DCS system maintenance system based on networking intelligent monitoring and method thereof

Abstract

The invention provides a DCS system maintenance system based on networking intelligent monitoring and a method thereof, and relates to the technical field of intelligent manufacturing and high-voltage circuit breaker health management. Firstly, an independent classification identifier and a unique increasing number are distributed for each type of high-voltage circuit breaker, and multi-dimensional operation data of temperature, pressure, vibration and current are collected at fixed time intervals and are stored in a high-voltage circuit breaker state data set according to the number and timestamp of the high-voltage circuit breaker. And then constructing a dynamic weight distribution formula, outputting a health state score and recording the health state score to a health state database. And extracting operation data in a fault state to construct an abnormal data set, combining the health state score with the abnormal data set, and performing deep training and prediction by using an LSTM model so as to obtain a future fault occurrence probability of the high-voltage circuit breaker. Once the predicted fault occurrence probability of the high-voltage circuit breaker is higher than a fault threshold value, the high-voltage circuit breaker is immediately calibrated as a potential fault high-voltage circuit breaker, a detailed and targeted maintenance plan is generated, and meanwhile, the operation parameters of the potential fault high-voltage circuit breaker are rapidly adjusted.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent manufacturing and high-voltage circuit breaker health management, and in particular to a DCS system maintenance system and method based on networked intelligent monitoring. Background Art

[0002] In the automated control system for vaporized liquid argon, a critical component of modern industrial systems, high-voltage circuit breakers, as core equipment for ensuring a stable power supply, shoulder the crucial mission of controlling and protecting the circuits. Their reliable operation is crucial for maintaining the smooth operation of the system, ensuring the continuity of the liquid argon vaporization process, and protecting the safety of various electrical equipment within the system. From the cryogenic pump drive motor at the liquid argon storage end, to powering the heating device during the vaporization process, to the various sensors and actuators in the control system, high-voltage circuit breakers operate throughout the entire system, and their performance directly impacts the stability and safety of the system.

[0003] In automatic control systems for vaporizing liquid argon, high-voltage circuit breakers often face extremely low temperatures, high pressures, and strong electromagnetic interference. For example, the distillation tower's motor and cooling system may require a high-voltage circuit breaker to protect the circuit from overload or short circuit. The vaporization of liquid argon is accompanied by a large amount of heat exchange, which can cause drastic changes in the local ambient temperature and potentially degrade the insulation material of the high-voltage circuit breaker. The high-pressure environment within the system places higher demands on the circuit breaker's sealing performance and mechanical structure. Furthermore, the strong electromagnetic interference generated by the system's electrical equipment during operation can affect the control signal transmission of the high-voltage circuit breaker, thereby affecting its normal operation. These complex and harsh operating conditions greatly increase the risk of high-voltage circuit breaker failure.

[0004] Traditional maintenance strategies for high-voltage circuit breakers rely primarily on regular manual inspections and post-event repairs. In vaporized liquid argon automatic control systems, this manual inspection model faces numerous challenges. Firstly, the high-voltage circuit breakers in this system are often located in complex locations, with some areas potentially subject to safety risks such as frostbite and suffocation, making it difficult for inspectors to conduct comprehensive and in-depth inspections within a limited timeframe. Secondly, the accuracy of manual judgment of equipment status is significantly affected by factors such as the inspector's professional qualifications and work experience. Different inspectors have varying perceptions and criteria for identifying signs of high-voltage circuit breaker failure, resulting in a lack of unified standards and reliable evidence for assessing the health of high-voltage circuit breakers.

[0005] Under today's trend toward digital and intelligent industrial development, the drawbacks of existing high-voltage circuit breaker maintenance technologies are becoming increasingly prominent. First, there are limitations to data acquisition. Many existing monitoring systems can only monitor data in a single dimension, such as focusing solely on current or voltage. This prevents them from fully acquiring the operating information of high-voltage circuit breakers and making it difficult to accurately assess their health status. Secondly, in terms of fault prediction, most existing models rely on a single data source or simple algorithmic models, failing to fully consider the complex factors involved in high-voltage circuit breaker operation, resulting in poorly accurate prediction results. Furthermore, maintenance planning lags behind. Although some intelligent monitoring systems can monitor the status of high-voltage circuit breakers in real time, the models lack a dynamic adjustment mechanism, making it impossible to formulate reasonable maintenance plans based on changes in the equipment's operating status in a timely manner. This makes it difficult to effectively respond to dynamic changes in the operating status of high-voltage circuit breakers. Summary of the Invention

[0006] The object of the present invention is to provide a DCS system maintenance system and method based on networked intelligent monitoring to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention further provides a DCS system maintenance system based on networked intelligent monitoring, comprising:

[0008] A status acquisition module is used to number the high-voltage circuit breakers in the DCS system in an incremental manner, collect multi-dimensional operating data at fixed time intervals, and store them as high-voltage circuit breaker status data sets by number and time stamp. The operating data includes temperature, pressure, vibration, and current.

[0009] The status scoring module is used to construct a dynamic weight distribution formula based on the high-voltage circuit breaker status data set and the real-time changes in the high-voltage circuit breaker operating status, output the health status score in real time, and record it in the health status database;

[0010] The fault prediction module is used to analyze the high-voltage circuit breaker status data set, extract the operating data of the fault state to construct an abnormal data set, and build an LSTM model based on the real-time changes in the health status score and the abnormal data set to predict the future failure probability of the high-voltage circuit breaker;

[0011] The fault judgment module is used to pre-set a fault threshold, mark the high-voltage circuit breaker with a predicted fault probability higher than the fault threshold as a potential fault high-voltage circuit breaker, generate a maintenance plan and adjust the operating parameters of the potential fault high-voltage circuit breaker.

