Skip to content

Language on the job – From peashooters and packing fingers…

Is this a “punch” or a “punch”?

The tech­nic­al lan­guage used in oper­a­tions is com­plic­ated and often ambiguous.

Many man­u­fac­tur­ers sup­ply their pro­cessing and pack­aging machines to inter­na­tion­al cus­tom­ers. MADDOX is sup­posed to work for the same machine type in dif­fer­ent lan­guages. We are there­fore often asked wheth­er we can also auto­mat­ic­ally trans­late the know­ledge it con­tains into oth­er lan­guages. The answer is yes: An inter­face to DeepL, cur­rently the best trans­la­tion tool based on machine learn­ing, is high on our roadmap.

But what if there is another language in every company?

All plants have their own lan­guage, which is con­stantly developed by employ­ees over time. Machine oper­at­ors rarely use the tech­nic­al terms for machine parts that design­ers have writ­ten down in the manu­als. The lat­ter are often found with the tech­no­logy and are not mem­or­ized. Instead, an oper­at­or is usu­ally trained by an exper­i­enced col­league who explains and shows him the machine. New employ­ees have to invest a lot of time to learn this lan­guage. In the course of pro­fes­sion­al exper­i­ence, the exact word­ing of tech­nic­al terms can also be for­got­ten or mod­i­fied by tech­ni­cians. Terms that are too long are abbre­vi­ated. “Pack­age hold­ing fin­ger” and “trans­fer fin­ger” become “hold­ing lever” and “trans­fer lever,”. One hold­ing the pack­age, the oth­er mov­ing it along. Unknown parts are giv­en names accord­ing to their char­ac­ter­ist­ic appear­ance or their main func­tion. Some­times very dif­fer­ent parts are giv­en the same or very sim­il­ar titles, e.g. “Punch” and “Punch”. This can eas­ily lead to confusion.

Findings from a survey with our pilot customers

For a user sur­vey in the fact­ory of one of our pilot cus­tom­ers, we first worked through machine manu­als. To com­pile the users’ under­stand­ing of the machines, we then had tech­ni­cians, machine oper­at­ors and shift super­visors explain them to us in detail. It quickly became appar­ent that users have dif­fer­ent under­stand­ings of how a machine is built and how it works, and that there are many dif­fer­ent terms for a com­pon­ent. For a com­pon­ent that blows air, we heard many terms, includ­ing “blower tube,” “blow tube,” and “fan rod.” Tape des­ig­na­tion is also a good example. While tech­ni­cians coun­ted tapes by manu­al, namely from back to front, machine oper­at­ors num­ber the tapes as they usu­ally stand in front of the machine. Thus, they give the num­ber 1 to the belt closest to them and then num­ber through to the back. Of course, this can lead to misunderstandings.

MADDOX improves operational communication

In cooper­a­tion with engin­eer­ing psy­cho­lo­gists, we have developed vari­ous func­tions with which MADDOX not only enables exchanges for troubleshoot­ing, but also improves the qual­ity of communication:

  • To doc­u­ment faults and solu­tions, pho­tos can be eas­ily taken and annot­ated with text or fin­ger sketches. It is also easy to record and insert video and audio record­ings. This means that in many cases the often mis­lead­ing verbal descrip­tion can be omitted.
  • Our algorithms based on machine learn­ing treat know­ledge con­tent as a “con­tain­er” – regard­less of the con­tent. As a res­ult, con­tent that was pre­vi­ously dif­fi­cult to search for, such as pho­tos and videos, is found just as well as text. An error-prone and cum­ber­some keyword­ing is not necessary.
  • We also use the concept of work­ing as visu­ally as pos­sible for the loc­al­iz­a­tion of faults: The loc­a­tion of the fault can be graph­ic­ally marked on a stored photo or dia­gram of the plant and thus doc­u­mented unambiguously.
  • MADDOX presents stored know­ledge in a struc­ture and sequence that spe­cific­ally pro­motes under­stand­ing and avoids mis­un­der­stand­ings. For example, in the event of a mal­func­tion, a descrip­tion of the mal­func­tion is out­put first in order to cre­ate a com­mon under­stand­ing of the prob­lem and not to apply a cor­rect solu­tion to the wrong problem.

This Post Has 0 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top