Thread Content
Some time ago, during a technical exchange regarding the pressure swing adsorption process, the manufacturer of this technology raised a rather difficult issue. Below is a detailed description of it. Experts, does anyone know what’s going on here? In the control software package of the PSA unit, not only are the regular control requirements of the system met, but also an automatic arbitrary switching between multiple towers and an adaptive optimization control system, which are essential, are incorporated. To achieve automatic and arbitrary switching between multiple towers as well as adaptive optimization control functions, the control system used for PSA devices is required to possess, in addition to conventional control functions, the following capabilities: 1. The control system must have strong secondary development capabilities ; 2. Advanced programming languages are required, such as languages like BASIC, PASCAL, C, and similar advanced languages ; 3. It must provide the capability to define array variables with at least one dimension ; 4. It must have the capability to write subroutines; the main program should be able to call these subroutines and pass relevant real (actual points defined in the database) and row parameters ; 5. The PSA program utilizes a large number of variables and arrays during its development, requiring the creation of complex mathematical models; as a result, the size of the program is quite large. The main CPU must provide sufficient storage capacity for such high-level language programs, with a minimum storage capacity of 4MB generally being required ; 6. On the human-machine interface, the acquisition rate for all pressure data and timer data is ≤2 seconds. 7. On the human-machine interface, each real-time trend chart shall display no less than 16 data curves simultaneously, with a time resolution of ≤2 seconds, and the real-time trend display duration shall be no less than 40 minutes. 8. On the human-machine interface, each historical trend chart shall display no less than 16 data curves simultaneously, with a time resolution of ≤2 seconds, and the real-time trend display shall cover a period of at least 24 hours. 9. Historical data shall be retained for no less than 6 months. HONEYWELL found that it was not possible to develop the PSA program using PKS’s own software, as the PKS system did not have tools for programming in high-level languages; PKS only allowed programming using CM and SM (function blocks and sequential control blocks). Later, they used a formula management software originally designed for the food industry to carry out the programming, and they managed to complete it after all. However, the volume of code generated was unprecedentedly large. How to maintain it in the future is a major issue. As far as is known, this software is being used for the first time in the country. Its core functionality relies on the traditional function blocks and sequence blocks of the DCS system; since recipe-based software also does not provide advanced programming tools, it only offers functions for managing multiple recipes and scheduling multiple projects, thereby using different combinations of recipes to control the processes. Programs written with this software require at least 20 times more time for major modifications compared to programs written in high-level languages. Even if such a program is developed, maintenance is first and foremost a major problem. In fact, we just went to check this program some time ago, and there are still quite a few issues that require further modifications.
Personally, I think HONEYWELL’s lack of support for lower-level languages is exactly as the original poster said – configuration becomes very inconvenient. I’ve encountered the same situation: tasks that can be easily completed using higher-level languages turn out to be extremely complicated when done using CM or SM, and they require a large amount of programming code. If HONEYWELL can support the encapsulation of languages and function blocks, and there are fewer software bugs, then PKS is still a decent system in my opinion.
Why hasn’t HONEYWELL used advanced programming languages all along? In this regard, it is likely closely related to the usage habits of the relevant personnel in one’s own company, as well as compatibility with previous systems. The stability seems to be quite good
May I ask, does the PSA device belong to Tianyi or West China Institute? Follow this post. We currently have the PSA from these two manufacturers. All of them were made using PKS; two sets were made by Tianyi using TPS.
I’ve previously written an algorithm program for a PSA unit with 12 towers, but I was unable to develop a program for switching towers in case of failures – it’s extremely complex. I have here a PSA program developed by the manufacturer for 7 towers, which includes a function for switching towers in case of failures. The key functions of such a program need to be implemented using advanced programming languages; since the PKS system does not support advanced languages, I’m not sure how this can be achieved. Someone above mentioned that they have used PKS to build PSA units – how was that done? I would really like to ask for some advice. Since our factory is planning to expand and install another set of PSA units, the management aims to save costs by integrating the new PSA units into the PKS system for control.
If you’ve never used PKS, it’s impossible that you haven’t programmed in a high-level language! This is the basic function of DCS! It’s possible to write very complex programs using CL on both TDC3000 and TPS! If an AM module is used as well, even extremely complex programs can be created!
There are no issues with Honeywell PKS; Shanghai Huaxi uses PKS for PSA