[0012] Furthermore, numbering the high-voltage circuit breakers in an incremental manner includes the following steps:

[0013] Assign independent classification identification to high-voltage circuit breakers, and then assign unique incremental numbers in turn, that is, assign a unique high-voltage circuit breaker number W to each high-voltage circuit breaker.ij , where i represents the classification number of the high-voltage circuit breaker and j represents the increasing serial number under the classification.

[0014] Furthermore, storing the data as a high-voltage circuit breaker status dataset by number and time stamp includes the following steps:

[0015] At fixed time intervals t gd For the data collection cycle, collect multi-dimensional operating data and record the temperature of each high-voltage circuit breaker as T ij , pressure is P ij , vibration is V ij , the current is I ij , then number the high voltage circuit breaker W ij and the corresponding timestamp t, the operating data of each high-voltage circuit breaker is stored as a high-voltage circuit breaker status dataset in real time, where the time interval t gd Given by experience, the default is 5 minutes.

[0016] Furthermore, constructing a dynamic weight allocation formula includes the following steps:

[0017] Extract the operating data of the high-voltage circuit breaker status dataset and construct a dynamic weight distribution formula:

[0018] S ij (t) = ω1·f1(T ij (t))+ω2·f2(P ij (t))+ω3·f3(V ij (t))+ω4·f4(I ij (t))

[0019] Where S ij (t a ) represents the high voltage circuit breaker W ij Health status score at time t; f1, f2, f3 and f4 are the calculation functions of temperature, pressure, vibration and current respectively, ω1, ω2, ω3 and ω4 are the weight coefficients of the health status score of the high-voltage circuit breaker, and ω1+ω2+ω3+ω4=1, ω2>ω1>ω3>ω4>0;

[0020] The formula for the temperature calculation function is:

[0021]

[0022] Where, T ij，ref It is represented as the reference temperature of the high-voltage circuit breaker, α1 is a sensitivity constant ≥ 0, which represents the sensitivity of temperature to the health status of the high-voltage circuit breaker;

[0023] The formula for the pressure calculation function is:

[0024]

[0025] Where, P ij，ref It is expressed as the reference pressure of the high-voltage circuit breaker, α2 is a sensitivity constant ≥ 0, which represents the sensitivity of pressure to the health status of the high-voltage circuit breaker;

[0026] The formula for the vibration calculation function is:

[0027]

[0028] Where V ij，ref It is expressed as the reference vibration value of the high-voltage circuit breaker, γ1 and γ2 are sensitivity constants, where γ1<0, γ2>0, which represents the impact of vibration on the health status of the high-voltage circuit breaker;

[0029] The formula for the current calculation function is:

[0030]

[0031] Where, I ij，ref It is represented as the reference current of the high-voltage circuit breaker, and α3 is a sensitivity constant ≥ 0, which represents the impact of the current on the health status of the high-voltage circuit breaker.

[0032] Furthermore, the algorithm method of the sensitivity constant is:

[0033] In the stage where the health status is known to be good, collect n sets of operating data of the high-voltage circuit breaker, and record the collected temperature data of the high-voltage circuit breaker as T ijN , pressure data is P ijN , vibration data is V ijN , current data is I ijN , the health status score at this stage is S max is 1, where N = 1, 2, ..., n, n is a positive integer ≥ 2,

[0034] Compute the mean of temperature data:

[0035]

[0036] Where, It represents the average value of n sets of temperature data of the high-voltage circuit breaker collected during the stage when the high-voltage circuit breaker is known to be in good health, and n represents the number of operating data sets collected;

[0037] Construct the intermediate temperature formula:

[0038]

[0039] The calculation formula of the temperature sensitivity constant can be obtained as follows:

[0040]

[0041] Calculate the average of the pressure data:

[0042]

[0043] Where, It represents the average value of n sets of pressure data of the high-voltage circuit breaker collected during the known good health stage;

[0044] Construct the intermediate formula for pressure:

[0045]

[0046] The calculation formula of the pressure sensitivity constant can be obtained as follows:

[0047]

[0048] Calculate the average value of vibration data:

[0049]

[0050] Where, It represents the average value of n sets of vibration data of the high-voltage circuit breaker collected during the period of known good health status;

[0051] when When , construct the vibration intermediate formula:

[0052]

[0053] The calculation formula of the vibration sensitivity constant can be obtained as follows:

[0054]

[0055] when When , construct the vibration intermediate formula:

[0056]

[0057] The calculation formula of the vibration sensitivity constant can be obtained as follows:

[0058]

[0059] Calculate the average value of the current data:

[0060]

[0061] Where, It represents the average value of n sets of current data of the high-voltage circuit breaker collected during the known good health stage;

[0062] Construct the intermediate formula for current:

[0063]

[0064] The calculation formula of the current sensitivity constant can be obtained as follows:

[0065]

[0066] Furthermore, the method of constructing an LSTM model to predict the future failure probability of high-voltage circuit breakers is as follows:

[0067] The high-voltage circuit breaker status dataset is used to extract the operating data of each high-voltage circuit breaker's fault state. When the temperature, pressure, vibration, and current exceed the safe range, they are marked as faulty. A fault dataset is created. This dataset includes the operating data of the high-voltage circuit breaker marked as faulty from 12 hours before to 12 hours after the fault occurs, along with the corresponding high-voltage circuit breaker number and timestamp. Based on the operating data collected during this time period, the health status score of the high-voltage circuit breaker is calculated according to the dynamic weight allocation formula.

