Automatic intelligent alarm analysis system and method for marine engine room based on digital twinning

Abstract

The invention discloses a marine engine room automatic intelligent alarm analysis system and method based on digital twinning, and relates to the field of marine engine room automation. A twinborn digital model of a ship machinery entity is built, a data prediction model is built through obtained operation state parameter sampling values, operation state parameter prediction values are obtained through the data prediction model, and time sequence data sequences L and L'are built through the operation state parameter sampling values and the operation state parameter prediction values respectively; a difference value data sequence # imgabs0 # is obtained, the number of wavelet decomposition layers is set to be N, wavelet transformation processing is carried out on the difference value data sequence # imgabs1 #, a Markov chain model of the operation state parameter sampling value state is constructed, a state transition matrix is obtained, an observation probability matrix of the operation state parameter sampling value is constructed, and the observation probability matrix of the operation state parameter sampling value is obtained; and defining state values when st1 = 1 and st0 of the sampling value states of the operation state parameters, constructing a judgment function by using the state values, and judging the abnormal condition of the sampling value of the operation state parameters at the t moment according to a calculation result of the judgment function.

Description

Technical Field

[0001] The present invention relates to the field of ship engine room automation, and in particular to a ship engine room automation intelligent alarm analysis system and method based on digital twins. Background Art

[0002] Ship equipment maintenance is the key to ensuring the safe and reliable operation of ships. During the operation of a ship, it is necessary to monitor its operating status based on the real-time data generated during operation to ensure the stability and safety of the operation of the ship's mechanical structures.

[0003] At present, the monitoring and early warning methods for the operation status of the mechanical mechanism of the ship mostly adopt the threshold method. Specifically, for example, by setting the thresholds of the operation status parameters, such as oil temperature, water temperature, speed, etc., and sampling each parameter in real time, when any parameter exceeds the threshold, an alarm is issued. This method is only applicable to the case where the data volume is small and the data dimension is small, and often only a few operation parameters can be monitored. With the continuous deepening of the intelligence of ships, the application of intelligent modular mechanisms has gradually increased, resulting in the data generated during the operation of ships. Compared with traditional mechanical structures, both in terms of data volume and data dimension, the data has exploded. The traditional monitoring and early warning methods for the operation status of mechanisms are no longer applicable to the needs of today's modern ships, both in terms of the timeliness of monitoring and the level of intelligence. To this end, we propose a ship engine room automation intelligent alarm analysis system and method based on digital twins. Summary of the invention

[0004] The main purpose of the present invention is to provide a ship engine room automation intelligent alarm analysis system and method based on digital twins, which can effectively solve the problems in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] The intelligent alarm analysis method for ship engine room automation based on digital twin includes:

[0007] Collect the structural parameters of the ship mechanical entity to be monitored, and according to the acquired structural parameters, use the digital twin modeling technology to interact and synchronously map the mechanical entity with the virtual digital model in real time, build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity, collect the operation status parameter data of the mechanical entity during the operation of the ship in real time, and interact and synchronously map the acquired operation status parameter sampling values ​​to the twin digital model in real time;

[0008] The data prediction model is constructed by using the acquired running state parameter sampling values, and the running state parameter prediction values ​​are obtained by using the data prediction model. The time series data sequences L and L' are constructed by using the running state parameter sampling values ​​and the running state parameter prediction values ​​respectively, wherein L = (x i(t-k) ,...,x i(t-2) ,x i(t-1) ,x it ); L'=(x' i(t-k) ,...,x' i(t-2) ,x' i(t-1) ,x' it ); x it It is represented by the sampling value of the i-th operating state parameter at the current time t; x' it It is represented as the predicted value of the i-th type of operating state parameter at the current time t;

[0009] The expression of the data prediction model is:

[0010]

[0011] Where a is a constant coefficient, and a∈(0,1);

[0012] Perform difference processing on each data point in the time series data sequence L and the time series data sequence L' to obtain the difference data sequence Set the number of wavelet decomposition layers to N pairs of difference data sequences Perform wavelet transform processing;

