A heat pipe safety risk assessment method, system, device and program

Abstract

The application discloses a heat pipeline safety risk assessment method, system, device and program, relates to the technical field of risk assessment, and comprises the following steps: acquiring heat pipeline data; the heat pipeline data comprises static data and time series data of the heat pipeline; wherein the time series data comprises mechanical vibration data; preprocessing the heat pipeline data; feature extraction is performed based on the preprocessed heat pipeline data to obtain static data features and time series data features; the static data features and the time series data features are input into a machine learning fusion model to output a risk probability; vibration fatigue damage is calculated based on the mechanical vibration data features; the risk probability is corrected based on the vibration fatigue damage to obtain a corrected risk probability; and the risk level is determined based on the corrected risk probability. Through the processing scheme, the real risk state of the pipeline can be reflected, and the risk level can be accurately evaluated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of risk assessment, in particular to a heat pipe safety risk assessment method, system, device and program. BACKGROUND

[0002] With the rapid development of urban heating industry in China, the heat industry involves heat source production, pipeline transportation, heat user supply and other links, and each link has potential safety risks. Especially, as an important urban infrastructure, the operation safety of heat pipe is directly related to the life of residents, public safety and social stability. Once a failure accident occurs, it may cause heat leakage, personnel scalding, equipment damage and even secondary disasters, which will have a serious impact on society.

[0003] In recent years, safety accidents caused by aging, corrosion and external damage of heat pipes have occurred from time to time, such as heat loss caused by pipe rupture and damage to surrounding facilities, which makes the safety of heat pipes face more and more severe challenges. Therefore, it is imperative to strengthen the safety risk assessment of urban heat pipes, which is of great significance to ensure the safe and stable operation of urban heating systems.

[0004] The existing technology has obvious limitations in heat pipe risk assessment, mainly in the following aspects:

[0005] 1. Single evaluation dimension, difficult to fully reflect the risk. Traditional methods focus on static features (such as material, service life) or single dynamic parameters (such as pressure), ignoring the correlation between static inherent properties and dynamic operating state.

[0006] 2. Insufficient model adaptability, weak time sequence correlation capture ability. Existing technologies often use logistic regression, static decision tree and other models, which are difficult to handle the time dependence of time series data such as temperature and vibration.

[0007] 3. Weight mechanism is rigid and lacks dynamic adaptability. Most fusion models use fixed weights and do not consider the difference in working condition fluctuations. When the pipe vibration is severe, the dynamic risk proportion will be underestimated if the fixed weight is used, which will lead to delayed warning.

[0008] 4. Key risk sources are not fully covered, and mechanical vibration factors are ignored. Traditional evaluation focuses on temperature and pressure, and pays insufficient attention to mechanical impact parameters such as vibration acceleration. Vibration caused by loose supports and unbalanced equipment is often an important inducement for pipe fatigue cracking, resulting in blind spots in risk tracing.

[0009] Therefore, the above-mentioned existing heat pipe risk assessment method has obvious inconvenience and defects, and needs to be further improved. How to create a new heat pipe risk assessment method has become the current industry's goal for improvement. SUMMARY

[0010] Therefore, the embodiment of the present disclosure provides a heat pipe safety risk assessment method, which at least partially solves the problems in the prior art.

[0011] In a first aspect, the embodiment of the present disclosure provides a heat pipe safety risk assessment method, which comprises the following steps:

[0012] Obtaining heat pipe data; the heat pipe data comprises static data and time series data of the heat pipe; wherein the time series data comprises mechanical vibration data;

[0013] Pretreating the heat pipe data;

[0014] Extracting features based on the pretreated heat pipe data to obtain static data features and time series data features;

[0015] Inputting the static data features and the time series data features into a machine learning fusion model to output a risk probability;

[0016] Calculating vibration fatigue damage based on the mechanical vibration data features;

[0017] Risk correcting the risk probability based on the vibration fatigue damage to obtain a corrected risk probability;

[0018] Determining a risk level based on the corrected risk probability.

[0019] According to a specific implementation manner of the embodiment of the present disclosure, the pretreating the heat pipe data comprises filling missing values, removing outliers and standardizing the heat pipe data;

[0020] The method further comprises:

[0021] Verifying the authenticity of the static data;

[0022] Aligning the time series data;

[0023] Smoothing the mechanical vibration data.