[0068] Afterwards, the operating data of the fault dataset and the corresponding high-voltage circuit breaker number, timestamp, and calculated health status score are used as input features.

[0069] The time step is:

[0070]

[0071] Where timesteps represents the time step of the LSTM model, T represents the time duration of the fault state, and t gd It represents the time interval for collecting multi-dimensional operation data;

[0072] The fault data set is divided into multiple time series samples in chronological order. A sliding window with a window size of timesteps is set. Each time it slides, it takes one time step. The data at the previous time point is used as the input feature (training data), and the data at the next time point is used as the output label (prediction result), that is, the probability of future failure of the high-voltage circuit breaker, and the value range is [0, 1]. The number of LSTM hidden layer units is set to 64, and the activation function is the hyperbolic tangent activation function. The number of LSTM layers is 2, and the number of hidden units in each LSTM layer is set to 128. Two fully connected layers are added after the LSTM layer. The last one is The layer uses the sigmoid activation function to output the probability of fault occurrence. The loss function uses the binary cross entropy loss function. The Adam optimizer is used, the initial learning rate is set to 0.001, the batch size is set to 32-128, and the training rounds are set to 50-200 rounds. A Dropout layer is added after each LSTM layer and fully connected layer with a ratio of 0.2 and a gradient threshold of 5. The future fault probability of the high-voltage circuit breaker is used as the output label with a value range of [0,1]. An LSTM model is constructed. After training, the trained model is used to predict the next time point to obtain the predicted fault probability.

[0073] Furthermore, marking a high-voltage circuit breaker with a predicted fault probability higher than a fault threshold as a potential fault high-voltage circuit breaker includes the following steps:

[0074] Use the trained LSTM model to predict the test data set and obtain the future failure probability of each high-voltage circuit breaker. The fault threshold is pre-set as P thresh , for high voltage circuit breaker W ij , when the future failure probability of the high-voltage circuit breaker P ij >P thresh , then the high voltage circuit breaker W ij Marked as a potential fault high voltage circuit breaker.

[0075] Furthermore, generating a maintenance plan and adjusting operating parameters of a potential faulty high-voltage circuit breaker includes the following steps:

[0076] For high-voltage circuit breakers identified as potential faults, adjustment instructions are issued through the DCS to modify the operating parameters of the high-voltage circuit breakers, namely increasing the cooling efficiency of the high-voltage circuit breakers, reducing the working pressure, reducing the speed of the high-voltage circuit breakers, and reducing the load current of the high-voltage circuit breakers. At the same time, an early warning is issued to generate a maintenance plan.

[0077] The present invention further provides a DCS system maintenance method based on networked intelligent monitoring. The maintenance method is performed by the above-mentioned DCS system maintenance system based on networked intelligent monitoring, and the specific steps include:

[0078] Step 1: Number the high-voltage circuit breakers in the DCS system in an incremental manner, collect multi-dimensional operating data at fixed time intervals, and store them as high-voltage circuit breaker status datasets according to the number and time stamp. The operating data includes temperature, pressure, vibration, and current.

[0079] Step 2: Based on the high-voltage circuit breaker status dataset and the real-time changes in the high-voltage circuit breaker operating status, a dynamic weight distribution formula is constructed to output the health status score in real time and record it in the health status database;

[0080] Step 3: Analyze the high-voltage circuit breaker status dataset, extract the operating data of the fault state to construct an abnormal dataset, and build an LSTM model based on the real-time changes in the health status score and the abnormal dataset to predict the future failure probability of the high-voltage circuit breaker.

[0081] Step 4: Pre-set a fault threshold, mark the high-voltage circuit breakers with a predicted fault probability higher than the fault threshold as potential fault high-voltage circuit breakers, generate a maintenance plan and adjust the operating parameters of the potential fault high-voltage circuit breakers.

[0082] Compared with the prior art, the present invention has the following beneficial effects:

[0083] The present invention solves the problem of insufficient data collection in traditional high-voltage circuit breaker management by regularly collecting and real-time monitoring data on multiple dimensions, including temperature, pressure, vibration, and current. Using these multi-dimensional data, the operating status of the high-voltage circuit breaker can be more comprehensively evaluated. By combining the real-time status changes of the high-voltage circuit breaker, the weights of various parameters are dynamically adjusted and the health status score of the high-voltage circuit breaker is calculated, which can reflect the health status of the high-voltage circuit breaker in real time. Finally, an abnormal data set is constructed based on the fault status data, and the LSTM (long short-term memory) model is used to predict the probability of future failures of the high-voltage circuit breaker. This allows for early identification of potential faults, early warning, and adjustment of the high-voltage circuit breaker operating parameters, solving the problem of the traditional method's lack of real-time performance and reducing the occurrence and losses of high-voltage circuit breaker failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0085] Figure 2 It is a schematic diagram of the overall system module of the present invention. DETAILED DESCRIPTION

[0086] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0087] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0088] Example:

[0089] See also Figure 1 , the present invention provides a technical solution:

[0090] A DCS system maintenance system based on networked intelligent monitoring includes a status acquisition module, a status scoring module, a fault prediction module, and a fault determination module. The status acquisition module incrementally numbers high-voltage circuit breakers in the DCS system, collects multi-dimensional operating data at fixed time intervals, and stores the data as a high-voltage circuit breaker status dataset based on the number and timestamp. The operating data includes temperature, pressure, vibration, and current.

