A fault simulation analysis system and method for ship power system

By designing a fault simulation analysis system in the ship's power system and using data acquisition and simulation modules for real-time analysis, the problems of long and high fault diagnosis time and high cost in traditional fault detection methods are solved, and fast and accurate fault detection and status warning are achieved. .

CN116484637BActive Publication Date: 2024-11-19GUANGDONG OCEAN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310520648.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-05-10
Publication Date
2024-11-19
Estimated Expiration
2043-05-10

AI Technical Summary

Technical Problem

Traditional fault detection methods rely on regular maintenance, resulting in increased material loss and maintenance costs. The fault diagnosis time is too long and the efficiency is inefficient, so it is impossible to monitor the operating status of the ship's power system in real time.

Method used

Design a fault simulation and analysis system, including data acquisition module, power system simulation module and analysis module, simulate and operate through the real collected ship operation data, build a model for synchronous analysis, monitor in real time and conduct status early warning.

Benefits of technology

It solves the problem of huge data quantity and slow calculation, realizes fast and accurate fault analysis and status warning of ship power systems, improves fault detection efficiency, and reduces maintenance costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116484637B_ABST
    Figure CN116484637B_ABST
Patent Text Reader

Abstract

The present invention discloses a fault simulation analysis system and method applied to a ship power system, wherein the system comprises: a data acquisition module, a power system simulation module, a database and an analysis module; the data acquisition module is used to transmit the real operation data of the ship to the power system simulation module, and to collect the simulated operation data of the power system simulation module; the power system simulation module is used to simulate the real ship power system, and to perform simulated operation based on the real operation data of the ship; the database is used to store the real operation data of the ship and the simulated operation data; the analysis module is used to perform calculations based on the simulated operation data, obtain the operation status of the power system simulation module, and perform status warnings based on the operation status. The present invention collects data from different structural units in real time, builds models, and performs synchronous analysis, thereby solving the problems of large amounts of data and slow calculations.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of fault detection, and in particular relates to a fault simulation analysis system and method applied to a ship power system. Background Art

[0002] The normal operation of a ship is inseparable from the support of the power system. The operating status of the power system not only directly affects the operation of the ship, but also relates to the safety of the personnel on board and huge economic benefits. Therefore, fault detection of the ship's power system is of great significance.

[0003] Most of the traditional fault detection still relies on regular maintenance, which is easy to cause material loss and increase maintenance costs. Frequent disassembly and inspection will accelerate the friction between components, and in severe cases, it is easy to cause new faults. Some ship operation fault diagnosis uses the principle of active electronic mutual inductor to check the modules that may fail one by one, and then uses the limiting effect of the analytical model to screen the results one by one, determine the faulty module, and complete a fault diagnosis. However, this method requires a large amount of data settlement results, which makes the fault diagnosis time too long and inefficient. Therefore, a power system simulation model is needed for fault simulation. By establishing a model, the power system operation status is monitored in real time, and the fault is integrated and judged, so as to comprehensively evaluate the performance of the ship power system. Summary of the invention

[0004] The present invention aims to solve the deficiencies of the prior art and proposes a fault simulation and analysis system and method applied to a ship power system. The system simulates the main power equipment and auxiliary equipment in the ship power system, uses the actual collected ship operation data as the historical data of each structural unit, and collects data from different structural units in real time, builds models, and performs synchronous analysis.

[0005] To achieve the above object, the present invention provides the following solution: A fault simulation and analysis system applied to a ship power system, comprising: a data acquisition module, a power system simulation module, a database and analysis module;

[0006] The data acquisition module is used to transmit the real operation data of the ship to the power system simulation module, and to acquire the simulated operation data of the power system simulation module;

[0007] The power system simulation module is used to simulate the real ship power system and perform simulated operation based on the real operation data of the ship;

[0008] The database is used to store the real operation data of the ship and the simulated operation data;

[0009] The analysis module is used to perform calculations based on the simulated operation data to obtain the operating status of the power system simulation module and to issue a status warning based on the operating status.

