CN113706037A - Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor - Google Patents
Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor Download PDFInfo
- Publication number
- CN113706037A CN113706037A CN202111018667.0A CN202111018667A CN113706037A CN 113706037 A CN113706037 A CN 113706037A CN 202111018667 A CN202111018667 A CN 202111018667A CN 113706037 A CN113706037 A CN 113706037A
- Authority
- CN
- China
- Prior art keywords
- virtual sensor
- data
- monitoring data
- trailing suction
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 94
- 238000010276 construction Methods 0.000 title claims abstract description 45
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000012549 training Methods 0.000 claims description 43
- 238000004422 calculation algorithm Methods 0.000 claims description 28
- 238000011156 evaluation Methods 0.000 claims description 25
- 238000012795 verification Methods 0.000 claims description 20
- 238000010801 machine learning Methods 0.000 claims description 12
- 238000005259 measurement Methods 0.000 claims description 11
- 238000010219 correlation analysis Methods 0.000 claims description 9
- 238000009499 grossing Methods 0.000 claims description 5
- 238000012512 characterization method Methods 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 5
- 238000001914 filtration Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000007637 random forest analysis Methods 0.000 description 5
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 4
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 4
- 238000003066 decision tree Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 239000002689 soil Substances 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000003252 repetitive effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000002285 radioactive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000002002 slurry Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F3/00—Dredgers; Soil-shifting machines
- E02F3/04—Dredgers; Soil-shifting machines mechanically-driven
- E02F3/88—Dredgers; Soil-shifting machines mechanically-driven with arrangements acting by a sucking or forcing effect, e.g. suction dredgers
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F3/00—Dredgers; Soil-shifting machines
- E02F3/04—Dredgers; Soil-shifting machines mechanically-driven
- E02F3/88—Dredgers; Soil-shifting machines mechanically-driven with arrangements acting by a sucking or forcing effect, e.g. suction dredgers
- E02F3/8833—Floating installations
- E02F3/885—Floating installations self propelled, e.g. ship
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Strategic Management (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Economics (AREA)
- Biophysics (AREA)
- General Business, Economics & Management (AREA)
- Mining & Mineral Resources (AREA)
- Structural Engineering (AREA)
- Civil Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mechanical Engineering (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Ocean & Marine Engineering (AREA)
- Educational Administration (AREA)
- Medical Informatics (AREA)
- Development Economics (AREA)
Abstract
本发明公开了一种基于虚拟传感器的耙吸挖泥船的施工安全辅助监控方法，其中，辅助监控方法包括：获取耙吸挖泥船的多维监测数据；将所述多维监测数据输入至虚拟传感器模型，输出虚拟传感器数据；其中，根据所述虚拟传感器数据用于监控耙吸挖泥船的施工安全状况。提出了采用虚拟信号辅助物理传感器对挖泥船施工过程进行安全监控的方法。
The invention discloses a construction safety auxiliary monitoring method of a trailing suction dredger based on a virtual sensor, wherein the auxiliary monitoring method comprises: acquiring multi-dimensional monitoring data of the trailing suction dredger; inputting the multi-dimensional monitoring data into the virtual sensor The model outputs virtual sensor data; wherein, the virtual sensor data is used to monitor the construction safety status of the trailing suction dredger. A method for safety monitoring of the dredger construction process using virtual signals assisted by physical sensors is proposed.
Description
Technical Field
The invention relates to the field of dredging engineering, in particular to a construction safety auxiliary monitoring method of a trailing suction hopper dredger based on a virtual sensor, which can be applied to construction safety auxiliary monitoring of trailing suction hopper dredgers for channel excavation and water dredging.
Background
The dredger is an important device for dredging engineering, and is mainly responsible for dredging a seabed riverbed or dredging a riverway, and the common types of the dredger mainly include a cutter suction dredger, a drag suction dredger and the like, which generally operate in oceans with complex soil quality and severe environment. The loading load, uncertain soil quality and random load generated by wind wave resistance which are gradually changed in each construction period all influence the mechanical performance of each system of the dredger, the dredging benefit is reduced, and the construction safety is endangered. Among them, the specific impact can be reflected in 3 aspects: 1. the vibration and the like have influence on the working performance of the machine, and the service life is influenced by abrasion, torsional fatigue and the like; 2. the reduction of mechanical properties also affects the dredging yield, thus reducing the dredging benefit; 3. the machine may compromise crew safety in the event of an accident.
