from them? Thanks, Mariano Real

Dear Iram, thank you very much for your questions, I would like to be able to exchange experiences, I will answer your questions Did you have any problem with encrypted files? If I had a lot of problems, what gave me the most results is converting to an image and OCR, but until now I have only managed to recover with 95% efficiency. Is the search process focused on text, structure or images? It is divided, first it searches for key text, and based on the found key text it looks for structures. The database is great! I am currently working with something similar to that concept. Are you considering sharing those results with other laboratories at INTI using a filter to show them yours or by identification? (This could be useful to add that data to your software to avoid entering values ​​by hand or to check if the certificates are up to date.) Temperature could be something that your coworkers could easily enter into your code (I'm sure you could thank you if you were developing software in LATU and you could share those results to allow me to use them directly in my software) Yes, of course, the search engine you build allows you to search by equipment ID or Serial Number. I would search all the history for that Id. Are there any limitations or, generally speaking, there is no other classification option for other instruments? If you can put together any rating for any type of instrument, that's part of the training that needs to be done to make it work. Are the frames presented in your video directly generated by C #? If everything is generated in C # and the reports with the graphics too. Have you tested software consumption times minimizing your application while the search process is in progress? The consumption depends on where the Word or Excel files are, for data extraction, if they are hosted in a cloud, or local server, it is when it consumes the most resources. Do you need to display all the fields in the text box? It is not necessary, I just did it to show everything that can be extracted. Are you using the processor or a graphics card with computing power? Do you use parallel algorithm? (just in case the answer is yes: did you try to use the CUDA kernel to create and sync iterative processes directly on the GPUs? I answer both questions, I use PETSc and OpenMP, to take advantage of multiple processor CPUs and OpenCL to use GPU Speaking of the other projects you mentioned (6:10) in your video: I'd like to know a little more about how you're generating those reports. What do you think could be the essential idea to successfully complete your project and be able to generate these reports (as part of your INTI experience)? It is a project that started several years ago, which was extended due to lack of time and resources. They have several parts, since I made from the hardware (cables for the connection of the scales with the PCs, the data acquisition board for the sensors) and the software that acquires data, and passes the data directly to an Excel designed to do it automatically. The report. Because everything is done with the Laboratory Management Software that I made. The management software passes all the equipment and client data to Excel, and the other programs the calibration results. Remaining the Report Done.

Estimado Claudio, gracias por tu consulta ¿Este software puede ser aplicado a instrumentos de control de masa o de temperatura? Si se puede entrenar para cualquier tipo de instrumentos. ¿qué requisitos de confidencialidad fueron tenidos en cuenta, al momento de trabajar con diferentes marcas de equipamiento? Las consultas se divide en función de quien la solicite, sise solicita el historial de un Cliente muestra los datos del instrumento porque es del cliente. Si la consulta es de otro tipo no muestra identificaciones de equipo, ni de clientes.

Hi Carlos, thanks for the question, the idea will be to give more information to the client on how the instrument comes, here many clients do not analyze the results of their calibrations. Many studies can be done from the stability of the instruments, but also see the Laboratory's own capacity.

Hola Claudio, gracias por tu pregunta, Las consultas, se arman en funcion de que es lo que se pide. Si la consulta es sobre el Equipo de un Cliente, para ser enviado al cliente, se muestran todos los valores. Si la consulta es general, no se muestran valores de identificación del equipo, ni de Clientes. Por otro lado se puede entrenar para cualquier tipo de instrumento.