[0091] In the vaporized liquid argon automatic control system, there are various types of high-voltage circuit breakers. To achieve accurate management and data monitoring, each type of high-voltage circuit breaker needs to be clearly identified and numbered. Specifically, each type of high-voltage circuit breaker is first assigned an independent classification identifier i, and then a unique incremental number j is assigned in sequence, that is, each high-voltage circuit breaker is assigned a unique high-voltage circuit breaker number W. ij , where i represents the classification number of the high-voltage circuit breaker, such as oil circuit breaker, gas circuit breaker, vacuum circuit breaker, and j represents the incremental serial number under the classification, with a fixed time interval t gd The data collection cycle is the data collection cycle. In the power industry, temperature is usually an important factor affecting the health status of high-voltage circuit breakers. Especially when the high-voltage circuit breaker is in a high-temperature environment, the risk of damage to the high-voltage circuit breaker will increase. Pressure also has a significant impact on the operating status of the high-voltage circuit breaker, especially in power systems containing pumps, valves and other equipment. Abnormal pressure may cause high-voltage circuit breaker failure. Vibration is often a precursor to damage to high-voltage circuit breakers. Especially for high-voltage circuit breakers, excessive vibration may cause damage to components. Current is an operating parameter of electrical high-voltage circuit breakers. Its changes reflect changes in current load, power and other aspects, affecting the stable operation of high-voltage circuit breakers.

[0092] Therefore, multi-dimensional operating data is collected and the temperature of each high-voltage circuit breaker is recorded as T ij , pressure is P ij , vibration is V ij , the current is I ij , then number the high voltage circuit breaker W ij And the corresponding timestamp t, the operating data of each high-voltage circuit breaker is stored in real time as a high-voltage circuit breaker status dataset.

[0093] For example, taking a liquid argon vaporization station as an example, this method is used to monitor the operating status of high-voltage circuit breakers. Classification numbers are assigned according to the type of high-voltage circuit breakers. In the high-voltage circuit breakers of the distillation tower, unique high-voltage circuit breaker numbers are assigned in ascending order. The high-voltage circuit breaker number of the first distillation tower is W. 11 , the high voltage circuit breaker number of the second distillation tower is W 12 ,The 5-minute time interval is a common collection cycle adopted in many industry standards and practices.,Thus, the system collects multi-dimensional operating data of each high-voltage circuit breaker, including,temperature, pressure, vibration and current, at a 5-minute time interval, and stores them as a high-voltage circuit breaker status dataset according to the corresponding high-voltage circuit breaker number and timestamp.

[0094] The status scoring module constructs a dynamic weight distribution formula based on the high-voltage circuit breaker status dataset and the real-time changes in the high-voltage circuit breaker operating status. It outputs the health status score in real time and records it in the health status database.

[0095] Extract the operating data of the high-voltage circuit breaker status dataset during the operation of the liquid argon vaporization station and construct a dynamic weight distribution formula:

[0096] S ij (t) = ω1·f1(T ij (t))+ω2·f2(P ij (t))+ω3·f3(V ij (t))+ω4·f4(I ij (t))

[0097] Where S ij (t a ) represents the high voltage circuit breaker W ijHealth score at time t; f1, f2, f3, and f4 are calculated functions of temperature, pressure, vibration, and current, respectively. ω1, ω2, ω3, and ω4 are weight coefficients for the health score of the high-voltage circuit breaker, where ω1+ω2+ω3+ω4=1. The default values are ω1=0.3, ω2=0.4, ω3=0.2, and ω4=0.1. Different types of high-voltage circuit breakers have different sensitivities to different operating parameters. In different application scenarios, some parameters may be more critical to the health of the high-voltage circuit breaker than others. For example, a distillation tower operates at high temperature (liquid argon boiling point approximately -186°C) and high pressure (operating pressure can reach 2-3 MPa). Abnormal temperature and pressure fluctuations can directly lead to equipment damage or liquid argon leakage. Temperature and pressure may be more critical than vibration. In mechanical high-voltage circuit breakers, vibration may be a more important fault indicator. The weights of various parameters can be appropriately adjusted based on the characteristics of each high-voltage circuit breaker. For example, if a high-voltage circuit breaker is sensitive to temperature changes, a higher temperature weight may be assigned to it, but the order ω2>ω1>ω3>ω4>0 is always the same. Changes in each parameter directly affect the health score. Abnormalities in temperature, pressure, vibration, or current (exceeding safety thresholds) can significantly reduce the health score, reflecting a deterioration in the health of the high-voltage circuit breaker. By integrating dynamic weighted scores based on multiple factors, the health of the high-voltage circuit breaker can be assessed in real time.

[0098] The formula for the temperature calculation function is:

[0099]

[0100] Where, T ij，ref It is represented as the reference temperature of the high-voltage circuit breaker, α1 is a sensitivity constant ≥ 0, which represents the sensitivity of temperature to the health status of the high-voltage circuit breaker;

[0101] The formula for the pressure calculation function is:

[0102]

[0103] Where, P ij，ref It is expressed as the reference pressure of the high-voltage circuit breaker, α2 is a sensitivity constant ≥ 0, which represents the sensitivity of pressure to the health status of the high-voltage circuit breaker;

[0104] The formula for the vibration calculation function is:

[0105]

[0106] Where V ij，ref It is expressed as the reference vibration value of the high-voltage circuit breaker, γ1 and γ2 are sensitivity constants, where γ1<0, γ2>0, which represents the impact of vibration on the health status of the high-voltage circuit breaker;

[0107] The formula for the current calculation function is:

[0108]

[0109] Where, I ij，ref It represents the reference current of the high-voltage circuit breaker. α3 is a sensitivity constant ≥ 0, which represents the impact of the current on the health status of the high-voltage circuit breaker. The setting of the sensitivity constant reflects the sensitivity of the high-voltage circuit breaker to parameter changes under different working conditions. It is adjusted according to the operating data and actual maintenance status of the high-voltage circuit breaker to ensure that the health status assessment of the high-voltage circuit breaker under different operating conditions is more accurate.