Aiming at the construction safety problem of the trailing suction hopper dredger, in the related technology, a physical sensor is usually adopted to monitor the construction state, such as a densimeter and a flowmeter, so that a large amount of real-time monitoring data can be obtained. However, the uncertainty of the quality of the dredged soil, the diversity of construction conditions and the complexity of multiphase flow bring challenges to construction monitoring, the physical sensor for monitoring has risks of instability and failure, and after the physical sensor fails, the construction state of the trailing suction dredger cannot be continuously monitored, so that the safety risk exists. In addition, part of the construction parameters are too high in measurement cost or difficult to directly measure, for example, a concentration meter containing a radioactive source is often adopted in the measurement of the slurry concentration in the trailing suction pipe, so that the risks of endangering the safety of crews and polluting the environment exist, and serious potential safety hazards are brought.
Disclosure of Invention
In order to solve the technical problem, the invention provides a construction safety auxiliary monitoring method of a trailing suction hopper dredger based on a virtual sensor, which comprises the following steps:
acquiring multidimensional monitoring data of the trailing suction hopper dredger;
inputting the multidimensional monitoring data into a virtual sensor model, and outputting virtual sensor data;
and monitoring the construction safety condition of the trailing suction hopper dredger according to the virtual sensor data.
In some embodiments of the present invention, the acquiring of the multidimensional monitoring data of the trailing suction hopper dredger includes acquiring the monitoring data by physical sensors disposed at different positions on the trailing suction hopper dredger.
In some embodiments of the present invention, the outputting the virtual sensor data further includes:
and calculating a residual error between the virtual sensor data and the multidimensional monitoring data, and performing fault early warning according to a preset residual error.
In some embodiments of the present invention, the virtual sensor model is obtained by:
generating a training sample data set according to the multidimensional monitoring data of the trailing suction hopper dredger;
and training a virtual sensor model by using the training sample data set to obtain the virtual sensor model for predicting the monitoring data of the virtual sensor through related data.
In some embodiments of the present invention, the generating a training sample data set according to the multidimensional monitoring data of the trailing suction hopper dredger comprises:
preprocessing the multidimensional monitoring data, performing correlation analysis on the multidimensional monitoring data and the monitoring data of the virtual sensor, eliminating redundant features, and obtaining an optimized feature set which is used as the training sample data set.
In some embodiments of the present invention, wherein the rejecting redundant features comprises: and according to the correlation analysis result, removing redundant features by utilizing a correlation heat map.
In some embodiments of the present invention, before performing the correlation analysis, the method further includes: and smoothing the multidimensional monitoring data by adopting a filter to improve the smoothness of the data and reduce noise interference.
In some embodiments of the present invention, the training of the virtual sensor model using the training sample data set to obtain the virtual sensor model for predicting the monitoring data of the virtual sensor through the relevant data includes:
dividing the training sample data set into a training set and a verification set;
training the training set by utilizing various machine learning algorithms to obtain a virtual sensor model to be verified;
and verifying the virtual sensor model to be verified by using the verification set to obtain a characterization result for characterizing the accuracy of the virtual sensor, and selecting an optimal model as the virtual sensor model according to the verification result.
In some embodiments of the present invention, the selecting an optimal model as the virtual sensor model according to the verification result includes performing comprehensive evaluation on the verification result by using an improved comprehensive index, and selecting the optimal model as the virtual sensor model according to the comprehensive evaluation result.
In some embodiments of the present invention, the performing the comprehensive evaluation on the verification result by using the improved comprehensive index includes:
evaluating the verification result according to the performance measurement indexes of the cost type and the profit type to obtain an overall performance evaluation result;
the overall performance evaluation results were obtained by the following formula:
where ISI represents the overall performance evaluation, n represents the number of cost type indicators in the performance metric, m represents the number of profit type indicators in the performance metric, Pi、PjI, j values, P, representing performance metrics, respectivelymin，i、Pmax，iRespectively the minimum value and the maximum value P of the ith evaluation index in different prediction modelsmin，j、Pmax，jThe evaluation indexes are respectively the minimum value and the maximum value of the jth evaluation index in different prediction models.
