Exaquantum Traveling Sales Diary

It has been more than six months since my last essay. This is not because I was sick or had run out of topics; rather, I was busy working on some new ideas that I briefly mentioned in diary entry No. 9, namely (1) figuring out how to quantify the intuition and expertise of experienced plant operators, and (2) monitoring the relationships between multiple loops. I was running around to various customer sites to prove my theory by experimenting with their data. Let me take some time to explain these theories in detail.Consider the example of a Chinese restaurant. It serves a certain fried rice dish that is prepared by an excellent cook. Even if several people order the same fried rice dish at the same time, the servings all taste exactly the same. But suppose one day I taste a difference. I dare to ask the cook whether he used any different ingredients, and the cook replies that the ingredients were the same, there was no change in the seasonings, and everything was cooked in exactly the same way as before. Even if I claim to be a gourmet, I cannot figure out what made it so. Had the cook changed the kind of salt? Was he using too much seasoning? Was I just tired and craving salty food, or had I caught a cold and had a stuffy nose? Could it have been any one or all of these reasons?

Haven't you ever heard similar stories at your plant? Just because a cook is very good does not mean that he will always cook a dish exactly the same way, right down to using the same number of grains of salt or some other seasoning. How it is heated or how long it is fried can vary slightly. A skilled cook takes all those tiny differences into consideration and brings everything to a perfect balance. The variety and amount of the ingredients and seasonings, the heating temperature and time, the speed with which it is stirred, and other variables all determine the final taste of the fried rice. If this balance is not achieved, or if the person's "sensors" are malfunctioning, it is certain that the fried rice will taste differently. In industrial plants, almost all the controls such as volume, temperature, time, and mixing speed are automated by a DCS or PLC, and every control object is adjusted within a pre-defined threshold. Even so, you may come across situations where quality varies and you cannot pinpoint the problem. You know that it has nothing to do with the DCS or PLC and believe that a deficiency in operator skills or a failure of a measurement sensor may be at fault. The DCS can diagnose a sensor failure, but it cannot tell you what caused the failure or identify whether some difference in the expertise and skills of the individual operators is at fault. That's not the role of the DCS. Aside from that, it is only human beings who seek to improve quality or increase efficiency. For instance, you may be aware that all machines may break down sooner or later and face a choice between checking them periodically or simply repairing them when they fail. Wouldn't it be better if you could rest assured that the machine would notify you of an impending malfunction several hours in advance and, providing such an alarm has not been issued, rely on it to function well and operate at peak efficiency?

In my latest experiments, Exaquantum continued to accumulate relevant operational data while equipment was operating, and marked those points where it operated optimally. The program informed the operators whether the current process data showed the same optimal balance and, if not, identified the cause for the deviation. By marking when equipment was operating perfectly and comparing the current data against the perfect data, Exaquantum was able to issue a predictive alarm when there was any deviation in the equipment performance. How about if you buy the Exaquantum package so I can tell you how to do this? …. Ah, well, I'm generous when it comes to sharing secrets. And it is simple! All you need is a good amount of historical data.

This technique was developed recently, and it applies the multivariable analysis method. The historical operational data includes both process and equipment data. Exaquantum collects data from when the process is operating smoothly and the equipment is in good condition, and generates a co-relational matrix with this data. This provides a reference point that can be used to identify when a process and its equipment are well balanced. If it only involved 5 or 10 tags, any smart operator could create such a matrix. But Exaquantum is capable of handling an even larger number of tags: 100, 200, or even more. This same procedure can then be repeated with the current process and equipment status data to generate a second co-relational matrix. By comparing the two matrixes, Exaquantum can determine whether the current process and its equipment are well-balanced. If so, the plant is in a healthy state. If not, the Exaquantum identifies which tag caused the imbalance.

I assume that you now understand the theory. You may wish to know if it can really be put to practical use! Well, so far I have received historical data from 10 companies, analyzed them with Exaquantum, and identified the causes of failures or spotted problems quite successfully. I have predicted failures in silicon quality, polymerization, distillation columns, quality of films and sheet materials, and so on. I also proposed my methodology to other customers who have used this to improve their yield rate in rubber manufacturing, predict failures in continuous smelters, provide finishing criteria for polymerization, and manage the lifecycle of plant equipment. Would you like me to visit your plant and analyze process failures?Or, would you like to predict equipment malfunctions? Please do not hesitate to let me know. I am willing to travel anywhere!