[0010] Preferably, the data acquisition module includes: a collection unit, a display unit, and an abnormal alarm unit;

[0011] The acquisition unit is used to acquire the actual operation data of the ship and the simulated operation data;

[0012] The display unit is used to display the simulation operation data in real time;

[0013] The abnormal alarm unit is used to issue an alarm based on the simulated operation data and a preset limit value.

[0014] Preferably, the acquisition unit includes: a rotation speed sensor, a temperature sensor, a pressure sensor, a vibration sensor, and an acoustic emission sensor.

[0015] Preferably, the power system simulation module includes: a main power unit operation unit and an auxiliary unit operation unit;

[0016] The main power unit operation unit is used to simulate the operation state of the main power unit of the ship;

[0017] The auxiliary engine operation unit is used to simulate the operation status of the ship's auxiliary engine.

[0018] Preferably, the main power unit operation unit comprises: a main propulsion unit operation subunit and a diesel engine operation subunit;

[0019] The main propulsion unit operation subunit is used to simulate the operation status of the ship shaft system and propeller;

[0020] The diesel engine operation subunit is used to simulate the operation state of the ship diesel engine set.

[0021] Preferably, the auxiliary machine operation unit includes: a generator set operation subunit and a boiler operation subunit;

[0022] The generator set operation subunit is used to simulate the operation status of the standby generator set;

[0023] The boiler operation subunit is used to simulate the operation status of a steam boiler.

[0024] Preferably, the analysis module comprises: a main analysis unit and a plurality of sub-analysis units;

[0025] A plurality of said sub-analysis units are connected to said acquisition unit, and respectively process and analyze the data acquired by said power system simulation module to obtain sub-unit analysis results;

[0026] The main analysis unit is used to comprehensively evaluate the analysis results of the sub-units, obtain the operating status of the power system simulation module, and issue a status warning based on the operating status.

[0027] The present invention also provides a fault simulation analysis method applied to a ship power system, comprising:

[0028] Collect the actual operation data of the ship as the historical operation data of the simulation system, and collect the simulated operation data of the simulation system in real time;

[0029] Storing and displaying the simulation operation data in real time;

[0030] Constructing an analysis model, and performing calculation and analysis on the simulated operation data based on the analysis model to obtain a simulated operation status;

[0031] Based on the operating conditions, a fault analysis is performed on the simulation system to obtain an analysis result.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention discloses a fault simulation and analysis system and method applied to a ship power system. The system simulates the main power equipment and auxiliary equipment in the ship power system, uses the actually collected ship operation data as the historical data of each structural unit, collects data from different structural units in real time, builds models, and performs synchronous analysis, thereby solving the problems of huge data quantity and slow calculation. The analysis of different structural units is summarized and compared with a large amount of fault analysis data stored in a database to obtain the overall equipment operation status. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0035] Figure 1 This is a schematic diagram of the structure of a fault simulation and analysis system applied to a ship power system according to the first embodiment of the present invention;

[0036] Figure 2 This is a schematic diagram of the structure of a hybrid neural network according to a first embodiment of the present invention;

[0037] Figure 3 The figure is a flow chart of a fault simulation analysis method applied to a ship power system according to the second embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] Embodiment 1

[0041] like Figure 1 As shown, it is a schematic diagram of the structure of a fault simulation and analysis system applied to a ship power system according to an embodiment of the present invention, including: a data acquisition module, a power system simulation module, a database and analysis module;

[0042] Among them, the data acquisition module is mainly used to collect the actual operation data of the ship and input it into the power system simulation module as the historical operation data of the power system simulation module; and to collect the simulated operation data of the power system simulation module during the operation process in real time; and transmit the collected simulated operation data to the analysis module for analysis and calculation.