Through the technical scheme, the invention provides the auxiliary monitoring method for the construction safety of the trailing suction hopper dredger based on the virtual sensor by utilizing the internal relation among the monitoring data, the auxiliary physical sensor carries out construction safety monitoring on the trailing suction hopper dredger to warn operators to carry out inspection on related equipment, the fault reason is found in time and maintained, and the safety accident is avoided.
Drawings
FIG. 1 schematically illustrates a flow chart of construction safety monitoring of a virtual sensor model of a trailing suction hopper dredger according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart illustrating a method for virtual sensor assisted monitoring based on a machine learning algorithm according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating feature importance analysis of a training method of a virtual sensor model according to an embodiment of the present invention;
FIG. 4 is a diagram schematically illustrating the filtering effect of the training method of the virtual sensor model according to an embodiment of the present invention;
FIG. 5 is a graph schematically illustrating a comparison of predicted effects of different algorithmic models of a virtual sensor in accordance with an embodiment of the present invention;
FIG. 6 is a schematic diagram illustrating an auxiliary application of a virtual sensor to a physical sensor according to an embodiment of the present invention;
fig. 7 schematically shows a residual value variation trend diagram between the virtual signal and the physical signal according to an embodiment of the present invention.
Detailed Description
In order that the objects, technical solutions and advantages of the present invention will become more apparent, the present invention will be further described in detail with reference to the accompanying drawings in conjunction with the following specific embodiments.
It is to be understood that such description is merely illustrative and not intended to limit the scope of the present invention. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the invention. It may be evident, however, that one or more embodiments may be practiced without these specific details. Furthermore, in the following description, descriptions of well-known technologies are omitted so as to avoid unnecessarily obscuring the concepts of the present invention.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "comprising" as used herein indicates the presence of the features, steps, operations but does not preclude the presence or addition of one or more other features.
All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.
The invention discloses a construction safety auxiliary monitoring method of a trailing suction hopper dredger based on a virtual sensor, and aims to provide a virtual sensor technology based on a machine learning algorithm by utilizing an internal relation among monitoring data, wherein the virtual sensor can assist a physical sensor to carry out safety monitoring on construction of the trailing suction hopper dredger.
FIG. 1 schematically illustrates a flow chart of construction safety monitoring of a virtual sensor model of a trailing suction hopper dredger according to an embodiment of the present invention;
in some embodiments of the present invention, the construction safety auxiliary monitoring method of the drag suction dredger based on the virtual sensor, as shown in fig. 1, includes step S110, step S120 and step S130.
In some embodiments of the invention, step S110: and acquiring multidimensional monitoring data of the trailing suction hopper dredger.
In some embodiments of the invention, step S120: and inputting the multidimensional monitoring data into a virtual sensor model, and outputting virtual sensor data.
In some embodiments of the invention, step S130: monitoring the construction safety of a trailing suction hopper dredger according to the virtual sensor data
According to the method, the internal relation among the monitoring data is utilized, the physical sensor is assisted to carry out construction safety monitoring on the trailing suction hopper dredger, the warning operator is assisted to carry out inspection on related equipment, fault reasons are found in time and maintained, and safety accidents are avoided.
Fig. 2 schematically shows a flowchart of a virtual sensor assisted monitoring method based on a machine learning algorithm according to an embodiment of the present invention.
In some embodiments of the present invention, as shown in fig. 2, the virtual sensor assisted monitoring method based on machine learning algorithm disclosed in the present invention includes the following steps S111, S121, 131, 132 and 133.
In some embodiments of the invention, step S111: the real ship acquires the operation data in the construction process through monitoring instruments arranged on all parts of the dredger.
In some embodiments of the invention, step S121: and smoothing the data by adopting a data processing method such as a filter and the like.
In some embodiments of the invention, step S131: and selecting a training sample data set from the high-dimensional monitoring data through an algorithm and a correlation thermodynamic diagram, and performing importance analysis on the features.
In some embodiments of the present invention, step S132: and (5) comparing and testing the prediction effects of different kinds of machine learning algorithms on the target value.
In some embodiments of the invention, step S133: and selecting the model with optimal performance for the prediction analysis of the virtual sensor.