[0013] Define the data state set of the running state parameter sampling value as S, where S = {0, 1}, and when the running state parameter sampling value is normal data, the state value is 1, and when the running state parameter sampling value is abnormal data, the state value is 0, and set the data state of the running state parameter sampling value at time t to s t , a uv It represents the probability value of the running state parameter sampling value transferring from state u to state v. According to the definition content and the historical data of the running state parameter sampling value, the Markov chain model of the running state parameter sampling value state is constructed. Then the state transfer matrix is ​​A=(a uv ) 2×2 ; The state transition probability values ​​in the state transition matrix include a 00 、a 01 、a 10 、a 11 , the calculation formulas are: a 00 =1-a 01 ; a 10 =1-a 11 ; In the formula, The frequency of the data being transferred from an abnormal state to an abnormal state in the historical data of the running data sampling value in one step; The frequency of the data being transferred from an abnormal state to a normal state in one step in the historical data of the running data sampling value; The frequency of the data being transferred from a normal state to an abnormal state in one step in the historical data of the running data sampling value; The frequency of the data being transferred from the normal state to the normal state in one step in the historical data of the running data sampling value;

[0014] Construct the observation probability matrix P of the running state parameter sampling value = (p tk ) 1×2 , where k = 0, 1; when k = 0, it represents the probability that the sampling value of the running state parameter at time t is 0; when k = 1, it represents the probability that the sampling value of the running state parameter at time t is 1. Calculate the probability p of the sampling value of the running state parameter at time t being 1 t1 ; The calculation formula is:

[0015]

[0016] Where W(t,f) represents the wavelet transform coefficient of the difference data sequence at time t; It is expressed as the average value of the wavelet coefficients when the sampling value of the operating state parameter is normal data;

[0017] Calculate the running state parameter sampling value state s respectively t =1 and s t = 0 when the state value θ t (1) and θ t (0), where θ t (1) = a u1 ×p t1 θ t (0) = a u0 ×(1-p t1 );Use the state value to construct the judgment function f(θ t )=θ t (1)-θ t (0), and judge the abnormality of the sampling value of the running state parameter at time t according to the calculation result of the judgment function. The judgment principle is:

[0018] When f(θ t )≥0, Then t =1, indicating that the data status of the running data sampling value is normal and the running data sampling value is normal data;

[0019] When f(θ t)<0, Then t =0, indicating that the data state of the running data sampling value is abnormal state, and the running data sampling value is abnormal data.

[0020] The method further comprises:

[0021] When it is determined that the data state of the sampling value of the i-th type of operating state parameter at time t is abnormal, an alarm process is performed.

[0022] The ship engine room automation intelligent alarm analysis system based on digital twins includes a ship twin digital model construction module, a data acquisition module, a data prediction model construction module, a time series construction module, a Markov chain model construction module, an observation probability matrix construction module, a judgment function construction module, and an abnormal warning module;

[0023] The ship twin digital model construction module is used to collect the structural parameters of the ship mechanical entity to be monitored, and based on the acquired structural parameters, the mechanical entity and the virtual digital model are interacted and synchronously mapped in real time by using the digital twin modeling technology to build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity;

[0024] The data acquisition module is used to collect the operating status parameter data of the mechanical entity during the operation of the ship in real time, and map the acquired operating status parameter sampling values ​​to the twin digital model interactively and synchronously in real time;

[0025] The data prediction model building module is used to build a data prediction model based on the acquired running state parameter sampling values, and use the data prediction model to obtain the running state parameter prediction value;

[0026] The time series construction module is used to construct time series data sequences L and L' using the running state parameter sampling values ​​and the running state parameter prediction values;

[0027] The Markov chain model building module is used to build a Markov chain model of the running state parameter sampling value state according to the historical data of the running state parameter sampling value obtained, and obtain the state transfer matrix A=(a uv ) 2×2 ;

[0028] The observation probability matrix building module is used to build the observation probability matrix P of the running state parameter sampling value. tk ) 1×2 ;

[0029] The judgment function construction module is used to construct the judgment function f(θ t )=θ t (1)-θ t(0), and judging the abnormality of the sampling value of the running state parameter at time t according to the calculation result of the judgment function;

[0030] The abnormal warning module is used to perform alarm processing when it is determined that the data state of the i-th type of operating state parameter sampling value at time t is abnormal.

[0031] The system also includes a memory, a processor, and a computer program stored on the memory and executable on the processor.