[0043] The data acquisition module mainly includes: an acquisition unit, a display unit, and an abnormal alarm unit; in this embodiment, the acquisition unit mainly includes sensors with different functions: speed sensor, temperature sensor, pressure sensor, vibration sensor, acoustic emission sensor and other acquisition devices, and the corresponding sensor type is set according to the type of operation data required by the analysis module. The acquisition unit is connected to the display unit and the abnormal alarm unit. The acquisition unit is mainly used to collect the real operation data and simulated operation data of the above-mentioned ship, and display the simulated operation data in real time through the display unit. The abnormal alarm unit sets the limit value of each subunit as the alarm value. When the real-time collected data is higher than the alarm value or lower than the alarm value, the abnormal alarm unit will sound an alarm.

[0044] The power system simulation module is mainly used to simulate the real ship power system, and the real ship operation data collected is used as historical data for simulation operation. The ship must be equipped with a complete set of power devices and auxiliary equipment that meet the requirements of the specifications before it can sail on the water. Therefore, in this embodiment, the power system simulation module mainly includes two parts: the main power unit operation unit and the auxiliary engine operation unit. Among them, the main power unit operation unit is used to simulate the operation state of the main power unit of the ship as the main power device of the ship power system; the main power unit operation unit is provided with: the main propulsion unit operation subunit and the diesel engine operation subunit; the main propulsion unit operation subunit is used to simulate the operation state of the ship shaft system and propeller; the diesel engine operation subunit is used to simulate the operation state of the ship diesel engine unit. The auxiliary engine operation unit is used to simulate the operation state of the ship auxiliary engine. In this embodiment, the auxiliary engine operation unit mainly includes: the generator set operation subunit and the boiler operation subunit; the main power of the ship is provided by the above-mentioned diesel engine operation subunit. Based on the consideration of ship safety and reliability, large ships are often equipped with multiple sets of generators to generate electricity when needed. In this embodiment, the generator set operation subunit is provided with 3 groups, which are used to simulate the operation status of the standby generator set. Steam boilers are required on ships powered by diesel engines to provide auxiliary steam. Therefore, a boiler operation subunit is provided to simulate the operation of the steam boiler.

[0045] The database is mainly used to store the actual operation data and simulated operation data of the ship, and provides a fault database for the analysis module to call the fault database during fault analysis to accurately detect the fault.

[0046] The analysis module is used to perform calculations based on the simulated operation data to obtain the operating status of the power system simulation module, and to issue status warnings based on the operating status. The analysis module mainly includes: a main analysis unit and several sub-analysis units; wherein the several sub-analysis units are connected to the acquisition unit to process and analyze the data collected by each sub-unit of the power system simulation module to obtain the sub-unit analysis results; the main analysis unit is connected to the several sub-analysis units to summarize the analysis results of the several sub-analysis units, conduct a comprehensive evaluation, obtain the operating status of the power system simulation module, and issue status warnings based on the operating status.

[0047] According to the sub-unit setting of the power system simulation module in this embodiment, this embodiment is provided with 4 sub-analysis units: a first sub-analysis unit, a second sub-analysis unit, a third sub-analysis unit and a fourth sub-analysis unit, which are respectively used to analyze the operation data collected by the above-mentioned main propulsion unit operation sub-unit, diesel engine operation sub-unit, generator set operation sub-unit and boiler operation sub-unit.

[0048] Main propulsion failures are often caused by poor materials or casting defects in the propeller, and under long-term seawater erosion, the blades will deform or break, causing the propeller to lose balance. In severe cases, the faults caused by the propeller will also cause deviations in the center line of the shaft system, causing ship vibrations, main engine failures, and other faults. Therefore, the first sub-analysis unit mainly uses health monitoring when analyzing and calculating the main propulsion, and divides its health status into: primary fault, secondary fault, and tertiary fault. Primary fault: After the fault occurs, it seriously affects the operation of the entire machine, and it occurs frequently. Secondary fault: It has a more serious impact on the entire machine, but the number of occurrences is not too many, and there is enough time to deal with it. Third-level fault: It has almost no impact on the operation of the entire machine, and there will be no impact even if it is not handled.

[0049] The working process of the first sub-analysis unit includes:

[0050] First, perform health monitoring on the main thrusters:

[0051] (1) Statistical distribution test of equipment health characterization parameters

[0052] In this embodiment, the distribution test selects the Kermogorov test method. When more than one distribution form has been tested, the distribution likelihood ratio test method is selected to select the most suitable distribution form.

