A method and system for dynamic evaluation of LNG receiving station operation discipline

Abstract

The present application belongs to the field of process safety management and industrial internet technology, and particularly relates to a kind of dynamic evaluation method and system for LNG receiving station operation discipline. The method comprises: real-time acquisition of DCS operation log, alarm record and informal communication data, extraction of three dynamic indexes of interruption recovery time R(t), response diversity D(t) and informal communication network density C(t); a five-state nonlinear system dynamics model is constructed, and state variables include R(t), D(t), C(t), trust level T(t) and work load W(t); Monte Carlo simulation is used to sample the key uncertainty parameters, and the statistical distribution of system safety margin S(t) is obtained by running simulation; when S(t) is lower than the preset threshold, a risk warning signal is output. The present application provides a specific method: by real-time acquisition of the deviation value of operation sequence and standard discipline template, the adjustment coefficient of intelligent adjustment effect degree to deviation change is calculated, and then the safety margin prediction value of the system under the current working condition is output, to fill the above technical gap, and to provide effective technical support for process safety management of high-risk enterprises such as LNG receiving station under the background of industrial 4.0.

Description

Technical Field

[0001] This invention belongs to the field of process safety management and industrial internet technology, specifically involving a method and system for dynamic evaluation of operational discipline in LNG receiving stations, and in particular a dynamic evaluation technology for process safety management performance that integrates system dynamics and Monte Carlo simulation, based on three-dimensional dynamic indicators of operational discipline (interruption recovery time, response diversity, and informal communication network density). Background Technology

[0002] In the era of Industry 4.0, the deep integration of technologies such as digital twins, industrial IoT, and artificial intelligence has significantly improved the level of automation in process industries. High-risk facilities, such as LNG receiving terminals, widely adopt advanced process control systems, predictive maintenance algorithms, and intelligent alarm management systems, greatly reducing the probability of human error in routine operations. However, this increased automation has not eliminated the crucial role of humans; instead, it has pushed operators from direct "operation executors" to the core positions of "system monitors" and "anomaly handlers." This becomes the kind of "irony" described by Bainbridge: the more we try to delegate to machines, the more we rely on humans. Therefore, existing technologies cannot simultaneously quantify the dynamic deviation of operational discipline in real time, assess the moderating effect of automation levels, and predict the system's safety margin accordingly. To address this, this invention provides a specific method: by collecting the deviation values ​​between the operational sequence and a standard discipline template in real time, calculating the moderating coefficient of the degree of automation intervention on the deviation change, and then outputting the predicted safety margin value of the system under the current operating conditions, this method fills the aforementioned technological gap and provides effective technical support for process safety management in high-risk enterprises such as LNG receiving terminals under the Industry 4.0 context. Summary of the Invention

[0003] This invention aims to provide a method and system for dynamic evaluation of operational discipline at LNG receiving terminals, addressing the problems in existing technologies such as the inability to dynamically quantify operational discipline, unclear effects of intelligent regulation, and lack of real-time prediction of safety margins. The invention discloses a method and system for dynamic evaluation of operational discipline at LNG receiving terminals, aiming to solve the problems in existing technologies such as the inability to dynamically quantify operational discipline, unclear effects of intelligent regulation, and lack of real-time prediction of system safety margins. The key technical solutions include: First, real-time acquisition of DCS operation logs, alarm records, and informal communication data based on a 5G audio-visual interaction system from operators in the LNG receiving terminal control room, extracting three dynamic indicators: interruption recovery time R(t), response diversity D(t), and informal communication network density C(t). Second, a five-state nonlinear system dynamic model is constructed, using R(t), D(t), C(t), trust level T(t), and workload W(t) as state variables. A set of differential equations characterizes the dynamic evolution of recovery time, response diversity, communication density, trust level, and workload with automation level A, anomaly intensity E, and time pressure P, and introduces a trust-performance closed-loop feedback and a workload-communication inhibition loop. Third, Monte Carlo simulation is used to probabilistically sample key uncertainty parameters, and simulation is run to obtain the statistical distribution of the system safety margin S(t), where the safety margin is defined as... Finally, a warning threshold (e.g., 0.53) and an alarm threshold (e.g., 0.50) are set, and a risk warning signal is output when the safety margin falls below the corresponding threshold. The evaluation system implementing the above method includes a data acquisition module, an index calculation module, a model simulation module, a warning output module, and a visualization module. Compared with existing technologies, this invention achieves three-dimensional dynamic quantification of operational discipline for the first time, filling the gap in the lack of dynamic quantification methods for the fourth-level indicators of the API RP 754 standard; through system dynamics and Monte Carlo coupling, the impact of parameter uncertainty on the safety margin is quantified, providing a probability distribution rather than a single deterministic value; through simulations of four typical scenarios, the moderating effect of automation level on operational discipline is quantitatively revealed, finding that the optimal automation level range is 0.80–0.85; it can be embedded in the execution layer (non-DCS system) of LNG receiving stations to achieve real-time online monitoring and risk warning, providing effective technical support for process safety management of high-risk enterprises under the Industry 4.0 background. Attached Figure Description