FIG. 3 schematically illustrates a feature importance analysis diagram of an embodiment of the present invention.
In some embodiments of the present invention, the virtual sensor model is trained using a training sample data set to obtain a virtual sensor model for predicting monitoring data of the virtual sensor through related data, and early-stage operation data is divided into a training set and a prediction set.
In alternative embodiments of the present invention, predictive analysis is performed using machine learning algorithms of different nature or a combination of several algorithms.
In some optional embodiments of the present invention, in order to further clarify factors affecting the target, a random forest algorithm is used to analyze the feature importance.
In some alternative embodiments of the present invention, as shown in fig. 3, the main machine power is the most important in all features, since the torsional vibration is generated by the imbalance of the torsion angles of the shaft sections during the process of transmitting power to the propeller by the shaft system, so that the related power plays an important role. In some optional embodiments of the invention, the virtual signal obtained by the virtual sensor disclosed by the invention can assist the physical sensor in monitoring the construction state in real time, so that a stable and reliable virtual signal is provided for the construction parameters which are difficult to measure directly, and the limitation of the measuring means on construction monitoring is overcome.
In some optional embodiments of the invention, the virtual sensor disclosed by the invention is used for diagnosing and early warning and analyzing mechanical faults by comparing and analyzing residual values of the physical signal and the virtual signal, and provides important guarantee for construction safety control of dredging operation.
In some embodiments of the present invention, multi-dimensional monitoring data of a trailing suction hopper dredger is obtained, wherein the multi-dimensional monitoring data includes, but is not limited to, load, drag depth, flow, density, engine speed, host pump power, stern draft. Specifically, the 793 latitude trailing suction hopper construction monitoring data shown in table 1 represents 793 basic operating parameters during construction of the trailing suction hopper.
TABLE 1 trailing suction hopper dredger construction monitoring data
In some embodiments of the present invention, a training sample data set is generated from the multidimensional monitoring data of the trailing suction hopper dredger, wherein the training sample data set is divided into a training set and a validation set.
In some embodiments of the present invention, an algorithm is used to screen out features with high correlation with the target from the multidimensional monitoring data, and specifically, correlation operation is performed on the target parameters and the multidimensional monitoring data.
In some alternative embodiments of the present invention, the correlation operation uses the Lasso algorithm.
In this embodiment, Lasso is a regression model (Least absolute value regression and selection operator), and the method is a compression estimation, and a penalty function is constructed to obtain a more refined model, so that the model compresses some regression coefficients, that is, the sum of absolute values of the forcing coefficients is smaller than a certain fixed value, and some regression coefficients are set to be zero, so that the advantage of self-contraction is retained.
In some embodiments of the invention, redundant features are removed by using a correlation thermodynamic diagram, an optimal feature set is obtained, and the Lasso algorithm is adopted to screen the monitoring data of the trailing suction hopper dredger.
In some embodiments of the present invention, the virtual sensor model is trained by using the training sample data set to obtain a virtual sensor model for predicting monitoring data of the virtual sensor through related data, and the early-stage operation data is divided into a training set and a prediction set.
In some embodiments of the invention, predictive analysis is performed using machine learning algorithms of different nature or a combination of several algorithms.
In some optional embodiments of the present invention, in order to further clarify factors affecting the target, a random forest algorithm is used to analyze the feature importance.
In this embodiment, the random forest algorithm is a classifier comprising a plurality of decision trees, and the class of its output is determined by the mode of the class output by the individual trees. Leo Breiman and Adele Cutler developed algorithms that inferred random forests. This term was derived from random decision forests (random decision trees) proposed by Tin Kam Ho of Bell laboratories in 1995. This approach combines the "boosting" idea of Breimans with the "random subspace method" of Ho to build a set of decision trees.
In some embodiments of the present invention, obtaining multidimensional monitoring data for the trailing suction hopper dredger comprises obtaining construction data related to dredging operations, including navigation speed, load capacity, etc., from monitoring data from physical sensors disposed at different locations on the trailing suction hopper dredger.
In some embodiments of the invention, the generating of the training sample dataset from the multidimensional monitoring data of the trailing suction hopper dredger comprises: preprocessing the multidimensional monitoring data, performing correlation analysis on the multidimensional monitoring data and the monitoring data of the virtual sensor, eliminating redundant features, and obtaining an optimized feature set which is used as the training sample data set.