[0053] (2) Determine the health control limits of health characterization parameters under various operating conditions

[0054] When it has been determined that the equipment is in a stable operating state, the health characterization parameter control limits under the corresponding operating conditions are determined based on the health characterization parameter statistical distribution test results and operating condition information.

[0055] (3) Analysis based on health control limits and health monitoring data

[0056] In this embodiment, over-limit analysis, offset analysis and trend analysis are performed respectively. Among them, over-limit analysis means that under stable working conditions, the health monitoring data falls outside the health control limit, indicating that the health of the equipment has a large abnormal fluctuation. Offset analysis means that the health monitoring data appears on one side of the health limit multiple times, indicating that the health status of the equipment has deviated. Trend analysis refers to the trend analysis of the health monitoring data to see whether it is continuously rising or falling.

[0057] Then perform a health assessment of the main thrusters:

[0058] When no fault occurs, the health characteristic value of the device will fluctuate within a certain range. The higher the health level, the smaller the fluctuation of its health characteristic value, and vice versa. Therefore, this embodiment uses the size of the fluctuation of the health characteristic value of the device to describe the health level, and the calculation method is as follows:

[0059] A=6σ

[0060] Where σ represents the standard deviation of the health characteristic value, and A represents the health of the main thruster.

[0061] Calculate health based on health:

[0062]

[0063] In the formula, LMAX represents the upper limit of health control, LMIN represents the lower limit of health control, and C represents the health level. In this embodiment, the value range of C is between [0, 1]. The C value is divided into primary fault, secondary fault, and tertiary fault.

[0064] The second sub-analysis unit is used to calculate and analyze the operation data of the diesel engine operation sub-unit, and the specific method includes:

[0065] (1) The acquisition unit acquires vibration signals, instantaneous speed signals, temperature signals, pressure signals and other signals of the diesel engine when a fault occurs, and uniformly converts the above signals into time domain signals.

[0066] (2) Obtain the characteristic vector of the above signal, specifically:

[0067] Perform EMD decomposition on the above signal to obtain the first three order IMF components;

[0068] Based on the IMF components obtained in the first three orders, the skewness and kurtosis values ​​are calculated;

[0069] The marginal spectra of the first three-order IMF components are obtained, and the frequency band is divided at 10 kHz to calculate the normalized energy; and the eigenvector is constructed.

[0070] In this embodiment, the characteristic vector is constructed by calculating the centroid frequency value and the centroid amplitude, wherein the centroid frequency calculation method includes:

[0071]

[0072] In the formula, f i Indicates the frequency point, G(f i ) represents the energy of the frequency point.

[0073] The center of gravity amplitude calculation method includes:

[0074]

[0075] (3) Use BP neural network to identify, diagnose and analyze diesel engine faults.

[0076] In this embodiment, the number of input layer nodes is set to 27, the number of hidden layer nodes is set to 28, and the number of output layer nodes is set to 4. By selecting training samples, the BP neural network is trained and optimized. The fault is diagnosed based on the optimized BP neural network.

[0077] The third sub-analysis unit is used to perform calculation and analysis on the collected operation data of the generator set operation sub-unit.

[0078] Common faults of ship generator sets mainly include mechanical faults and electrical faults. Mechanical faults include: rotor bearing faults, overheating, vibration, noise and other problems; electrical faults include: no voltage generation during startup or no-load, voltage value is too low or too high, voltage fluctuation and other problems. Among them, rotor eccentricity fault is one of the main faults of ship generators. Inaccurate stator and rotor matching position, bearing damage, stator core movement or deformation, etc. can cause rotor eccentricity. Therefore, this embodiment takes the rotor eccentricity fault as an example, and adopts the method based on ant algorithm optimization training neural network to diagnose and analyze the fault.