[0004] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0005] Figure 1 This is an overall flowchart of the method of the present invention, which shows the complete steps from data acquisition, index calculation, model simulation to safety margin early warning;

[0006] Figure 2 The present invention is a modular architecture diagram of the system, including a data acquisition module, an indicator calculation module, a model simulation module, an early warning output module, and a visualization display module, as well as their connection relationships with external data sources and output terminals;

[0007] Figure 3 The five-state stock flow diagram of this invention depicts the causal relationship and feedback loop between interruption recovery time R(t), response diversity D(t), informal communication network density C(t), trust level T(t), and workload W(t);

[0008] Figure 4 This is an example diagram of the real-time monitoring curve and early warning of the safety margin in this invention, which shows the trajectory of the safety margin over time and the triggering of the early warning threshold.

[0009] Figure 5 This is a schematic diagram illustrating the effect of time pressure on operator response diversity in this invention.

[0010] Figure 6 This is a schematic diagram of the self-reinforcing feedback loop between operator trust level and system safety margin in this invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0012] This embodiment uses an LNG receiving station as an application object to provide a detailed description of the dynamic evaluation method and system for operational discipline of LNG receiving stations provided by the present invention.

[0013] First, step S1 is executed. With enterprise authorization, the operation logs for the past 12 months are extracted from the DCS system of an LNG receiving terminal. The sampling period is 1 second, and the recorded fields include operator ID, timestamp, operation type (parameter adjustment / alarm confirmation / mode switching / emergency shutdown), operation object (equipment tag number such as valve, pump, and compressor), and values ​​before and after the operation. Simultaneously, alarm records (alarm type, timestamp, confirmation time, and handling operation) are obtained from the intelligent operation management platform in the central control room. Furthermore, using the 5G audio and video interaction system deployed at the station, informal communications in the central control room are recorded for four weeks with the informed consent of the operators.

[0014] Then, proceed to step S2 to calculate three dynamic indicators:

[0015] Interruption recovery time R(t): For each interruption event (alarm sounding, walkie-talkie call, external inquiry), a stopwatch is used to record the time difference from the occurrence of the event to the operator completing the first effective operation (confirming the alarm, adjusting parameters, replying to the command). The average of 10 consecutive interruption events is taken as R(t) for this time period. The steady-state value under normal operating conditions is approximately 7.8 seconds.

[0016] Response diversity D(t): For abnormal operating conditions (such as pressure fluctuations during unloading operations), the Shannon diversity index is calculated by statistically analyzing the different response strategies adopted by operators (such as adjusting valve opening, switching to standby pumps, reducing external output, etc.). Under normal operating conditions, D(t) is approximately 0.52.

[0017] Informal communication network density C(t): Based on 5G audio and video recordings, the number of informal communication connections L between operators in the control room is counted (more than one informal communication between each pair of personnel is counted as one connection), with a team size of N=4, according to the formula... Calculate the density. Under normal operating conditions, C(t) is approximately 0.42.

[0018] Perform step S3 to construct a five-state nonlinear system dynamics model. The state variables are... The trust level T(t) and workload W(t) are the mediating variables. The system of differential equations is as follows:

[0019] , ,

[0020] ,

[0021] ,

[0022] , ,

[0023] , .

[0024] The parameter settings are as follows:

[0025] Fixed parameters (based on literature and physical constraints): , , , , , , , .

[0026] Adjustable deterministic parameters (calibrated using measured data from a specific LNG operating condition): 25, , , , , , and initial state , , , , After calibration, the steady-state safety margin under normal operating conditions is 0.57±0.01, consistent with field observations.