[0110] In actual operation, various data of high-voltage circuit breakers may produce errors due to environmental interference, measurement instrument accuracy, etc. Collecting multiple sets of data and taking the average value can comprehensively consider the operating status of the high-voltage circuit breaker at different times, so that the sensitivity constant can adapt to this dynamic change and more comprehensively and accurately reflect the relationship between the health status of the high-voltage circuit breaker and the data indicators.

[0111] Therefore, when the health status is known to be good, n sets of operating data of the high-voltage circuit breaker are collected, and the collected temperature data of the high-voltage circuit breaker is recorded as T ijN , pressure data is P ijN , vibration data is V ijN , current data is I ijN , the health status score at this stage is S max is 1, where N = 1, 2, ..., n, n is a positive integer ≥ 2,

[0112] Compute the mean of temperature data:

[0113]

[0114] Where, It represents the average value of n sets of temperature data of the high-voltage circuit breaker collected during the stage when the high-voltage circuit breaker is known to be in good health, and n represents the number of operating data sets collected;

[0115] Construct the intermediate temperature formula:

[0116]

[0117] The calculation formula of the temperature sensitivity constant can be obtained as follows:

[0118]

[0119] Calculate the average of the pressure data:

[0120]

[0121] Where, It represents the average value of n sets of pressure data of the high-voltage circuit breaker collected during the known good health stage;

[0122] Construct the intermediate formula for pressure:

[0123]

[0124] The calculation formula of the pressure sensitivity constant can be obtained as follows:

[0125]

[0126] Calculate the average value of vibration data:

[0127]

[0128] Where, It represents the average value of n groups of vibration data of the high-voltage circuit breaker collected in the stage of known good health status; When , construct the vibration intermediate formula:

[0129]

[0130] The calculation formula of the vibration sensitivity constant can be obtained as follows:

[0131]

[0132] when When , construct the vibration intermediate formula:

[0133]

[0134] The calculation formula of the vibration sensitivity constant can be obtained as follows:

[0135]

[0136] Calculate the average value of the current data:

[0137]

[0138] Where, It represents the average value of n sets of current data collected for a high-voltage circuit breaker in a known good health state; construct the intermediate current formula:

[0139]

[0140] The calculation formula of the current sensitivity constant can be obtained as follows:

[0141]

[0142] Take the high voltage circuit breaker W of the distillation tower 11 For example, to calculate the health status score at a certain moment, the collected data includes: timestamp: 2025-01-09 10:00:00, T 11 =450, P 11 =2.5, V 11 =2.0,I 11 =200, reference temperature T ij , ref =450, reference pressure P ij，ref =2, reference vibration value V ij，ref =1.5, reference current I ij，ref =180. Under normal circumstances, the temperature operation has a more significant impact on the high-voltage circuit breaker of the distillation tower. Temperature fluctuations are related to combustion efficiency, thermal expansion performance and the structural safety of the distillation tower. Therefore, the temperature weight is set to ω1 = 0.35. During the operation of the distillation tower, the internal pressure is the key parameter to maintain the normal operation of the distillation tower. Too high or too low pressure may cause the distillation tower to malfunction, and in severe cases may cause an explosion. Therefore, the pressure weight is usually set to be the second, ω2 = 0.3. During the long-term operation of the high-voltage circuit breaker of the distillation tower, the abnormal changes in vibration This may cause wear of the high-voltage circuit breaker, loose connections, or other mechanical damage. The impact of vibration on the high-voltage circuit breaker of the distillation tower is relatively small compared to temperature and pressure, so the vibration weight is set to ω3 = 0.15. The current usually reflects the load condition of the high-voltage circuit breaker in the distillation tower. Excessive or insufficient current may indicate abnormal operation of the high-voltage circuit breaker in the distillation tower, indirectly affecting the working condition of the distillation tower. Therefore, the current weight is set to ω4 = 0.2. Among them, in the stage where the health status is known to be good, five sets of operating data of the high-voltage circuit breaker are collected, and the data average is calculated:

[0143] Temperature average:

[0144]

[0145] Average pressure:

[0146]

[0147] Vibration average value:

[0148]

[0149] Average current:

[0150]

[0151] Known S max =1, the temperature sensitivity constant calculation formula can be obtained:

[0152]

[0153] The pressure sensitivity constant calculation formula can be obtained:

[0154]

[0155] The vibration sensitivity constant calculation formula can be obtained:

[0156]

[0157] The current sensitivity constant calculation formula can be obtained:

[0158]

[0159] Among them, the result of the temperature calculation function is:

[0160]

[0161] The result of the pressure calculation function is:

[0162]

[0163] The result of the vibration calculation function is:

[0164] f3(V 11 (t))=1+γ2·(V 11 (t)-V 11，ref )=1+32.2×(2.0-1.5)=17.1

[0165] The result of the current calculation function is:

[0166] f4(I 11 (t))=1+α3·(I 11 (t)-I 11，ref )=1+0.57×(200-180)=12.4

[0167] Therefore, the health status score of the high-voltage circuit breaker of the distillation tower at this moment is:

[0168] S 11 (t) = ω1·f1(T 11 (t))+ω2·f2(P 11 (t))+ω3·f3(V 11 (t))+ω4·f4(I 11 (t))

[0169] =0.35×1+0.3×5.228+0.15×17.1+0.2×12.4

[0170] =0.35+1.568+2.565+2.48=6.963

[0171] Finally, the health status score is output in real time and recorded in the health status database.