In some embodiments of the present invention, the rejecting redundant features comprises: and according to the correlation analysis result, removing redundant features by utilizing a correlation heat map.
In some embodiments of the invention, the trailing suction ship is constructed in the ocean and is influenced by multiple factors such as sea waves, gusts, mechanical vibration and the like, and the output result of the sensor is extremely unstable.
In some embodiments of the present invention, monitoring instrument device layout, ambient temperature, sensitivity, etc. is susceptible to introducing noise, and further comprises, prior to performing the correlation analysis: and smoothing the multidimensional monitoring data by adopting a filter to improve the smoothness of the data and reduce noise interference.
In some optional embodiments of the invention, a Savitzky-Golay filter is adopted to carry out smooth filtering on monitoring data, so that the smoothness of the data is improved, and noise interference is reduced.
In some alternative embodiments of the invention, a Savitzky-Golay filter (often referred to simply as an S-G filter) was originally proposed by Savitzky and Golay in 1964, published in journal of Analytical Chemistry. The method is widely applied to data stream smoothing and denoising, and is a filtering method based on local polynomial least square fitting in the time domain.
Fig. 4 schematically shows a filtering effect diagram of a training method of a virtual sensor model according to an embodiment of the present invention.
In some embodiments of the invention, by adopting the technical scheme, the noise can be filtered, the shape and the width of the signal can be ensured to be unchanged, the output result of the sensor of the trailing suction hopper dredger can be better stabilized, and the filtering effect is shown in fig. 4.
In some embodiments of the present invention, the training of the virtual sensor model using the training sample data set to obtain the virtual sensor model for predicting the monitoring data of the virtual sensor through the related data includes steps S201 and S202.
In some embodiments of the invention, step S201: and dividing the training sample data set into a training set and a verification set.
In some embodiments of the invention, step S202: and training the training set by using various machine learning algorithms to obtain a virtual sensor model to be verified.
In some optional embodiments of the present invention, 4 machine learning algorithms with different properties are used to perform predictive analysis on data, specifically including an LSTM algorithm, an XGBoost algorithm, an SVR forward algorithm, and a DBN algorithm.
In this embodiment, the LSTM, also called Long Short-Term Memory network (Long Short-Term Memory), is a time-cycle neural network, and is designed to solve the Long-Term dependence problem of general RNNs (cyclic neural networks), all RNNs have a chain form of a repetitive neural network module, and in a standard RNN, the repetitive structural module has only a very simple structure, such as a tanh layer.
In the embodiment, XGBoost is also called eXtreme Gradient Boosting (eXtreme Gradient Boosting), and is often used in some competitions, and the effect is significant. The XGboost is an improved GBDT (generalized boosting decision tree) algorithm and can be used for classification and regression.
In this embodiment, the SVR is also called Support Vector Regression (Support Vector Regression), and the SVR is an important application branch in an SVM (Support Vector machine), and mainly implements linear Regression by constructing a linear decision function in a high-dimensional space after dimension increasing, and when an e-insensitive function is used, the basis of the SVR is mainly an e-insensitive function and a kernel function algorithm.
In this embodiment, the DBN is also called Deep Belief networks (Deep Belief networks), which is a network of probabilistic graphical knowledge expression and inference model based on multi-layer neurons.
In some embodiments of the present invention, model evaluation indexes such as goodness of fit R2, RMSE, MAPE, and the like are selected to evaluate and analyze the predicted effect of the model.
In this embodiment, R2Is a statistical determinant coefficient for measuring Goodness of Fit (Goodness of Fit), which is the degree of Fit of a regression line to an observed value. The statistic for measuring goodness of fit is the coefficient of likelihood (also known as the deterministic coefficient) R2，R2Maximum value of 1, R2The closer the value of (1) is, the better the fitting degree of the regression straight line to the observed value is; otherwise, R2The smaller the value of (a) is, the worse the fitting degree of the regression line to the observed value is.
In the present embodiment, RMSE is also called Root Mean Square Error (Root Mean Square Error), also called Root-Mean-Square deviation (RMSD), which is a commonly used measure of the difference between measured values.