[0024] According to a specific implementation manner of the embodiment of the present disclosure, the extracting features based on the pretreated heat pipe data to obtain static data features and time series data features comprises:

[0025] Extracting static data features, comprising:

[0026] Using the standardized value of the pretreated static data as the static data features;

[0027] Calculating the coupling features of the heat pipe material and the service life based on the following formula:

[0028] ;

[0029] wherein, is the coupling characteristic of heat pipe material and service life; is the standardized value of pipe material yield strength; is the standardized value of pipe service life; 0.5 is the attenuation coefficient;

[0030] The time series data feature extraction includes:

[0031] The vibration-fault correlation characteristic is calculated based on the following formula:

[0032] ;

[0033] wherein, is the vibration-fault correlation characteristic; is the number of fatigue cracks of similar pipes in the past 3 years; is the average vibration fatigue damage of the pipe in the past 30 days; is the fatigue damage threshold.

[0034] According to a specific implementation manner of the embodiment of the present disclosure, the static data feature and the time series data feature are input into a machine learning fusion model, and a risk probability is output, which includes:

[0035] An LSTM model and an XGBoost model are established;

[0036] The time series data feature is input into the LSTM model, and a first probability value is output ;

[0037] The static data feature is input into the XGBoost model, and a second probability ;

[0038] The first probability and the second probability are fused based on the following formula:

[0039] ;

[0040] wherein, is the fusion probability; is the LSTM model weight, is the XGBoost model weight, and ; is the LSTM model probability; is the XGBoost model probability.

[0041] According to a specific implementation of an embodiment of the present disclosure, the method further comprises: optimizing the weights α and β through cross-validation.

[0042] The optimization of the weights α and β through cross-validation comprises:

[0043] Selecting pipeline sample historical data containing known failure and normal labels;

[0044] Dividing the sample historical data into a training set and a validation set;

[0045] Calculating different weights Corresponding fusion probability Selecting α and β that make the prediction accuracy the highest as initial weights.

[0046] According to a specific implementation of an embodiment of the present disclosure, the calculation of the vibration fatigue damage based on the mechanical vibration data features comprises:

[0047] The vibration fatigue damage is calculated based on the following formula:

[0048] ;

[0049] Wherein, is the vibration fatigue damage; is the vibration load grade number; is the grade vibration sequence number; is the cycle number of the 24-hour vibration of the first grade vibration; is the material fatigue life corresponding to the first grade vibration;

[0050] The material fatigue life is calculated based on the following formula :

[0051] ;

[0052] Wherein, is the material symmetric cyclic fatigue limit; is the material fatigue index, which can be obtained by consulting industry standards; is the stress amplitude corresponding to the first grade vibration.

[0053] According to a specific implementation of an embodiment of the present disclosure, the risk probability is corrected based on the vibration fatigue damage, to obtain a corrected risk probability, comprising:

[0054] When the vibration fatigue damage is greater than or equal to 1, the risk is corrected based on the following formula:

[0055] ;

[0056] wherein, is a revised risk probability; is a risk probability; is a vibration correction coefficient, ;

[0057] The vibration correction coefficient is calculated based on the following formula :

[0058] ;

[0059] wherein, is a vibration fatigue damage;

[0060] When the vibration fatigue damage is greater than or equal to a threshold value, the vibration correction coefficient , .

[0061] In a second aspect, the embodiments of the present disclosure provide a heat pipe safety risk assessment system, the system comprising:

[0062] a data acquisition module configured to acquire heat pipe data; the heat pipe data comprising static data and time series data of the heat pipe; wherein the time series data comprises mechanical vibration data;

[0063] a preprocessing module configured to preprocess the heat pipe data;

[0064] a feature extraction module configured to perform feature extraction based on the preprocessed heat pipe data, to obtain static data features and time series data features;

[0065] a machine learning module configured to input the static data features and the time series data features into a machine learning fusion model, to output a risk probability; to calculate a vibration fatigue damage based on the mechanical vibration data features; and to revise the risk probability based on the vibration fatigue damage, to obtain a revised risk probability;

[0066] a risk judgment module configured to determine a risk level based on the revised risk probability.

[0067] In a third aspect, the embodiments of the present disclosure further provide an electronic device, the electronic device comprising:

[0068] at least one processor; and,

[0069] a memory in communication with the at least one processor; wherein,

[0070] The memory stores instructions capable of being executed by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one processor to perform the heat pipe safety risk assessment method in the first aspect or any implementation manner of the first aspect.

[0071] In a fourth aspect, the embodiments of the present disclosure further provide a non-transitory computer-readable storage medium storing computer instructions, which, when executed by at least one processor, cause the at least one processor to perform the heat pipe safety risk assessment method in the first aspect or any implementation manner of the first aspect.

[0072] In a fifth aspect, the embodiments of the present disclosure further provide a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, which, when executed by a computer, cause the computer to perform the heat pipe safety risk assessment method in the first aspect or any implementation manner of the first aspect.