[0172] The fault prediction module analyzes the high-voltage circuit breaker status dataset, extracts the operating data of the fault status to construct an abnormal dataset, and constructs an LSTM model based on the real-time changes in the health status score and the abnormal dataset to predict the future failure probability of the high-voltage circuit breaker.

[0173] The high-voltage circuit breaker status dataset is used to extract the operating data of each high-voltage circuit breaker's fault state. When the temperature, pressure, vibration, and current exceed the safe range, they are marked as faulty. A fault dataset is created. This dataset includes the operating data of the high-voltage circuit breaker marked as faulty from 12 hours before to 12 hours after the fault occurs, along with the corresponding high-voltage circuit breaker number and timestamp. Based on the operating data collected during this time period, the health status score of the high-voltage circuit breaker is calculated according to the dynamic weight allocation formula.

[0174] Afterwards, the operating data of the fault dataset and the corresponding high-voltage circuit breaker number, timestamp, and calculated health status score are used as input features.

[0175] The time step is:

[0176]

[0177] Where timesteps represents the time step of the LSTM model, T represents the time duration of the fault state, and t gd It represents the time interval for collecting multi-dimensional operation data;

[0178] Extract a set of 24-hour temperature, pressure, vibration, and current operating data of a known healthy stage and perform Min-Max normalization processing:

[0179]

[0180] Where, T norm Represented as normalized temperature data, min(T ij ) represents the minimum value of the temperature data in this group, maxT ij Shown is the maximum value of the temperature data in this group;

[0181]

[0182] Where, P norm Expressed as normalized pressure data, min(P ij) represents the minimum value of the pressure data in this group, maxP ij Shown is the maximum value of the pressure data in this group;

[0183]

[0184] Where V norm Expressed as normalized vibration data, min(V ij ) represents the minimum value of the vibration data in this group, maxV ij Shown as the maximum value of the vibration data in this group;

[0185]

[0186] middle, I norm Expressed as normalized current data, min(I ij ) represents the minimum value of the current data in this group, maxI ij Shown is the maximum value of the current data in this group;

[0187] The fault dataset is divided into multiple time series samples in chronological order. A sliding window with a window size of timesteps is set, sliding one time step at a time. The data at the previous time point serves as the input features (training data), and the data at the next time point serves as the output label (prediction result), which is the probability of future failure of the high-voltage circuit breaker, with a value range of [0, 1]. Since the LSTM model is suitable for time series data, a sliding window is created. The window size depends on the requirements of the prediction model. The data sampling interval is 5 minutes (adjustable according to actual needs), so each window contains data from the past 60 minutes, or 12 data points. High-voltage circuit breakers marked as faulty are labeled 1, and those marked as healthy are labeled 0. The input layer includes multiple cloud top data (temperature, pressure, vibration, current) and health status scores of the high-voltage circuit breakers. The LSTM layer uses two LSTM layers, with 128 hidden units in each layer. The LSTM layer processes the input time series data and captures long-term dependencies in the data. The batch size was set to 32-128, and the number of training epochs was set to 50-200. Dropout layer: A dropout layer was added after the LSTM layer with a ratio of 0.2 to prevent overfitting. Fully connected layers: Two fully connected layers, the last of which uses a sigmoid activation function to output the probability of future high-voltage circuit breaker failure. Loss function: A binary cross-entropy loss function was used, suitable for binary classification problems. Optimizer: The Adam optimizer was used with an initial learning rate of 0.001. Output layer: A sigmoid activation function was used to output the probability of high-voltage circuit breaker failure in the range [0, 1].

[0188] The fault judgment module pre-sets the fault threshold, and marks the high-voltage circuit breakers with a predicted fault probability higher than the fault threshold as potential fault high-voltage circuit breakers, generates a maintenance plan and adjusts the operating parameters of the potential fault high-voltage circuit breakers.

[0189] Using the trained LSTM model, we predict the future probability of each high-voltage circuit breaker failure on the test dataset. In machine learning and deep learning, setting the threshold to 0.5 is a common practice, especially in binary classification problems. The model output is the probability of a high-voltage circuit breaker failure. The model output probability value is typically in the range [0, 1], representing the probability of a high-voltage circuit breaker failure.

[0190] If the output probability is greater than 0.5, it means that the high-voltage circuit breaker is more likely to fail.

[0191] If the output probability is less than 0.5, it means that the high-voltage circuit breaker is more likely to be in a normal operating state.

[0192] During the operation of the liquid argon vaporization station, in order to promptly and effectively deal with possible failures of the high-voltage circuit breaker, the fault threshold is pre-set as P thresh =0.5, but according to the actual situation, the threshold can be adjusted according to business needs. For high-voltage circuit breaker W ij , when the future failure probability of the high-voltage circuit breaker P ij >P thresh , then the high voltage circuit breaker W ij High-voltage circuit breakers identified as potential failures. Based on the potential failure prediction results for high-voltage circuit breakers, the DCS issues a series of instructions to adjust the operating parameters of the high-voltage circuit breakers to reduce their workload and lower the risk of failure. Therefore, for high-voltage circuit breakers identified as potential failures, the DCS issues adjustment instructions to modify the operating parameters of the high-voltage circuit breakers, increasing the cooling efficiency of the high-voltage circuit breakers, reducing the operating pressure, reducing the speed of the high-voltage circuit breakers, and reducing the load current of the high-voltage circuit breakers. At the same time, an early warning is issued to generate a maintenance plan.