[0079] The specific methods of ant algorithm optimization training BP neural network include:

[0080] (1) Set the initialization conditions: set time t = 0 and number of iterations NC = 0, set the maximum number of iterations NC max , the number of ants is K, let the set I pi The pheromone τ of element j in (1≤i≤M) j (I pi )(0)=C, and Δτ j (I pi )=0, place all ants in the ant nest.

[0081] (2) Start all ants. Each ant starts from the first set and selects an element from each set in turn according to the path selection rule.

[0082] Path selection rule: For set I pi , any ant randomly selects its j-th element according to the probability calculated by the following formula.

[0083]

[0084] Until all the ants reach the food source.

[0085] (3) When all ants have selected an element in each set and returned to the ant nest along the original path, assuming that the time taken for this process is n time units, the pheromone of the selected element is adjusted accordingly.

[0086] τ j (I pi )(t+n)=pτj (I pi )(t)+ΔT j (I pi )

[0087]

[0088] Where ρ represents the persistence of pheromone, Indicates that the kth ant is in the set I in this cycle pi The pheromone left on the j-th element of

[0089]

[0090] In the formula, Q is a constant, e k is the output error of the weights and thresholds selected by the kth ant as the weights and thresholds of the neural network, e k =|On-Op|. On and Op represent the actual output and expected output of the neural network respectively.

[0091] (4) Repeat the above steps until all ants converge to one path or the maximum number of iterations NC is reached. max , output the optimal solution, and the loop ends.

[0092] Finally, faults are detected based on the optimized neural network.

[0093] In this embodiment, the number of neurons in the input layer of the neural network is set to 6, the number of neurons in the output layer is set to 2, and the number of neurons in the hidden layer is set to 8.

[0094] The fourth sub-analysis unit is used to calculate and analyze the collected operation data of the boiler operation sub-unit. The boiler structure of the ship is complex, with many components and strong coupling between the equipment. Failure in any process may cause the boiler to fail to work properly. Therefore, it is very difficult to monitor the status and detect faults of each unit in the boiler system. This embodiment constructs a hybrid neural network to detect and analyze faults in the boiler system. In this embodiment, the hybrid neural network is constructed based on the SOM neural network and the BP neural network. First, the collected data is input into the SOM neural network, and the collected data is preliminarily classified. The preliminary classification results are used as the input of the BP neural network, and the network is trained and optimized until the hybrid neural network obtains the optimal accuracy. It is used as the final hybrid neural network model to diagnose the fault and obtain the diagnosis result.

[0095] like Figure 2As shown, the hybrid neural network includes an input layer, a competition layer, a hidden layer and an output layer. In this embodiment, the number of input layer nodes is set to 7; the number of nodes in the competition layer is intuitively important for network training, and an inappropriate number of nodes will cause a series of subsequent problems: therefore, the number of neurons in the competition layer is set to 25 after repeated tests, and the topological structure of the competition layer adopts a hexagonal topological structure; the hidden layer is designed as a single hidden layer; the number of output layer neuron nodes is set to 9. The hidden layer activation function is set to the Logsig function, and the output layer activation function is set to the Perelin function.

[0096] The main analysis unit receives the fault analysis results of the above four sub-analysis units, matches the types with the fault database in the database, obtains the overall fault situation of the ship power system, and issues an early warning based on the fault situation.

[0097] Embodiment 2

[0098] like Figure 3 As shown, the present invention also provides a fault simulation analysis method applied to a ship power system. This embodiment will describe the method flow in detail in conjunction with a simulation system. The method includes:

[0099] S1. Collecting the actual operation data of the ship as the historical operation data of the simulation system, and collecting the simulated operation data of the simulation system in real time;

[0100] In this embodiment, sensors with different functions are mainly used, such as speed sensors, temperature sensors, pressure sensors, vibration sensors, acoustic emission sensors and other acquisition devices, and corresponding sensor types are set according to the type of operating data required for fault analysis.

[0101] S2, storing and displaying the simulation operation data in real time;

[0102] The simulation operation data is displayed in real time through the display unit, and by setting the alarm value, an alarm is issued when the real-time collected data is higher than the alarm value or lower than the alarm value.