[0073] The heat pipe safety risk assessment method in the embodiments of the present disclosure comprehensively reflects the real risk state of the pipe through multi-dimensional risk assessment fusion. The XGBoost model is adapted to static characteristics, accurately captures cumulative effects such as long-term aging and material decay, the LSTM model is good at processing time series data, effectively identifies short-term dynamic risks such as temperature sudden change, pressure fluctuation, and vibration anomaly, and the two types of models cooperate to improve the assessment accuracy; through feature extraction such as vibration acceleration, frequency, and amplitude, the forward-looking of risk early warning is improved, an intuitive risk probability is output, and the risk source is quickly located by the operation and maintenance personnel, measures are taken in advance, the accident rate is reduced, and strong support is provided for the safety operation and maintenance of the heat pipe. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 A heat pipe safety risk assessment method flowchart is provided for the embodiments of the present disclosure.

[0075] Figure 2 A heat pipe safety risk assessment method flowchart is provided for the embodiments of the present disclosure.

[0076] Figure 3 A heat pipe safety risk assessment system structure diagram is provided for the embodiments of the present disclosure.

[0077] Figure 4 An electronic device diagram is provided for the embodiments of the present disclosure. DETAILED DESCRIPTION

[0078] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0079] The following detailed description illustrates embodiments of the disclosure by way of example and not by way of limitation. The description makes reference to the accompanying drawings in which: The description will be more fully understood with reference to the drawings and the following detailed description, in which: It should be noted that the description is merely illustrative of the disclosure and should not be taken in a limiting sense. The description together with the drawings makes the disclosure accessible to those skilled in the art. Various embodiments of the disclosure are described hereinafter with reference to the drawings. While the disclosure is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It should be understood, however, that the intention is not to limit the disclosure to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the disclosure as defined by the appended claims.

[0080] It is to be understood that the aspects described herein can be implemented in various forms of hardware, software, firmware, special purpose processes, or a combination of these, and that the described aspects can be implemented differently depending upon the particular desires of the implementor. It is to be further understood that the aspects described herein can be implemented by firmware or software operating in conjunction with an application specific integrated circuit or a general purpose central processing unit. It is to be further understood that the aspects described herein can be implemented by a combination of hardware and software. It is to be further understood that the aspects described herein can be implemented as a computer program product, which can be executed on a computer or a general purpose processor. It is to be further understood that the aspects described herein can be implemented as a computer program product, which can be executed on a computer or a general purpose processor.

[0081] In addition, in the following description, specific details are provided to thoroughly understand examples. However, one of ordinary skill in the art will understand that the described aspects can be practiced without these specific details.

[0082] Figure 1 A schematic diagram of a heat pipe safety risk assessment method process is provided for the embodiments of the disclosure.

[0083] Figure 2 A corresponding heat pipe safety risk assessment method process block diagram is provided for Figure 1 A corresponding heat pipe safety risk assessment method process block diagram is provided for

[0084] As shown in Figure 1 At step S110, heat pipe data is acquired; the heat pipe data includes static data and time series data of the heat pipe.

[0085] More specifically, the static data includes: nominal diameter, design wall thickness, actual residual wall thickness, material yield strength and service life of the heat pipe; the time series data includes: real-time temperature of the medium in the heat pipe, real-time pressure of the medium, soil corrosion rate and mechanical vibration data, the mechanical vibration data includes: instantaneous vibration acceleration, vibration frequency and vibration amplitude of the outer wall of the pipe.

[0086] The heat pipe data also includes: the number of cracks generated by vibration in the same material and working condition pipe in the past three years.

[0087] The specific information of the collected data is shown in Table 1.

[0088] Table 1: Heat pipe data

[0089]

[0090] The specific acquisition method of the heat pipe data can include sensor collection, manual collection, data query and table research, and the like, and does not serve as the focus of the present application and will not be described here in detail.

[0091] In the embodiments of the present application, the acquired data further includes other data necessary for calculation, which is not limited here.

[0092] More specifically, next turn to step S120.

[0093] At step S120, the heat pipe data is preprocessed.

[0094] In the embodiments of the present application, the preprocessing of the heat pipe data includes: missing value filling, abnormal value elimination and standardization processing of the heat pipe data; the method further includes: real-time inspection of the static data; data alignment of the time series data; signal smoothing processing of the mechanical vibration data.

[0095] More specifically, the missing value filling includes: the missing value filling of the time series data adopts forward filling, that is, using the valid observation value at the previous time of the missing value to fill the current missing value; the static data adopts the mean value filling of the same type pipe, and when there is no same batch data, the recommended value of the industry standard is used.