[0193] See also Figure 2 The present invention further provides a DCS system maintenance method based on networked intelligent monitoring, which is executed by the above-mentioned DCS system maintenance system based on networked intelligent monitoring, and includes:

[0194] Step 1: Number the high-voltage circuit breakers in the DCS system in an incremental manner, collect multi-dimensional operating data at fixed time intervals, and store them as high-voltage circuit breaker status datasets according to the number and time stamp. The operating data includes temperature, pressure, vibration, and current.

[0195] Step 2: Based on the high-voltage circuit breaker status dataset and the real-time changes in the high-voltage circuit breaker operating status, a dynamic weight distribution formula is constructed to output the health status score in real time and record it in the health status database;

[0196] Step 3: Analyze the high-voltage circuit breaker status dataset, extract the operating data of the fault state to construct an abnormal dataset, and build an LSTM model based on the real-time changes in the health status score and the abnormal dataset to predict the future failure probability of the high-voltage circuit breaker.

[0197] Step 4: Pre-set a fault threshold, mark the high-voltage circuit breakers with a predicted fault probability higher than the fault threshold as potential fault high-voltage circuit breakers, generate a maintenance plan and adjust the operating parameters of the potential fault high-voltage circuit breakers.

[0198] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0199] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0200] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A DCS system maintenance system based on networked intelligent monitoring, characterized in that: include: A status acquisition module is used to number the high-voltage circuit breakers in the DCS system in an incremental manner, collect multi-dimensional operating data at fixed time intervals, and store them as high-voltage circuit breaker status data sets by number and time stamp. The operating data includes temperature, pressure, vibration, and current. The status scoring module is used to construct a dynamic weight distribution formula based on the high-voltage circuit breaker status data set and the real-time changes in the high-voltage circuit breaker operating status, output the health status score in real time, and record it in the health status database; The fault prediction module is used to analyze the high-voltage circuit breaker status data set, extract the operating data of the fault state to construct an abnormal data set, and build an LSTM model based on the real-time changes in the health status score and the abnormal data set to predict the future failure probability of the high-voltage circuit breaker; The fault judgment module is used to pre-set a fault threshold, mark the high-voltage circuit breaker with a predicted fault probability higher than the fault threshold as a potential fault high-voltage circuit breaker, generate a maintenance plan and adjust the operating parameters of the potential fault high-voltage circuit breaker.

2. A DCS system maintenance system based on networked intelligent monitoring according to claim 1, characterized in that: Numbering the high-voltage circuit breakers in an incremental manner includes the following steps: Assign independent classification identification to high-voltage circuit breakers, and then assign unique incremental numbers in turn, that is, assign a unique high-voltage circuit breaker number W to each high-voltage circuit breaker. ij , where i represents the classification number of the high-voltage circuit breaker and j represents the increasing serial number under the classification.

3. A DCS system maintenance system based on networked intelligent monitoring according to claim 2, characterized in that: Storing the data as a high-voltage circuit breaker status dataset by number and timestamp includes the following steps: At fixed time intervals t gd For the data collection cycle, collect multi-dimensional operating data and record the temperature of each high-voltage circuit breaker as T ij , pressure is P ij , vibration is V ij , the current is I ij , then number the high voltage circuit breaker W ij and the corresponding timestamp t, the operating data of each high-voltage circuit breaker is stored as a high-voltage circuit breaker status dataset in real time, where the time interval t gd Given by experience, the default is 5 minutes.

4. A DCS system maintenance system based on networked intelligent monitoring according to claim 3, characterized in that: Constructing a dynamic weight distribution formula includes the following steps: Extract the operating data of the high-voltage circuit breaker status dataset and construct a dynamic weight distribution formula: S ij (t)=ω1·f1(T ij (t))+ω2·f2(P ij (t))+ω3·f3(V ij (t))+ω4·f4(I ij (t)) Where S ij (t a ) represents the high voltage circuit breaker W ij Health status score at time t; f1, f2, f3 and f4 are the calculation functions of temperature, pressure, vibration and current respectively, ω1, ω2, ω3 and ω4 are the weight coefficients of the health status score of the high-voltage circuit breaker, and ω1+ω2+ω3+ω4=1, ω2>ω1>ω3>ω4>0; The formula for the temperature calculation function is: Where, T ij，ref It is represented as the reference temperature of the high-voltage circuit breaker, α1 is a sensitivity constant ≥ 0, which represents the sensitivity of temperature to the health status of the high-voltage circuit breaker; The formula for the pressure calculation function is: Where, P ij，ref It is expressed as the reference pressure of the high-voltage circuit breaker, α2 is a sensitivity constant ≥ 0, which represents the sensitivity of pressure to the health status of the high-voltage circuit breaker; The formula for the vibration calculation function is: Where V ij，ref It is expressed as the reference vibration value of the high-voltage circuit breaker, γ1 and γ2 are sensitivity constants, where γ1<0, γ2>0, which represents the impact of vibration on the health status of the high-voltage circuit breaker; The formula for the current calculation function is: Where, I ij，ref It is represented as the reference current of the high-voltage circuit breaker, and α3 is a sensitivity constant ≥ 0, which represents the impact of the current on the health status of the high-voltage circuit breaker.