[0032] The present invention has the following beneficial effects:

[0033] Compared with the prior art, a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity is built, and a data prediction model is constructed by using the acquired operation state parameter sampling values, and the operation state parameter prediction values ​​are obtained by using the data prediction model, and the operation state parameter sampling values ​​and the operation state parameter prediction values ​​are respectively used to construct the time series data sequences L and L', and the difference data sequence is obtained. Set the number of wavelet decomposition layers to N pairs of difference data sequences Perform wavelet transform processing, construct a Markov chain model of the running state parameter sampling value state, obtain the state transfer matrix, construct the observation probability matrix of the running state parameter sampling value, and define the running state parameter sampling value state s t =1 and s t =0, and use the state value to build a judgment function. According to the calculation result of the judgment function, the abnormal situation of the sampling value of the operating state parameter at time t is judged. The big data processing technology is used to identify the abnormal state parameters in the operation process of ship mechanical components in real time, which can improve the timeliness and intelligence of fault warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of the ship engine room automation intelligent alarm analysis method based on digital twin of the present invention;

[0035] Figure 2 This is a structural block diagram of the ship engine room automation intelligent alarm analysis system based on digital twins of the present invention. DETAILED DESCRIPTION

[0036] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0037] The ship engine room automation intelligent alarm analysis system based on digital twins includes a ship twin digital model construction module, a data acquisition module, a data prediction model construction module, a time series construction module, a Markov chain model construction module, an observation probability matrix construction module, a judgment function construction module, and an abnormal warning module;

[0038] The ship twin digital model construction module is used to collect the structural parameters of the ship mechanical entity to be monitored. Based on the acquired structural parameters, the digital twin modeling technology is used to interact and synchronously map the mechanical entity with the virtual digital model in real time, and to build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity.

[0039] The data acquisition module is used to collect the operating status parameter data of the mechanical entity during the operation of the ship in real time, and map the acquired operating status parameter sampling values ​​to the twin digital model in real time interactively and synchronously;

[0040] The data prediction model building module is used to build a data prediction model based on the acquired running state parameter sampling values, and use the data prediction model to obtain the running state parameter prediction value;

[0041] The time series construction module is used to construct time series data sequences L and L' using the running state parameter sampling values ​​and the running state parameter prediction values;

[0042] The Markov chain model building module is used to build a Markov chain model of the running state parameter sampling value state according to the historical data of the running state parameter sampling value obtained, and obtain the state transfer matrix;

[0043] The observation probability matrix building module is used to build the observation probability matrix of the running state parameter sampling values;

[0044] The determination function building module is used to build a determination function using the state value, and to determine the abnormality of the running state parameter sampling value at time t according to the calculation result of the determination function;

[0045] The abnormal warning module is used to perform alarm processing when it is determined that the data state of the i-th type of operating state parameter sampling value at time t is abnormal.

[0046] The specific implementation process of the technical solution of the present invention includes the following steps:

[0047] Step 1: Collect the structural parameters of the ship mechanical entity to be monitored. Based on the acquired structural parameters, use the digital twin modeling technology to conduct real-time interaction and synchronous mapping between the mechanical entity and the virtual digital model, and build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity.

[0048] Step 2: Collect the operating status parameter data of the mechanical entities during the operation of the ship in real time, and map the acquired operating status parameter sampling values ​​to the twin digital model interactively and synchronously in real time.

[0049] Step 3: Build a data prediction model based on the obtained running status parameter sampling values, and use the data prediction model to obtain the running status parameter prediction value, where the expression of the data prediction model is:

[0050]

[0051] Where a is a constant coefficient and a∈(0,1).

[0052] Step 4: Use the running state parameter sampling value and the running state parameter prediction value to construct the time series data sequence L and L', where L = (x i(t-k) ,...,x i(t-2) ,x i(t-1) ,x it ); L'=(x' i(t-k) ,...,x' i(t-2) ,x' i(t-1) ,x' it ); x it It is represented by the sampling value of the i-th operating state parameter at the current time t; x' it It is represented as the predicted value of the i-th type of operating state parameter at the current time t;

[0053] Step 5: Perform difference processing on each data point in the time series data sequence L and the time series data sequence L' to obtain the difference data sequence

[0054] Step 6: Set the number of wavelet decomposition layers to N pairs of difference data sequences Perform wavelet transform processing.