[0096] The abnormal value elimination includes: based on the 3σ criterion, the abnormal value elimination of the time series data and the static data.

[0097] The standardization processing includes:

[0098] The parameters of the risk correlation direction positive correlation (that is, the greater the value, the lower the risk), for example, the heat pipe design wall thickness, the actual residual wall thickness and the material yield strength, are standardized based on the following formula respectively:

[0099]

[0100] Wherein, is the standardized value; is the actual value of the parameter, is the industry safety lower limit of the parameter; is the industry safety upper limit of the parameter; and The industry standard file can be consulted.

[0101] The parameters with a negative correlation with risk (i.e., the greater the value, the higher the risk) such as the service life of the heat pipe, the real-time temperature of the medium, the real-time pressure of the medium, the soil corrosion rate, the instantaneous vibration acceleration of the outer wall of the pipe, the vibration frequency and the vibration amplitude and the historical failure are standardized based on the following formula respectively:

[0102]

[0103] wherein, is the standardized value; is the actual value of the parameter, is the lower limit of the industry safety of the parameter; is the upper limit of the industry safety of the parameter; and The industry standard documents can be consulted.

[0104] Static data is the inherent property of the pipe, and if there is an input error, it is easy to cause misjudgment of the material resistance to risk.

[0105] Therefore, the authenticity of the static data needs to be verified, which specifically includes: cross comparison of the collected static data with the pipe design drawings, factory reports and historical detection records. When the deviation of the nominal diameter of the pipe from the design drawing is more than ±5mm, it is considered to be a authenticity error, and needs to be resampled; when the deviation of the yield strength of the material from the factory report is more than ±10MPa, it is considered to be a authenticity error, and needs to be resampled; when the change amount of the actual residual wall thickness compared with the last detection value is greater than 0.5mm / year, it is considered to be a authenticity error, and needs to be resampled, and the detection equipment precision needs to be reviewed.

[0106] The machine learning model needs to unify the time series data with the same time granularity, and the time axis misalignment will cause feature-time matching error.

[0107] Therefore, the time series data needs to be aligned, which specifically includes the following steps:

[0108] The high-frequency data (real-time temperature of the medium, real-time pressure of the medium) is down-sampled by hour, and the time axis is aligned with the low-frequency data (soil corrosion rate);

[0109] The entire time series data is divided into continuous and non-overlapping time windows in units of hours, and each window corresponds to 1 hour; all sampling values in a single window are extracted;

[0110] The hourly mean value of each window is calculated as the down-sampled feature value.

[0111] The hourly mean value of the data is calculated based on the following formula respectively:

[0112]

[0113] wherein, is the mean value of data in the first hour window; is the sampling frequency of 1 hour; is all the sampling values of data in the time window; is the serial number of the sampling; in the present application, the data sampling frequency for calculating the hourly mean value is higher than 1 time per hour.

[0114] The mechanical vibration data is subjected to signal smoothing, including: the mechanical vibration data is subjected to signal smoothing by a five-point sliding average method.

[0115] Next, go to step S130.

[0116] At step S130, feature extraction is performed based on the preprocessed heat pipe data to obtain static data features and time series data features.

[0117] In the embodiments of the present application, the feature extraction based on the preprocessed heat pipe data to obtain static data features and time series data features includes:

[0118] The static data feature extraction includes:

[0119] The standardized value of the preprocessed static data is used as the static data feature;

[0120] The coupling feature of the heat pipe material and the service life is calculated based on the following formula:

[0121] ;

[0122] wherein, is the coupling feature of the heat pipe material and the service life; is the standardized value of the yield strength of the pipe material; is the standardized value of the service life of the pipe; 0.5 is the attenuation coefficient, which is obtained according to experience;

[0123] The time series data feature extraction includes:

[0124] The vibration-failure correlation feature is calculated based on the following formula:

[0125] ;

[0126] wherein, is the vibration-failure correlation feature; is the number of fatigue cracks of the same type of pipe in the past 3 years, which is obtained by statistical historical data; is the average vibration fatigue damage of the pipe in the past 30 days (D ); The fatigue damage criticality is 0.2. With 3 years as the fatigue life benchmark, the vibration damage intensity in the past 30 days contributes to the degree of vibration-failure correlation, and the scale connection between long-term fatigue life and short-term damage statistics is realized.

[0127] To convert the coupling relationship between damage rate and failure time scale into a comparable correlation feature, a fatigue damage criticality is introduced. When the fatigue damage criticality is 0.2, the vibration fatigue contribution to failure enters a significant correlation interval, and the vibration-failure correlation intensity of different pipelines is measured based on this benchmark. Experts in the field refer to standards such as API 579-1 / ASME FFS-1 “Fitness for Service” and ASME B31.8S “Gas Pipeline System Integrity Management” to determine that the 5 months before the pipeline accumulates to a critical failure state due to vibration fatigue is the critical point at which vibration fatigue has a significant driving effect on failure. That is, , and the benchmark value is 0.2.