5. A DCS system maintenance system based on networked intelligent monitoring according to claim 4, characterized in that: The algorithm method for the sensitivity constant is: In the stage where the health status is known to be good, collect n sets of operating data of the high-voltage circuit breaker, and record the collected temperature data of the high-voltage circuit breaker as T ijN , pressure data is P ijN , vibration data is V ijN , current data is I ijN , the health status score at this stage is S max is 1, where N = 1, 2, ..., n, n is a positive integer ≥ 2, Compute the mean of temperature data: Where, It represents the average value of n sets of temperature data of the high-voltage circuit breaker collected during the stage when the high-voltage circuit breaker is known to be in good health, and n represents the number of operating data sets collected; Construct the intermediate temperature formula: The calculation formula of the temperature sensitivity constant can be obtained as follows: Calculate the average of the pressure data: Where, It represents the average value of n sets of pressure data of the high-voltage circuit breaker collected during the known good health stage; Construct the intermediate formula for pressure: The calculation formula of the pressure sensitivity constant can be obtained as follows: Calculate the average value of vibration data: Where, It represents the average value of n sets of vibration data of the high-voltage circuit breaker collected during the known good health stage; when When , construct the vibration intermediate formula: The calculation formula of the vibration sensitivity constant can be obtained as follows: when When , construct the vibration intermediate formula: The calculation formula of the vibration sensitivity constant can be obtained as follows: Calculate the average value of the current data: Where, It represents the average value of n sets of current data of the high-voltage circuit breaker collected during the known good health stage; Construct the intermediate formula for current: The calculation formula of the current sensitivity constant can be obtained as follows:

6. A DCS system maintenance system based on networked intelligent monitoring according to claim 1, characterized in that: The method for constructing an LSTM model to predict the future failure probability of a high-voltage circuit breaker is as follows: The high-voltage circuit breaker status dataset is used to extract the operating data of each high-voltage circuit breaker's fault state. When the temperature, pressure, vibration, and current exceed the safe range, they are marked as faulty. A fault dataset is created. This dataset includes the operating data of the high-voltage circuit breaker marked as faulty from 12 hours before to 12 hours after the fault occurs, along with the corresponding high-voltage circuit breaker number and timestamp. Based on the operating data collected during this time period, the health status score of the high-voltage circuit breaker is calculated according to the dynamic weight allocation formula. Afterwards, the operating data of the fault dataset and the corresponding high-voltage circuit breaker number, timestamp, and calculated health status score are used as input features. The time step is: Where timesteps represents the time step of the LSTM model, T represents the time duration of the fault state, and t gd It represents the time interval for collecting multi-dimensional operation data; The fault data set is divided into multiple time series samples in chronological order. A sliding window with a window size of timesteps is set. Each time it slides, it takes one time step. The data at the previous time point is used as the input feature (training data), and the data at the next time point is used as the output label (prediction result), that is, the probability of future failure of the high-voltage circuit breaker, and the value range is [0, 1]. The number of LSTM hidden layer units is set to 64, and the activation function is the hyperbolic tangent activation function. The number of LSTM layers is 2, and the number of hidden units in each LSTM layer is set to 128. Two fully connected layers are added after the LSTM layer. The last one is The layer uses the sigmoid activation function to output the probability of fault occurrence. The loss function uses the binary cross entropy loss function. The Adam optimizer is used, the initial learning rate is set to 0.001, the batch size is set to 32-128, and the training rounds are set to 50-200 rounds. A Dropout layer is added after each LSTM layer and fully connected layer with a ratio of 0.2 and a gradient threshold of 5. The future fault probability of the high-voltage circuit breaker is used as the output label with a value range of [0,1]. An LSTM model is constructed. After training, the trained model is used to predict the next time point to obtain the predicted fault probability.

7. The DCS system maintenance system based on networked intelligent monitoring according to claim 1 is characterized by: Marking a high-voltage circuit breaker with a predicted fault probability higher than a fault threshold as a potential fault high-voltage circuit breaker includes the following steps: Use the trained LSTM model to predict the test data set and obtain the future failure probability of each high-voltage circuit breaker. The fault threshold is pre-set as P thresh , for high voltage circuit breaker W ij , when the future failure probability of the high-voltage circuit breaker P ij >P thresh , then the high voltage circuit breaker W ij Marked as a potential fault high voltage circuit breaker.

8. A DCS system maintenance system based on networked intelligent monitoring according to claim 7, characterized in that: Generating a maintenance plan and adjusting the operating parameters of a potentially faulty high-voltage circuit breaker includes the following steps: For high-voltage circuit breakers identified as potential faults, adjustment instructions are issued through the DCS to modify the operating parameters of the high-voltage circuit breakers, namely increasing the cooling efficiency of the high-voltage circuit breakers, reducing the working pressure, reducing the speed of the high-voltage circuit breakers, and reducing the load current of the high-voltage circuit breakers. At the same time, an early warning is issued to generate a maintenance plan.

9. A DCS system maintenance method based on networked intelligent monitoring, characterized in that: The maintenance method is performed by a DCS system maintenance system based on networked intelligent monitoring according to any one of claims 1 to 8, and the specific steps include: Step 1: Number the high-voltage circuit breakers in the DCS system in an incremental manner, collect multi-dimensional operating data at fixed time intervals, and store them as high-voltage circuit breaker status datasets according to the number and time stamp. The operating data includes temperature, pressure, vibration, and current. Step 2: Based on the high-voltage circuit breaker status dataset and the real-time changes in the high-voltage circuit breaker operating status, a dynamic weight distribution formula is constructed to output the health status score in real time and record it in the health status database; Step 3: Analyze the high-voltage circuit breaker status dataset, extract the operating data of the fault state to construct an abnormal dataset, and build an LSTM model based on the real-time changes in the health status score and the abnormal dataset to predict the future failure probability of the high-voltage circuit breaker. Step 4: Pre-set a fault threshold, mark the high-voltage circuit breakers with a predicted fault probability higher than the fault threshold as potential fault high-voltage circuit breakers, generate a maintenance plan and adjust the operating parameters of the potential fault high-voltage circuit breakers.

Images