In this embodiment, MAPE is also called Mean Absolute Percentage Error (Mean Absolute Percentage Error) and is commonly used as a statistical indicator for measuring the prediction accuracy.
Fig. 5 schematically shows a model prediction effect diagram of a training method of a virtual sensor model according to an embodiment of the present invention.
In some embodiments of the present invention, as shown in fig. 5, the DBN model with the best overall effect (the sum of coefficients is the lowest) is selected as the prediction model of the virtual sensor.
In some embodiments of the invention, the virtual sensor model to be verified is verified by using a verification set to obtain a characterization result for characterizing the accuracy of the virtual sensor, and an optimal model is selected as the virtual sensor model according to the verification result.
In some embodiments of the present invention, the selecting an optimal model as the virtual sensor model according to the verification result includes performing comprehensive evaluation on the verification result by using an improved comprehensive index.
And selecting an optimal model as a virtual sensor model according to the comprehensive evaluation result, generally selecting indexes such as R2, RMSE, MAPE and the like to evaluate the model prediction effect, selecting an Improved Synthesis Index (ISI) index to carry out comprehensive evaluation, and selecting the optimal model to be used for the virtual sensor to obtain a virtual signal.
In some embodiments of the present invention, the comprehensively evaluating the verification result by using the improved comprehensive index includes: and evaluating the verification result according to the performance measurement indexes of the cost type and the profit type to obtain an overall performance evaluation result.
In some embodiments of the invention, the overall performance evaluation result is obtained by:
ISI represents an overall performance evaluation result, n represents the number of cost type indexes in performance measurement, m represents the number of profit type indexes in the performance measurement, Pi and Pj represent the ith and jth values of the performance measurement respectively, Pmin, i and Pmax are the minimum value and the maximum value in different prediction models of the ith evaluation index respectively, and Pmin, j and Pmax and j are the minimum value and the maximum value in different prediction models of the jth evaluation index respectively.
In this embodiment, the cost type indicators include mean absolute error MAE, mean absolute percentage error MAPE, mean square error MSE, root or root mean square error RMSE, log of mean square error MSLE, and median absolute error MedAE.
In the present embodiment, the performance type indicator includes a goodness-of-fit R2.
The invention also discloses an application method of the drag suction dredger virtual sensor model, which comprises the step S301 and the step S302.
In some embodiments of the invention, step S301: and acquiring multidimensional monitoring data of the trailing suction hopper dredger.
In some embodiments of the invention, step S302: and inputting the multidimensional monitoring data into the virtual sensor model, and outputting the virtual sensor data.
In the embodiment, the virtual sensor model is obtained by training a training method of the virtual sensor model of the trailing suction hopper dredger, and the virtual sensor is adopted for auxiliary monitoring aiming at physical quantities which are difficult to directly measure.
In this embodiment, when the physical signal is unstable, the parameters obtained by the physical sensor are relatively discrete and cannot be used as a reference, and at this time, continuous and stable monitoring of the construction process can be achieved by referring to the virtual signal.
In some optional embodiments of the invention, under the influence of an unstable construction environment, the physical sensor has a certain measurement deviation, even has risks of instrument failure and monitoring data all being 0, and when the physical sensor cannot acquire construction parameters, the virtual sensor is adopted to perform safety monitoring on construction.
Fig. 6 is a schematic diagram illustrating an auxiliary application of a virtual sensor to a physical sensor in a training method of a virtual sensor model according to an embodiment of the present invention.
In some optional embodiments of the present invention, as shown in fig. 6, the torsional vibration values between serial numbers 1264 and 1791 are all 0 due to instrument failure, which affects the control of the dredger operation condition during the period.
At the moment, other physical sensors are still in a normal working state and can obtain stable detection parameters, so that the virtual sensor can still normally represent parameters of torsional vibration, specifically, target parameters can be predicted in real time through the internal relation among monitoring data, stable virtual signals are obtained, and accurate monitoring of the construction state is continuously achieved.
Fig. 7 schematically shows a residual value variation trend diagram between the virtual signal and the physical signal according to an embodiment of the present invention.
In some embodiments of the present invention, fig. 7 illustrates a trend of variation of residual value for torsional vibration as a parameter.