[0128] Next, go to step S140.

[0129] At step S140, the static data features and time series data features are input into a machine learning fusion model, and a risk probability is output.

[0130] In the embodiment of the present application, the input of the static data features and the time series data features into the machine learning fusion model and the output of the risk probability include: establishing an LSTM model and an XGBoost model; inputting the time series data features into the LSTM model to output a first probability value ; inputting the static data features into the XGBoost model to output a second probability ; and fusing the first probability and the second probability based on the following formula:

[0131]

[0132] wherein, is the fusion probability; is the LSTM model weight, is the XGBoost model weight, and ; is the LSTM model probability; is the XGBoost model probability.

[0133] More specifically, a single LSTM model cannot effectively process static features (such as wall thickness not changing over time), and a single XGBoost model cannot process time series fluctuations (such as instantaneous peak values of vibration); the fusion model can take into account both types of features, and the weighting weight can adapt to the intensity of time series fluctuations, improving the accuracy of risk probability calculation.

[0134] The model architecture is established, specifically including the following steps:

[0135] 1. Establishing an LSTM model and an XGBoost model.

[0136] The LSTM model includes an input layer, an LSTM layer, a fully connected layer, and an output layer.

[0137] The input layer is 288x3, i.e., each input sample contains 288 consecutive time point data (corresponding to 24 hours, with 1 collection every 5 minutes), and each time step contains 3 features: the standardized value of medium temperature, the standardized value of medium pressure, and the standardized value of mechanical vibration acceleration.

[0138] The LSTM layer has 64 neurons, and the dropout is set to 0.2 to prevent overfitting.

[0139] The fully connected layer has 32 neurons, which compresses the feature dimension from 64 to 32 to simplify subsequent calculations.

[0140] The output layer has a sigmoid activation function, and the output probability is , which represents the probability of pipeline failure within the next 1 hour. When the output value is closer to 1, it indicates a higher confidence level of "risk occurrence", and when it is closer to 0, it indicates a higher confidence level of "low risk".

[0141] 2. The XGBoost model's hyperparameters are set as follows: learning rate is set to 0.1, tree depth is set to 3, estimators are set to 100, and loss function is set to log loss.

[0142] The input static features (normalized residual wall thickness, normalized material yield strength, normalized service life, vibration-failure correlation feature, and material-life coupling feature) are input into the XGBoost model, and the output static risk probability is .

[0143] The vibration-failure correlation feature is a static feature. Static features are features that change slowly and do not fluctuate frequently with time, usually reflecting the inherent properties or long-term cumulative state of the pipeline. The average vibration fatigue damage of the pipeline in the past 30 days is a static feature. Although it is based on data from the past 30 days, its statistical period is "month", which is much lower than real-time dynamic data (such as second-level vibration acceleration), and it is a static summary of monthly vibration damage, rather than a real-time fluctuation value. It is a long-term correlation quantification of vibration fatigue accumulation and failure occurrence, and its value does not change frequently with hourly or daily vibration fluctuations, and has the stability of static features.

[0144] 3. Single static risk or dynamic risk cannot fully assess pipeline safety, for example: a pipeline with excellent static properties ( = 0.2) but intense real-time vibration ( = 0.8) may have a high actual risk; a pipeline with poor static properties ( = 0.8) but stable real-time operation ( = 0.2) may have a lower short-term risk.

[0145] Therefore, the outputs of the LSTM model and the XGBoost model need to be fused based on the following formula:

[0146]

[0147] wherein, is the fusion probability; is the LSTM model weight, is the XGBoost model weight, and ; is the LSTM model probability; is the XGBoost model probability.

[0148] The weights α, β are optimized through cross-validation, including the following steps:

[0149] Select pipeline sample historical data containing known failure and normal labels;

[0150] Divide the sample historical data into a training set and a validation set (7:3);

[0151] Test the fusion probability corresponding to different α (0.1~0.9) , calculate the accuracy rate of the fusion probability and the actual label;

[0152] Select the α, β that make the prediction accuracy the highest as the initial weights.

[0153] Calculate the accuracy rate of and the actual label based on the following method:

[0154]

[0155] wherein, is the prediction accuracy; is the actual failure and predicted failure; is the actual normal and predicted normal; is the actual normal but predicted failure; is the actual failure but predicted normal.