[0027] The distribution of uncertainty parameters (used in Monte Carlo simulations) includes: diversity adaptation rate. It follows a triangular distribution (0.20, 0.30, 0.40); communication activation rate It follows a triangular distribution (0.25, 0.30, 0.40); baseline communication density It follows a uniform distribution [0.30, 0.40]; the anomaly-manual coupling coefficient It follows a triangular distribution (0.35, 0.40, 0.50); ​​anomaly weakening coefficient. It follows a triangular distribution (0.30, 0.35, 0.45); confidence adjustment rate. It follows a truncated normal distribution (mean 0.15, standard deviation 0.03, range [0.10, 0.25]); workload adaptation rate It follows a triangular distribution (0.15, 0.20, 0.30). Monte Carlo number. , with a fixed random seed rng(42).

[0028] Design four typical scenarios:

[0029] Scenario 1 (Normal Operating Conditions): , , The simulation duration is constant, 460 minutes (one shift).

[0030] Scenario 2 (Automation Failure): Within t∈[20,40] minutes, the automation level A drops to 0.25, and the anomaly intensity E rises to 0.50; simultaneously, an anomaly pulse t∈[50,55] is added. The simulation duration is 460 minutes.

[0031] Scenario 3 (High Time Pressure); , The time pressure P increases to 0.80 within t∈[25,35] minutes. The simulation duration is 460 minutes.

[0032] Scenario 4 (Long-term dependency): Automation level A increases linearly from 0.70 to 1.00 (1440 minutes). , .

[0033] Run Monte Carlo simulations for each scenario and extract key metrics:

[0034] Scenario 1: Average safety margin for the last 20% of the time. Scenarios 2 and 3 take the global minimum safety margin. Scenario 4: Finding the endpoint safety margin The formula for calculating the safety margin is:

[0035]

[0036] The simulation results are as follows: For Scenario 1 (normal operating conditions), the mean safety margin is 0.5518, the standard deviation is 0.0089, the 5th percentile is 0.5364, the 50th percentile is 0.5532, and the 95th percentile is 0.5657. For Scenario 2 (automation failure), the mean safety margin is 0.4288, the standard deviation is 0.0083, the 5th percentile is 0.4144, the 50th percentile is 0.4290, and the 95th percentile is 0.4419. For Scenario 3 (high time pressure), the mean safety margin is 0.4564, the standard deviation is 0.0087, the 5th percentile is 0.4416, the 50th percentile is 0.4570, and the 95th percentile is 0.4702. The mean safety margin for scenario 4 (long-term dependence) is 0.4906, with a standard deviation of 0 (the same for all samples). The 5th, 50th, and 95th percentiles are all 0.4906.

[0037] Execute step S4. Based on the simulation results, set the warning threshold. Alarm threshold In real-time monitoring, when the calculated safety margin falls below 0.53, the system issues a yellow warning signal on the central control room screen, prompting the operator to strengthen monitoring; when it falls below 0.50, a red alarm signal is issued, suggesting immediate intervention measures (such as reducing production load, increasing personnel support, or switching to manual control mode). Taking scenario 2 as an example, the average safety margin is 0.4288, far below the alarm threshold, and the system continuously outputs a red alarm. Corresponding post-event analysis shows that after automation failure, the average operator recovery time increased to 22 seconds (3 times the normal value), response diversity decreased to 0.62, and trust level decreased from 0.72 to 0.52, verifying the effectiveness of the early warning system of this invention.

[0038] The evaluation system provided by this invention includes a data acquisition module, an index calculation module, a model simulation module, an early warning output module, and a visualization module. In an actual deployment at an LNG receiving terminal, the data acquisition module acquires raw data from the DCS system and the 5G audio-visual system at a 1-second interval; the index calculation module calculates R(t), D(t), and C(t) in real time; the model simulation module calls the differential equation solver every 5 minutes to update the safety margin prediction value; the early warning output module compares the current S(t) with the threshold and pushes early warning information through the central control room screen and operator mobile terminals; and the visualization module generates a real-time safety margin curve and a historical comparison chart.

