US 12,591,231 B2Granted

Method for detecting operational failures

AssigneePetroleo Brasileiro S.A. - Petrobras

InventorsBruno Martins Santos, Leonardo Fialho De Mello, Leandro Borges Grimaldi Molleri, Rodrigo Rosa Fernandes De Carvalho, Diego Telles Fernandes, Adriana De Souza Ferreira, Luiz Eduardo Borges Dos Santos, Vivian Passos De Souza, Ludmila Ribeiro De Francisco

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

The present invention is related to the field of detection of operational failures, more specifically with the monitoring of head loss in equipment such as reactors, heat exchangers and adsorption vessels. The invention uses a method for continuous monitoring of equipment head loss, in order to assist in the early identification and action on this type of problem, avoiding or reducing losses resulting from its occurrence.

Claims14

Dependent claims

A method for detecting operational failures, wherein the method comprises:

calculating current from a plant and limit from an equipment design normalized equipment head loss values from equation (1):
Δ
⁢
Pnormalized
=
(
Δ
⁢
P
)
×
(
W
ref
⁢
head
W
head
)
1.89
×
(
R
gas
/
ref
⁢
head
R
gas
/
head
)
0.88
wherein ΔPnormalized is normalized head loss, ΔP is measured head loss, Wref head is reference head flow rate, Whead is current head flow rate, Rgas/ref head is reference gas constant, and Rgas/head is current gas constant;
calculating an evolution rate of the current normalized ΔP, from an angular coefficient of a line of historical values;
calculating a moving average of current and limit normalized ΔPs, using historical values;
designing a time for the equipment to reach the normalized ΔP limit value, using the moving average values calculated and the evolution rate calculated, based on equations (2) and (3):
Time
⁢
Δ
⁢
P
⁢
limit
=
(
Δ
⁢
P
normalized
⁢
limit
⁢
(
moving
⁢
average
)
-
Δ
⁢
P
current
⁢
normalized
⁢
(
moving
⁢
average
)
)
rate
,
and
⁢
Date
⁢
Δ
⁢
P
⁢
limit
=
(
Time
⁢
Δ
⁢
P
⁢
limit
)
+
(
current
⁢
date
)
wherein
ΔPnormalized limit (moving average) is limit normalized head loss moving average,
ΔPcurrent normalized (moving average) is current normalized head loss moving average, rate is evolution rate, Time ΔP limit is predicted time to reach limit, and
Date ΔP limit is predicted date of reaching limit;
calculating a remaining campaign time from Equation (4):
Remaining campaign time=(expected date of the end of the campaign)−(current date);
calculating R from Equation (5):
R
=
Time
⁢
Δ
⁢
P
⁢
limit
Remaining
⁢
campaign
⁢
time
wherein R is risk ratio;
issuing an alert level based on R;
and
based on the issued alert level, taking action on or scheduling shutdown of the equipment.

122 more elements

Dependencies are inferred from the claim text.

Facts

Application: 18/082,001Filed 15 Dec 2022
Priority: BR 10 2021 025514 516 Dec 2021
Published: US 2023/0195096 A122 Jun 2023
Granted: 31 Mar 2026
Term ends: Not estimatedThe source records don’t support an estimate.

Classification4 codes

G05B 23/024
Process history based detection method, e.g. whereby history implies the availability of large amounts of data: quantitative history assessment,…
G05B 23/0221
Characterised by the fault detection method dealing with either existing or incipient faults: preprocessing measurements, e.g. data collection…
G05B 23/0267
Characterized by the response to fault detection: fault communication, e.g. human machine interface [HMI]
G05B
Control or regulating systems in general