[0156] Cross-validation can avoid the bias of subjective weight setting, ensure that the fusion result is optimal in historical data, and provide a benchmark for subsequent dynamic adjustment.

[0157] The weights a and b are dynamically adjusted, specifically including the following steps:

[0158] When the standard deviation of the time sequence feature fluctuates more than 0.3, the influence of real-time running fluctuations (such as sudden temperature rise, pressure pulse, vibration impact) on the current risk is more significant, and the weight is increased (increased by 0.1); .

[0159] When the standard deviation of the time sequence feature does not fluctuate more than 0.3, the pipeline operation is stable, and the long-term influence of static inherent properties (such as wall thickness, material aging) is more critical, and the weight is increased (increased by 0.1). .

[0160] 4. Train and verify the model:

[0161] Collect 1000 pieces of heat pipe data (including 200 pieces of fault data), and divide them into training set, verification set and test set according to the ratio of 7:2:1;

[0162] Train the machine learning fusion model based on the training set;

[0163] Verify the machine learning fusion model based on the verification set, and when the prediction accuracy (P) of the test set is ≥0.92 and the F1 score is ≥0.9, the verification is passed. .

[0164] Test the machine learning fusion model based on the test set.

[0165] Calculate the F1 score based on the following method:

[0166]

[0167] Wherein, is the proportion of actual faults in predicted faults, . is the proportion of correctly predicted actual faults in actual faults, .

[0168] Next, go to step S150.

[0169] At step S150, the vibration fatigue damage is calculated based on the mechanical vibration data features.

[0170] In the embodiment of the application, the vibration fatigue damage is calculated based on the mechanical vibration data features, including:

[0171] The vibration fatigue damage is calculated based on the following formula:

[0172]

[0173] wherein, is the vibration fatigue damage; is the vibration load class number, the vibration load level is divided based on industry standards and vibration acceleration amplitude range, for example, less than 5 m / s 2 is classified as level one, between 5-15 m / s 2 is evaluated as level two, and greater than 15 m / s 2 is evaluated as level three; is the vibration load index; is the actual number of times of the first class vibration load; is the fatigue limit number corresponding to the first class vibration load;

[0174] The material fatigue life is calculated based on the following formula :

[0175]

[0176] wherein, is the material symmetric cycle fatigue limit, which can be obtained by consulting industry standards, for example, Q235 steel MPa; is the material fatigue index, which can be obtained by consulting industry standards; is the stress amplitude of the first vibration load.

[0177] Next, go to step S160.

[0178] At step S160, the risk probability is risk corrected based on the vibration fatigue damage, to obtain a corrected risk probability.

[0179] More specifically, when the vibration fatigue damage is less than 1, the risk correction is based on the following formula:

[0180]

[0181] wherein, is the corrected risk probability; is the risk probability; is the vibration correction coefficient, ; and

[0182] The vibration correction coefficient is calculated based on the following formula :

[0183]

[0184] wherein, is the vibration fatigue damage.

[0185] When the vibration fatigue damage is greater than the vibration correction coefficient , .

[0186] The machine learning output is not quantified separately, and under the same P, the pipeline with high vibration damage has a higher actual risk, can accurately reflect the long-term damage of vibration to the pipeline, can correct the deviation and avoid underestimating the vibration risk.

[0187] Next, go to step S170.

[0188] At step S170, the risk level is determined based on the corrected risk probability.

[0189] More specifically, the risk level is determined based on the preset risk level threshold, when it is determined to be low risk; when it is determined to be medium risk; when it is determined to be high risk; and when it is determined to be extremely high risk.

[0190] In the embodiments of the present application, the method further comprises:

[0191] When it is determined to be low risk, the vibration acceleration is detected once a month, and the data is recorded;

[0192] When it is determined to be medium risk, a rubber shock absorber is installed at the connection between the pump group and the pipeline;

[0193] When it is determined to be high risk, the vibration section pipeline is replaced with a thick-walled pipeline, and a shock absorber is installed;

[0194] When it is determined to be extremely high risk, the vibration section pipeline is shut down urgently, replaced with a fatigue-resistant material, and then operated after re-pressurization.

[0195] The new fault data is supplemented to the training set, and the machine learning model is retrained and the threshold is optimized every quarter.

[0196] The heat pipe safety risk assessment method provided by the application comprehensively reflects the real risk state of the pipe through multi-dimensional risk assessment fusion. The XGBoost model is adapted to static characteristics, accurately captures the cumulative effects of long-term aging, material attenuation, etc., the LSTM model is good at processing time series data, effectively identifies short-term dynamic risks such as temperature sudden change, pressure fluctuation, vibration anomaly, and the like, and the two types of models cooperate to improve the assessment accuracy; through feature extraction such as vibration acceleration, frequency and amplitude, the forward-looking of risk warning is improved, the intuitive risk probability is output, the risk source is quickly located by the operation and maintenance personnel, measures are taken in advance, the accident rate is reduced, and strong support is provided for the safety operation and maintenance of the heat pipe.