In this embodiment, the residual value is small in the initial stage, and as the construction progresses, the machine has a slight torsional vibration phenomenon, and the residual value gradually increases.
Based on this, it is alert operating personnel need carry out the inspection of relevant equipment, in time discovers the fault reason and maintains, avoids the emergence of incident.
The invention combines the working principle of the trailing suction dredger in the dredging engineering and the machine learning algorithm, forms a virtual sensor by utilizing the internal relation among the monitoring data, and provides a method for carrying out safety monitoring on the construction process of the dredger by adopting a virtual signal auxiliary physical sensor.
On one hand, the virtual sensor can assist the physical sensor to monitor a target which is difficult to directly measure, so that the stability and reliability of signals in a monitoring system are kept, and the damage of a poor detection means to the environment is reduced; and secondly, the residual error of the physical sensor and the virtual sensor is utilized to monitor the mechanical fault, the fault is found in time, and the diagnosis and early warning are carried out on the mechanical operation state, so that the dredging construction safety control precision is improved.
So far, the embodiments of the present disclosure have been described in detail with reference to the accompanying drawings. It is to be noted that, in the attached drawings or in the description, the implementation modes not shown or described are all the modes known by the ordinary skilled person in the field of technology, and are not described in detail. In addition, the above definitions of the components are not limited to the specific structures, shapes or manners mentioned in the embodiments, and those skilled in the art may easily modify or replace them.
It is also noted that, unless otherwise indicated, the numerical parameters set forth in the specification and attached claims are approximations that can vary depending upon the desired properties sought to be obtained by the present disclosure. In particular, all numbers expressing dimensions, range conditions, and so forth, used in the specification and claims are to be understood as being modified in all instances by the term "about". Generally, the expression is meant to encompass variations of ± 10% in some embodiments, 5% in some embodiments, 1% in some embodiments, 0.5% in some embodiments by the specified amount.
It will be appreciated by a person skilled in the art that various combinations and/or combinations of features described in the various embodiments and/or in the claims of the invention are possible, even if such combinations or combinations are not explicitly described in the invention. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present invention may be made without departing from the spirit or teaching of the invention. All such combinations and/or associations fall within the scope of the present invention.
The above embodiments are provided to further explain the objects, technical solutions and advantages of the present invention in detail, it should be understood that the above embodiments are only examples of the present invention and are not intended to limit the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111018667.0A CN113706037A (en) | 2021-08-31 | 2021-08-31 | Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111018667.0A CN113706037A (en) | 2021-08-31 | 2021-08-31 | Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113706037A true CN113706037A (en) | 2021-11-26 |
Family
ID=78658583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111018667.0A Pending CN113706037A (en) | 2021-08-31 | 2021-08-31 | Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113706037A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102411657A (en) * | 2011-10-31 | 2012-04-11 | 江苏科技大学 | Design method of filter for dynamic positioning of trailing suction hopper dredger in free navigation state |
| CN110045608A (en) * | 2019-04-02 | 2019-07-23 | 太原理工大学 | Based on the twin mechanical equipment component structural dynamic state of parameters optimization method of number |
| CN110782078A (en) * | 2019-10-18 | 2020-02-11 | 武汉科技大学 | Learning method for predicting mud discharging rate of trailing suction hopper dredger |
| CN110795891A (en) * | 2019-10-18 | 2020-02-14 | 武汉科技大学 | A learning method for operating parameters of trailing suction dredger |
| CN110826790A (en) * | 2019-10-31 | 2020-02-21 | 天津大学 | An intelligent prediction method for construction productivity of cutter suction dredger |
| AU2020101453A4 (en) * | 2020-07-23 | 2020-08-27 | China Communications Construction Co., Ltd. | An Intelligent Optimization Method of Durable Concrete Mix Proportion Based on Data mining |
