Method for monitoring and analyzing a process
View Patent ↗The invention relates to a method for monitoring and analyzing a production process. In the method, a large number of variables are measured from the process, with the aid of these variables, fingerprints according to a good and a poor process situation, relative to runnability, are defined in several sub-processes and are then stored in a memory, the stored fingerprints are compared with fingerprints obtained in a normal process situation, on the basis of the comparison, the difference, displayed graphically to the user, between the recorded good situation and the momentary process situation is defined.
1. A method for monitoring and analyzing a process, the method comprising the steps of:
measuring a number of variables from the process,
with the aid of these variables, forming a fingerprint according to a good process situation with respect to runnability, and then storing the fingerprint in a memory,
comparing the stored fingerprint with fingerprints obtained in an online momentary process situation,
on the basis of the comparison, deriving the difference between the stored good process situation and the momentary process situation, and graphically displaying the difference to the user,
characterized in that the fingerprint according to a good process situation is made separately in several sub-processes, and at least one specific index in the selected sub-process is defined according to a poor process situation with respect to runnability, in order to detect a machine-specific critical situation.
2. A method, according to claim 1 , in a paper machine, characterized in that said specific index relates to one of the following:
mass-mixing in a short circulation sub-process,
a condition of felts in a press section, and
an electrochemical state of a wet-end of the process.
3. A method according to claim 1 , characterized in that the fingerprint according to a poor process situation is narrower in its area than the fingerprints according to a good process situation.
4. A method according to claim 1 , characterized in that the fingerprint according to a poor process situation is calculated from at the most six variables.
5. A method, according to claim 1 , using a neural network, characterized in that the fingerprint according to a good process situation is calculated in a teaching stage from at least ten variables.
6. A method, according to claim 1 , using a neural network, characterized in that the system is used under remote control.