[0197] Figure 3 The heat pipe safety risk assessment system 300 provided by the application is shown, which comprises a data acquisition module 310, a preprocessing module 320, a feature extraction module 330, a machine learning module 340 and a risk judgment module 350.

[0198] The data acquisition module 310 is used for acquiring heat pipe data; the heat pipe data comprises static data and time series data of the heat pipe; wherein the time series data comprises mechanical vibration data;

[0199] The preprocessing module 320 is used for preprocessing the heat pipe data;

[0200] The feature extraction module 330 is used for feature extraction based on the preprocessed heat pipe data, to obtain static data features and time series data features;

[0201] The machine learning module 340 is used for inputting the static data features and the time series data features into a machine learning fusion model, and outputting a risk probability; vibration fatigue damage is calculated based on the mechanical vibration data features; the risk probability is corrected based on the vibration fatigue damage, to obtain a corrected risk probability;

[0202] The risk judgment module 350 is used for judging a risk level based on the corrected risk probability.

[0203] Referring to Figure 4 The electronic device 40 comprises:

[0204] at least one processor; and

[0205] a memory in communication connection with the at least one processor; wherein

[0206] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the heat pipe safety risk assessment method in the foregoing method embodiments.

[0207] The embodiment of the present disclosure further provides a non-transitory computer readable storage medium storing computer instructions for causing a computer to execute the heat pipe safety risk assessment method in the foregoing method embodiment.

[0208] The embodiment of the present disclosure further provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to execute the heat pipe safety risk assessment method in the foregoing method embodiment.

[0209] Reference is made below to Figure 4 which shows a structural schematic diagram of an electronic device 40 suitable for use to implement the embodiment of the present disclosure. The electronic device in the embodiment of the present disclosure can include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (Personal Digital Assistant), a PAD (Tablet Personal Computer), a PMP (Portable Multimedia Player), a vehicle terminal (e.g., a car navigation terminal), and the like, and a fixed terminal such as a digital TV, a desktop computer, and the like. Figure 4 The electronic device shown is merely an example and should not bring any limitation to the function and use range of the embodiment of the present disclosure.

[0210] As shown in Figure 4 , the electronic device 40 can include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 401 which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 40 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0211] Generally, the following devices can be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, and the like; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, and the like; a storage device 408 including, for example, a magnetic tape, a hard disk, and the like; and a communication device 409. The communication device 409 can allow the electronic device 40 to communicate with other devices wirelessly or by wire to exchange data. Although the electronic device 40 having various devices is shown in the figure, it should be understood that it is not required to implement or have all the devices shown. More or less devices can be alternatively implemented or had.

[0212] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a storage device 408, or installed from a ROM 402. When the computer program is executed by the processing device 401, it performs the functions defined in the methods of embodiments of this disclosure.

[0213] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0214] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0215] The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire at least two Internet Protocol (IP) addresses; send a node evaluation request including the at least two IP addresses to a node evaluation device, wherein the node evaluation device selects an IP address from the at least two IP addresses and returns it; and receive the IP address returned by the node evaluation device; wherein the acquired IP address indicates an edge node in a content delivery network.

[0216] Alternatively, the aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: receive a node evaluation request including at least two Internet Protocol (IP) addresses; select an IP address from the at least two IP addresses; and return the selected IP address; wherein the received IP address indicates an edge node in the content delivery network.

[0217] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0218] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0219] The units described in the embodiments of this disclosure can be implemented in software or in hardware. The name of a unit does not necessarily limit the unit itself; for example, the first acquisition unit can also be described as "a unit that acquires at least two Internet Protocol addresses".

[0220] It should be understood that the various parts of this disclosure can be implemented in hardware, software, firmware, or a combination thereof.

[0221] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for safety risk assessment of thermal pipelines, characterized in that, The method includes the following steps: Acquire thermal pipeline data; the thermal pipeline data includes static data and time-series data of the thermal pipeline; wherein, the time-series data includes mechanical vibration data; The thermal pipeline data is preprocessed; Feature extraction is performed on the preprocessed thermal pipeline data to obtain static data features and time-series data features. The static data features and time-series data features are input into a machine learning fusion model, which outputs the risk probability. Vibration fatigue damage is calculated based on the characteristics of the mechanical vibration data. Based on the vibration fatigue damage, the risk probability is corrected to obtain the corrected risk probability. The risk level is determined based on the revised risk probability. The risk probability is corrected based on the vibration fatigue damage to obtain the corrected risk probability, including: Vibration fatigue damage At that time, risk adjustment is performed based on the following formula: ； in, This is the corrected risk probability; For risk probability; This is the vibration correction factor. ; The vibration correction factor is calculated based on the following formula. : ； in, Vibration fatigue damage; Vibration fatigue damage When, vibration correction factor , .