[0055] Step 7: Define the data state set of the running state parameter sampling value as S, where S = {0, 1}, and when the running state parameter sampling value is normal data, the state value is 1, and when the running state parameter sampling value is abnormal data, the state value is 0, and set the data state of the running state parameter sampling value at time t to s t , a uv represents the probability value of the running state parameter sampling value transferring from state u to state v, where the state transition probability value in the state transition matrix includes a 00 、a 01 、a 10 、a 11 , the calculation formulas are: a 00 =1-a 01 ; a 10 =1-a 11 ; In the formula, The frequency of the data being transferred from an abnormal state to an abnormal state in the historical data of the running data sampling value in one step; The frequency of the data being transferred from an abnormal state to a normal state in one step in the historical data of the running data sampling value; The frequency of the data being transferred from a normal state to an abnormal state in one step in the historical data of the running data sampling value; The frequency at which the data is transferred from the normal state to the normal state in the historical data of the operating data sampling value in one step.

[0056] Step 8: Construct a Markov chain model of the state of the running state parameter sampling value according to the definition content and the historical data of the running state parameter sampling value, and obtain the state transfer matrix A = (a uv ) 2×2 .

[0057] Step 9: Construct the observation probability matrix P of the running state parameter sampling value = (p tk ) 1×2 , where k = 0, 1; when k = 0, it represents the probability that the sampling value of the running state parameter at time t is 0; when k = 1, it represents the probability that the sampling value of the running state parameter at time t is 1. Calculate the probability p of the sampling value of the running state parameter at time t being 1 t1 ; The calculation formula is:

[0058]

[0059] Where W(t,f) represents the wavelet transform coefficient of the difference data sequence at time t; It is expressed as the average value of the wavelet coefficients when the sampling value of the operating state parameter is normal data;

[0060] Step 10: Calculate the running state parameter sampling value state s respectively t =1 and s t = 0 when the state value θ t (1) and θ t (0), where θ t (1) = a u1 ×p t1 θ t (0) = a u0 ×(1-p t1 ).

[0061] Step 11: Use the state value to construct the judgment function f(θ t )=θ t (1)-θ t(0), and judge the abnormality of the sampling value of the running state parameter at time t according to the calculation result of the judgment function. The judgment principle is:

[0062] When f(θ t )≥0, Then t =1, indicating that the data status of the running data sampling value is normal and the running data sampling value is normal data;

[0063] When f(θ t )<0, Then t =0, indicating that the data state of the running data sampling value is abnormal state, and the running data sampling value is abnormal data.

[0064] Step 12: When it is determined that the data state of the sampling value of the i-th type of operating state parameter at time t is abnormal, an alarm process is performed.

[0065] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A digital twin-based intelligent alarm analysis method for ship engine room automation, characterized in that: include: Collect the structural parameters of the ship mechanical entity to be monitored, and according to the acquired structural parameters, use the digital twin modeling technology to interact and synchronously map the mechanical entity with the virtual digital model in real time, build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity, collect the operation status parameter data of the mechanical entity during the operation of the ship in real time, and interact and synchronously map the acquired operation status parameter sampling values ​​to the twin digital model in real time; The data prediction model is constructed by using the acquired running state parameter sampling values, and the running state parameter prediction values ​​are obtained by using the data prediction model. The time series data sequences L and L' are constructed by using the running state parameter sampling values ​​and the running state parameter prediction values ​​respectively, wherein L = (x i(t-k) ,...,x i(t-2) ,x i(t-1) ,x it ); L'=(x' i(t-k) ,...,x' i(t-2) ,x' i(t-1) ,x' it ); x it It is represented by the sampling value of the i-th operating state parameter at the current time t; x' it It is represented as the predicted value of the i-th type of operating state parameter at the current time t; Perform difference processing on each data point in the time series data sequence L and the time series data sequence L' to obtain the difference data sequence Set the number of wavelet decomposition layers to N pairs of difference data sequences Perform wavelet transform processing; Define the data state set of the running state parameter sampling value as S, where S = {0, 1}, and when the running state parameter sampling value is normal data, the state value is 1, and when the running state parameter sampling value is abnormal data, the state value is 0, and set the data state of the running state parameter sampling value at time t to s t , a uv It represents the probability value of the running state parameter sampling value transferring from state u to state v. According to the definition content and the historical data of the running state parameter sampling value, the Markov chain model of the running state parameter sampling value state is constructed. Then the state transfer matrix is ​​A=(a uv ) 2×2 ; Construct the observation probability matrix P of the running state parameter sampling value = (p tk ) 1×2 , where k = 0, 1; when k = 0, it represents the probability that the sampling value of the running state parameter at time t is 0; when k = 1, it represents the probability that the sampling value of the running state parameter at time t is 1. Calculate the probability p of the sampling value of the running state parameter at time t being 1 t1 ; Calculate the running state parameter sampling value state s respectively t =1 and s t = 0 when the state value θ t (1) and θ t (0), where θ t (1) = a u1 ×p t1 θ t (0) = a u0 ×(1-p t1 );Use the state value to construct the judgment function f(θ t )=θ t (1)-θ t (0), and judge the abnormality of the sampling value of the operating state parameter at time t according to the calculation result of the judgment function.