[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method and system for dynamic evaluation of operational discipline at LNG receiving terminals for Industry 4.0, characterized in that, Includes the following steps: S1: Real-time collection of DCS operation logs, alarm records, and informal communication data based on the 5G audio and video interaction system from operators in the LNG receiving terminal control room, extracting the following dynamic indicators of operational discipline: Interruption recovery time R(t): The time difference between the occurrence of the interference event and the completion of the first effective operation by the operator, in seconds; Response diversity D(t): The diversity of operator response strategies under abnormal operating conditions is calculated using the Shannon diversity index. in Let i be the frequency of occurrence of the i-th response strategy within the time period; Informal communication network density C(t): The density of informal communication connections within a work group is calculated using social network analysis methods. Where L represents the number of informal communication connections that actually occurred during the time period, and N represents the number of people in the work group; S2: Construct a five-state nonlinear system dynamics model, where the state variables are... ,in The level of operator trust in the automation system, The cognitive workload of the operator is represented by the model, which is described by the following set of differential equations: ， 。 ， ， ， ， ， 。 in, To the level of automation, It is of abnormal strength. Due to time pressure, It is Gaussian white noise. For system safety margin, These are model parameters; S3: Using the Monte Carlo simulation method, key uncertainty parameters in the model are probabilistically sampled, and the simulation is run to obtain the system safety margin. The statistical distribution of the safety margin, defined as: Among them, the weighting coefficient Determined using the analytic hierarchy process (AHP). This is the maximum allowable recovery time. S4: Based on safety margin Based on the calculation results, set the early warning threshold and alarm threshold, when A yellow risk warning signal is output when the level is below the warning threshold. A red alarm signal is output when the value is below the alarm threshold.

2. The method according to claim 1, wherein a warning signal is issued when the safety margin is lower than the warning threshold; and an alarm signal is issued when the safety margin is lower than the alarm threshold; wherein the alarm threshold is lower than the warning threshold.

3. The method according to claim 1, characterized in that: The Monte Carlo simulation used a sampling number N=500, with a fixed random number seed to ensure reproducible results.

4. The method according to claim 1, characterized in that: The fixed parameter in the model parameters takes the value of , , , , , , , Second.

5. The method according to claim 1, characterized in that: The key uncertainty parameter includes the diversity adaptation rate. Communication activation rate Benchmark communication density Abnormal - Manual Coupling Coefficient Abnormal weakening coefficient Trust adjustment rate Workload adaptation rate and initial state Their probability distributions are triangular, uniform, or truncated normal, respectively.

6. A dynamic evaluation system for operational discipline at an LNG receiving terminal, used to implement the method according to any one of claims 1-5, characterized in that, include: The data acquisition module is characterized by comprising: a 5G audio and video data parsing unit, used to receive voice streams from the 5G audio and video interaction system of the LNG receiving station in real time, and to filter out informal communication segments of operators through keyword matching (keywords include words such as "don't know", "maybe", "never mind" that indicate operational uncertainty or tendency to violate regulations); and an event association unit synchronized with the DCS system and intelligent operation management platform, used to bind the filtered informal communication segments with the DCS operation logs and alarm records within the same time window to form a "discipline deviation event triplet". The indicator calculation module, connected to the data acquisition module, is configured to calculate the interruption recovery time R(t), response diversity D(t), and informal communication network density C(t) based on the acquired data. The model simulation module is connected to the index calculation module, and has the built-in five-state nonlinear system dynamics model configured to perform Monte Carlo simulation and output the dynamic evolution results of the safety margin S(t). The early warning output module is connected to the model simulation module and is configured to compare the safety margin S(t) with the preset early warning threshold and alarm threshold. When S(t) is lower than the threshold, the corresponding risk early warning signal is output. The visualization module is connected to the model simulation module and is configured to generate time series plots, box plots, and cumulative distribution function plots of the safety margin.

7. The system according to claim 6, characterized in that: The data acquisition module collects operator ID, operation timestamp, operation type, operation object, and values ​​before and after the operation from the DCS operation log during the sampling period.

8. The system according to claim 6, characterized in that: The interrupt recovery time R(t) in the indicator calculation module is taken as the average value of 10 consecutive interrupt events.

9. The system according to claim 6, characterized in that: The early warning output module is also configured to record the time, duration, and safety margin value of each early warning event, forming an operational discipline risk log.

10. The system according to claim 6, wherein the visualization module is further configured to: display the current safety margin value in real time and map the value to the angle between the pointer deflection angle on the dashboard and the background color ring; when the safety margin is lower than a first threshold, the pointer automatically swings to the red area and the color ring angle is reduced to 50% of the original angle; when the safety margin is lower than a second threshold, the dashboard displays a predicted countdown number, which is dynamically calculated by the ratio of the adjustment coefficient to the deviation.

Images