[0103] S3, construct an analysis model, perform calculation and analysis on the simulation operation data based on the analysis model, and obtain the simulation operation status;

[0104] In this embodiment, different models with fast response and high accuracy are constructed according to the conditions of different units, and the different units are analyzed simultaneously. Finally, the overall fault result is obtained by summarizing the information.

[0105] The construction method of the first analysis model includes:

[0106] First, perform health monitoring:

[0107] (1) Statistical distribution test of equipment health characterization parameters

[0108] In this embodiment, the distribution test selects the Kermogorov test method. When more than one distribution form has been tested, the distribution likelihood ratio test method is selected to select the most suitable distribution form.

[0109] (2) Determine the health control limits of health characterization parameters under various operating conditions

[0110] When it has been determined that the equipment is in a stable operating state, the health characterization parameter control limits under the corresponding operating conditions are determined based on the health characterization parameter statistical distribution test results and operating condition information.

[0111] (3) Analysis based on health control limits and health monitoring data

[0112] In this embodiment, over-limit analysis, offset analysis and trend analysis are performed respectively. Among them, over-limit analysis means that under stable working conditions, the health monitoring data falls outside the health control limit, indicating that the health of the equipment has a large abnormal fluctuation. Offset analysis means that the health monitoring data appears on one side of the health limit multiple times, indicating that the health status of the equipment has deviated. Trend analysis refers to the trend analysis of the health monitoring data to see whether it is continuously rising or falling.

[0113] Then do a health assessment:

[0114] When no fault occurs, the health characteristic value of the device will fluctuate within a certain range. The higher the health level, the smaller the fluctuation of its health characteristic value, and vice versa. Therefore, this embodiment uses the size of the fluctuation of the health characteristic value of the device to describe the health level, and the calculation method is as follows:

[0115] A=6σ

[0116] Where σ represents the standard deviation of the health characteristic value, and A represents the health of the main thruster.

[0117] Calculate health based on health:

[0118]

[0119] In the formula, LMAX represents the upper limit of health control, LMIN represents the lower limit of health control, and C represents the health level. In this embodiment, the value range of C is between [0, 1]. The C value is divided into primary fault, secondary fault, and tertiary fault.

[0120] The construction method of the second analysis model includes:

[0121] (1) The acquisition unit acquires vibration signals, instantaneous speed signals, temperature signals, pressure signals and other signals of the diesel engine when a fault occurs, and uniformly converts the above signals into time domain signals.

[0122] (2) Obtain the characteristic vector of the above signal, specifically:

[0123] Perform EMD decomposition on the above signal to obtain the first three order IMF components;

[0124] Based on the IMF components obtained in the first three orders, the skewness and kurtosis values ​​are calculated;

[0125] The marginal spectra of the first three-order IMF components are obtained, and the frequency band is divided at 10 kHz to calculate the normalized energy; and the eigenvector is constructed.

[0126] In this embodiment, the characteristic vector is constructed by calculating the centroid frequency value and the centroid amplitude, wherein the centroid frequency calculation method includes:

[0127]

[0128] In the formula, f i Indicates the frequency point, G(f i ) represents the energy of the frequency point.

[0129] The center of gravity amplitude calculation method includes:

[0130]

[0131] (3) Use BP neural network to identify, diagnose and analyze diesel engine faults.

[0132] In this embodiment, the number of input layer nodes is set to 27, the number of hidden layer nodes is set to 28, and the number of output layer nodes is set to 4. By selecting training samples, the BP neural network is trained and optimized. The fault is diagnosed based on the optimized BP neural network.

[0133] The construction method of the third analysis model includes:

[0134] The fault diagnosis and analysis is carried out by using the method of optimizing and training neural network based on ant algorithm.

[0135] The specific methods of ant algorithm optimization training BP neural network include:

[0136] (1) Set the initialization conditions: set time t = 0 and number of iterations NC = 0, set the maximum number of iterations NC max , the number of ants is K, let the set I pi The pheromone τ of element j in (1≤i≤M) j (I pi)(0)=C, and Δτ j (I pi )=0, place all ants in the ant nest.