2. The method for analyzing ship engine room automation intelligent alarm based on digital twin according to claim 1 is characterized in that: The method further comprises: When it is determined that the data state of the sampling value of the i-th type of operating state parameter at time t is abnormal, an alarm process is performed.

3. The ship engine room automation intelligent alarm analysis method based on digital twin according to claim 1 is characterized in that: The expression of the data prediction model is: Where a is a constant coefficient and a∈(0,1).

4. The method for analyzing ship engine room automation intelligent alarm based on digital twin according to claim 1 is characterized in that: The state transition probability values ​​in the state transition matrix include a 00 、a 01 、a 10 、a 11 , the calculation formulas are: a 00 =1-a 01 ; a 10 =1-a 11 ; In the formula, The frequency of the data being transferred from an abnormal state to an abnormal state in the historical data of the running data sampling value in one step; The frequency of the data being transferred from an abnormal state to a normal state in one step in the historical data of the running data sampling value; The frequency of the data being transferred from a normal state to an abnormal state in one step in the historical data of the running data sampling value; The frequency at which the data is transferred from the normal state to the normal state in the historical data of the operating data sampling value in one step.

5. The method for analyzing ship engine room automation intelligent alarm based on digital twin according to claim 1 is characterized in that: The probability p that the sampling value of the running state parameter at time t is 1 t1 The calculation formula is: Where W(t,f) represents the wavelet transform coefficient of the difference data sequence at time t; It is expressed as the average value of the wavelet coefficients when the sampling value of the operating status parameter is normal data.

6. The ship engine room automation intelligent alarm analysis method based on digital twin according to claim 1 is characterized in that: The principles for determining abnormal situations are as follows: When f(θ t )≥0, Then t =1, indicating that the data status of the running data sampling value is normal and the running data sampling value is normal data; When f(θ t )<0, Then t =0, indicating that the data state of the running data sampling value is abnormal state, and the running data sampling value is abnormal data.

7. The ship engine room automation intelligent alarm analysis system based on digital twin is characterized by: The system is used to implement the steps of the ship engine room automation intelligent alarm analysis method based on digital twins according to any one of claims 1 to 6, including a ship twin digital model construction module, a data acquisition module, a data prediction model construction module, a time series construction module, a Markov chain model construction module, an observation probability matrix construction module, a judgment function construction module, and an abnormal warning module; The ship twin digital model construction module is used to collect the structural parameters of the ship mechanical entity to be monitored, and based on the acquired structural parameters, the mechanical entity and the virtual digital model are interacted and synchronously mapped in real time by using the digital twin modeling technology to build a twin digital model of the mechanical entity for simulating the operation process of the ship mechanical entity; The data acquisition module is used to collect the operating status parameter data of the mechanical entity during the operation of the ship in real time, and map the acquired operating status parameter sampling values ​​to the twin digital model interactively and synchronously in real time; The data prediction model building module is used to build a data prediction model based on the acquired running state parameter sampling values, and use the data prediction model to obtain the running state parameter prediction value; The time series construction module is used to construct time series data sequences L and L' using the running state parameter sampling values ​​and the running state parameter prediction values; The Markov chain model building module is used to build a Markov chain model of the running state parameter sampling value state according to the historical data of the running state parameter sampling value obtained, and obtain the state transfer matrix A=(a uv ) 2×2 ; The observation probability matrix building module is used to build the observation probability matrix P of the running state parameter sampling value. tk ) 1×2 ; The judgment function construction module is used to construct the judgment function f(θ t )=θ t (1)-θ t (0), and determine the abnormality of the sampling value of the operating state parameter at time t according to the calculation result of the determination function; The abnormal warning module is used to perform alarm processing when it is determined that the data state of the i-th type of operating state parameter sampling value at time t is abnormal.

8. The digital twin-based ship engine room automation intelligent alarm analysis system according to claim 7 is characterized in that: The system also includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, the steps of the digital twin-based ship engine room automated intelligent alarm analysis method described in any one of claims 1 to 6 can be implemented.

Images