2. The method for safety risk assessment of thermal pipelines according to claim 1, characterized in that, The preprocessing of the thermal pipeline data includes: filling missing values, removing outliers, and standardizing the thermal pipeline data; The method further includes: The authenticity of the static data is verified. Perform data alignment on time-series data; Perform signal smoothing processing on mechanical vibration data.

3. The method for safety risk assessment of thermal pipelines according to claim 1, characterized in that, The feature extraction based on the preprocessed thermal pipeline data yields static data features and time-series data features, including: Perform static data feature extraction, including: The standardized values ​​of the preprocessed static data are used as static data features. The coupling characteristics between the material and service life of thermal pipelines are calculated based on the following formula: ； in, This refers to the coupling characteristics between the material and service life of thermal pipelines; This is the standardized value of the yield strength of the pipe material; This represents the standardized value for the service life of the pipeline; 0.5 is the attenuation coefficient. Feature extraction from time-series data includes: Vibration-fault correlation characteristics are calculated based on the following formula: ； in, Vibration-fault correlation characteristics; The number of fatigue cracks in similar pipelines over the past 3 years; This represents the average vibration fatigue damage of the pipeline over the past 30 days. It is the critical point for fatigue damage.

4. The method for safety risk assessment of thermal pipelines according to claim 1, characterized in that, The process of inputting static data features and time-series data features into a machine learning fusion model and outputting risk probabilities includes: Establish LSTM and XGBoost models; The time-series data features are input into the LSTM model, which outputs the first probability value. ; The static data features are input into the XGBoost model, which outputs the second probability. ; The first probability is based on the following formula. and the second probability To merge: ； in, For fusion probability; For the weights of the LSTM model, These are the weights of the XGBoost model, and ; For LSTM model probabilities; For the XGBoost model probability.

5. The method for safety risk assessment of thermal pipelines according to claim 4, characterized in that, The method further includes: optimizing weights α and β through cross-validation; The optimization of weights α and β through cross-validation includes: Historical data of pipeline samples containing known fault and normal labels were selected; The historical data of the samples are divided into a training set and a validation set; Calculate different weights Corresponding fusion probability We select α and β, which maximize the prediction accuracy, as the initial weights.

6. The method for safety risk assessment of thermal pipelines according to claim 1, characterized in that, The calculation of vibration fatigue damage based on the features of the mechanical vibration data includes: Vibration fatigue damage is calculated based on the following formula: ； in, Vibration fatigue damage; This represents the vibration load level number; For vibration level sequence number; For the first The number of vibration cycles over 24 hours; For the first Material fatigue life corresponding to vibration level; The fatigue life of materials is calculated based on the following formula. : ； in, The symmetrical cyclic fatigue limit of the material; The material fatigue index can be obtained by consulting industry standards. For the first Stress amplitude corresponding to the level of vibration.

7. A safety risk assessment system for thermal pipelines, characterized in that, The system includes: A data acquisition module is configured to acquire thermal pipeline data; the thermal pipeline data includes static data and time-series data of the thermal pipeline; wherein, the time-series data includes mechanical vibration data; A preprocessing module is configured to preprocess the thermal pipeline data; The feature extraction module is configured to extract features based on preprocessed thermal pipeline data to obtain static data features and time-series data features. The machine learning module is configured to input static data features and time-series data features into a machine learning fusion model and output a risk probability; calculate vibration fatigue damage based on the features of the mechanical vibration data; and perform risk correction on the risk probability based on the vibration fatigue damage to obtain a corrected risk probability. The risk assessment module is configured to determine the risk level based on the revised risk probability. The risk probability is corrected based on the vibration fatigue damage to obtain the corrected risk probability, including: Vibration fatigue damage At that time, risk adjustment is performed based on the following formula: ； in, This is the corrected risk probability; For risk probability; This is the vibration correction factor. ; The vibration correction factor is calculated based on the following formula. : ； in, Vibration fatigue damage; Vibration fatigue damage When, vibration correction factor , .

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, cause the at least one processor to perform the thermal pipeline safety risk assessment method as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, The computer program product includes a computing program stored on a non-transitory computer-readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform the thermal pipeline safety risk assessment method as described in any one of claims 1 to 6.

Images