Chinese dumplings (Gyoza) in Utsunomiya

The other day, I had some business in the Utsunomiya area of Tochigi Prefecture. I took along with me a recent university graduate who had joined Yokogawa in April and is receiving on-the-job training in my group. Of course I did not expect him to help me make sales presentations to my customers, but I wanted him to have a positive experience on his first business trip.

Utsunomiya is well-known for its Chinese dumplings (called "gyoza" in Japanese), so it was not at all difficult to go online and find a number of restaurants specializing in this delicacy. I tried to create a short list based on customer reviews, but eventually gave up on that because the evaluations varied so much. Even though I wanted to sample a wide variety of dumplings, this was not practical, so I just made a mental note of the names of some of the better known restaurants. But then I came across the Ki-Rasse food court in the basement of a major supermarket that was just a 10 minute walk from Utsunomiya Station. Several dumpling restaurants had booths there, and so I thought, "That's convenient! I can save time and wander around to the different booths, all in one place. " Well, yes, I admit that I immediately formed the plan to visit every one of these booths. By the time we finished our meetings with the customers and returned in a taxi to Utsunomiya Station, it was already five o'clock. I thought nobody could blame us for going out then for some beer and dumplings, instead of heading straight back to the office! So off went to Ki-Rasse, where I expected to find small booths all clustered together. But it was somewhat different: the floor was divided into two main areas, and the first of these areas had just one restaurant that was shared by 27 of Utsunomiya's dumpling restaurants.

Each day, on a rotating schedule, seven of these restaurants served their food. I was disappointed to find none of the restaurants from my Internet search represented there. Feeling a bit discouraged by my misfortune, we went on to the other area where I found booths from five of the famous restaurants. It was set up so that you can order and pay for your meal at any of these booths, and have the food delivered to your table.

After a brief discussion with my colleague, we agreed to place orders for pan-fried and boiled dumplings from three of the famous restaurants."Mi" was the most famous of these, and they also have a kiosk in the train station."Me" is known for its pan-fried hane-tsuki (wing-shaped) dumplings. And "Ko" serves dumplings with garlic, a rarity for the Utsunomiya area. And how could we eat them without beer? We started with the pan-fried and boiled dumplings from "Ko." The strong garlic in the dumplings brought smiles to our faces and exclamations of delight. We then gulped down the dumplings from "Mi" as soon as they were served. Our impression was that, without the garlic, they were rather delicate in taste, lacking a "punch."This kind of dumpling is like those served in China, and most of the restaurants in Utsunomiya cook them in the traditional Chinese way. Those used to eating dumplings with garlic or garlic chives may find the Utsunomiya variety less appealing. In authentic Chinese restaurants, boiled dumplings, not pan-fried ones, are usually served. In Chinese households, dumplings are usually only fried when served later, as leftovers from a previous meal. While I was busy telling my colleague all this dumpling trivia, our hane-tsuki dumplings from "Me" arrived. These pan-friend dumplings were clumped together in a wing-like shape, and had a crispy outer surface and a very satisfying filling.

After eating six different kinds of dumplings, we cared less whether they contained garlic or garlic chives; all were delicious and the beer was as tasty as ever.But the beer was also filling and so we went looking for chu-hai, a mixed drink made from shochu (a liquor distilled from barley, sweet potatoes, or rice) and fruit juice.We looked in at "Ry" and found that they had it. Being polite, I placed an order for dumplings to go together with the chu-hai. My colleague found this absurd. Well, no matter what he thought, the dumplings disappeared into our stomachs in no time. In the end, we ranked the dumplings in the order of "Ko," "Me," "Mi," and "Ry." This is just our evaluation, and others may have a different take on this. But when you are in the Utsunomiya area and you do not know which restaurant to go to, Ki-Rasse will suit you best. To sample the greatest variety of dumplings, I also recommend that you go along with several others.

By the way, my colleague had originally applied to work in the Corporate R&D Headquarters, but he has come around and is beginning to think that it might be interesting to get a position where he can travel around the country.