| AU2020101854A4 (en) * | 2020-08-17 | 2020-09-24 | China Communications Construction Co., Ltd. | A method for predicting concrete durability based on data mining and artificial intelligence algorithm |
-
2021
- 2021-08-31 CN CN202111018667.0A patent/CN113706037A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102411657A (en) * | 2011-10-31 | 2012-04-11 | 江苏科技大学 | Design method of filter for dynamic positioning of trailing suction hopper dredger in free navigation state |
| CN110045608A (en) * | 2019-04-02 | 2019-07-23 | 太原理工大学 | Based on the twin mechanical equipment component structural dynamic state of parameters optimization method of number |
| CN110782078A (en) * | 2019-10-18 | 2020-02-11 | 武汉科技大学 | Learning method for predicting mud discharging rate of trailing suction hopper dredger |
| CN110795891A (en) * | 2019-10-18 | 2020-02-14 | 武汉科技大学 | A learning method for operating parameters of trailing suction dredger |
| CN110826790A (en) * | 2019-10-31 | 2020-02-21 | 天津大学 | An intelligent prediction method for construction productivity of cutter suction dredger |
| AU2020101453A4 (en) * | 2020-07-23 | 2020-08-27 | China Communications Construction Co., Ltd. | An Intelligent Optimization Method of Durable Concrete Mix Proportion Based on Data mining |
| AU2020101854A4 (en) * | 2020-08-17 | 2020-09-24 | China Communications Construction Co., Ltd. | A method for predicting concrete durability based on data mining and artificial intelligence algorithm |
Non-Patent Citations (2)
| Title |
|---|
| MENGXI ZHANG等: "Multiple mechanical properties prediction of hydraulic concrete in the form of combined damming by experimental data mining", 《CONSTRUCTION AND BUILDING MATERIALS》, vol. 1, pages 663 - 155 * |
| 黄荣舟: "基于深度循环网络的风电齿轮箱故障诊断方法研究", 《中国优秀硕士学位论文全文数据库（工程科技Ⅱ辑）》, no. 01, pages 11 - 22 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Digital twin-driven virtual sensor approach for safe construction operations of trailing suction hopper dredger | |
| JP7514242B2 (en) | Method and system for predicting risk of observable damage in wind turbine gearbox components | |
| CN105700518B (en) | A kind of industrial process method for diagnosing faults | |
| CN110807257A (en) | Method for predicting residual life of aircraft engine | |
| JP5292477B2 (en) | Diagnostic device and diagnostic method | |
| CN112116566B (en) | A Defect Diagnosis Method for Onshore Oil and Gas Pipelines Based on Hyperspectral Remote Sensing Technology | |
| CN106368816A (en) | Method for online abnormity detection of low-speed diesel engine of ship based on baseline deviation | |
| Xiao et al. | A model-based health indicator for leak detection in gas pipeline systems | |
| CN114738205B (en) | Method, device, equipment and medium for monitoring state of floating fan foundation | |
| CN109270458A (en) | Intelligent failure diagnosis method, system, Wind turbines and storage medium | |
| Wang et al. | The replacement of dysfunctional sensors based on the digital twin method during the cutter suction dredger construction process | |
| CN110990938B (en) | Soft measurement method for tunnel face in rock digging state | |
| JP2018034585A (en) | Analysis of propulsion performance of ship | |
| Li et al. | Anomaly identification of wind turbine blades based on Mel-spectrogram difference feature of aerodynamic noise | |
| CN118363308B (en) | Pipe jacking hoisting control system and control method based on hydrologic digital twinning | |
| CN119204378A (en) | Intelligent optimization system and method for unmanned boat navigation path based on ocean data | |
| CN113706037A (en) | Construction safety auxiliary monitoring method of trailing suction hopper dredger based on virtual sensor | |
| KR102108975B1 (en) | Apparatus and method for condition based maintenance support of naval ship equipment | |
| Liang et al. | Unsupervised anomaly detection in marine diesel engines using transformer neural networks and residual analysis | |
| Sanchez-Varela et al. | Determining the likelihood of incidents caused by human error during dynamic positioning drilling operations | |
| Wang et al. | Natural Gas Transmission Pipeline Leak Detection Model Based on Acoustic Emission and Machine Learning | |
| CN113033632A (en) | High-altitude platform fault diagnosis method based on wavelet analysis and multi-layer overrun learning machine | |
| Smith et al. | Now You SCC Me, Now You Don’t: Using Machine Learning to Find Stress Corrosion Cracking | |
| CN118129081B (en) | Active safety early warning system based on neural network | |
| CN119179912B (en) | Self-elevating platform sensor health index evaluation method, device, server and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211126 |
|
| RJ01 | Rejection of invention patent application after publication |