[0137] (2) Start all ants. Each ant starts from the first set and selects an element from each set in turn according to the path selection rule.

[0138] Path selection rule: For set I pi , any ant randomly selects its j-th element according to the probability calculated by the following formula.

[0139]

[0140] Until all the ants reach the food source.

[0141] (3) When all ants have selected an element in each set and returned to the ant nest along the original path, assuming that the time taken for this process is n time units, the pheromone of the selected element is adjusted accordingly.

[0142] τ j (I pi )(t+n)=pτ j (I pi )(t)+Δτ j (I pi )

[0143]

[0144] Where ρ represents the persistence of pheromone, Indicates that the kth ant is in the set I in this cycle pi The pheromone left on the j-th element of

[0145]

[0146] In the formula, Q is a constant, e k is the output error of the weights and thresholds selected by the kth ant as the weights and thresholds of the neural network, e k =|On-Op|. On and Op represent the actual output and expected output of the neural network respectively.

[0147] (4) Repeat the above steps until all ants converge to one path or the maximum number of iterations NC is reached. max , output the optimal solution, and the loop ends.

[0148] Finally, faults are detected based on the optimized neural network.

[0149] In this embodiment, the number of neurons in the input layer of the neural network is set to 6, the number of neurons in the output layer is set to 2, and the number of neurons in the hidden layer is set to 8.

[0150] The construction method of the fourth analysis model includes:

[0151] The hybrid neural network is constructed based on the SOM neural network and the BP neural network. First, the collected data is input into the SOM neural network, and the collected data is preliminarily classified. The preliminary classification results are used as the input of the BP neural network, and the network is trained and optimized until the hybrid neural network obtains the optimal accuracy. It is used as the final hybrid neural network model to diagnose the fault and obtain the diagnosis result.

[0152] The hybrid neural network includes an input layer, a competition layer, a hidden layer and an output layer. In this embodiment, the number of input layer nodes is set to 7; the number of nodes in the competition layer is intuitively important for network training, and an inappropriate number of nodes will cause a series of subsequent problems: therefore, the number of neurons in the competition layer is set to 25 after repeated tests, and the topological structure of the competition layer adopts a hexagonal topological structure; the hidden layer is designed as a single hidden layer; the number of output layer neuron nodes is set to 9. The hidden layer activation function is set to the Logsig function, and the output layer activation function is set to the Perelin function.

[0153] S4. Perform fault analysis on the simulation system based on the operating conditions and obtain analysis results.

[0154] By receiving the fault analysis results of the above four sub-analysis units and matching the types with the fault database in the database, the overall fault situation of the ship power system is obtained, and an early warning is issued based on the fault situation.

[0155] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A fault simulation and analysis system for ship power systems, It is characterized in that include: Data acquisition module, power system simulation module, database and analysis module; The data acquisition module is used to transmit the real operation data of the ship to the power system simulation module, and to acquire the simulated operation data of the power system simulation module; The power system simulation module is used to simulate the real ship power system and perform simulated operation based on the real operation data of the ship; The power system simulation module includes: a main power unit operation unit and an auxiliary unit operation unit; The main power unit operation unit is used to simulate the operation state of the main power unit of the ship; The auxiliary engine operation unit is used to simulate the operation status of the ship's auxiliary engine; The main power unit operation unit includes: a main propulsion unit operation subunit and a diesel engine operation subunit; The main propulsion unit operation subunit is used to simulate the operation status of the ship shaft system and propeller; The diesel engine operation subunit is used to simulate the operation state of the ship diesel engine set; The auxiliary machine operation unit includes: a generator set operation subunit and a boiler operation subunit; The generator set operation subunit is used to simulate the operation status of the standby generator set; The boiler operation subunit is used to simulate the operation status of the steam boiler; The database is used to store the real operation data of the ship and the simulated operation data; The analysis module is used to perform calculations based on the simulated operation data to obtain the operation status of the power system simulation module, and to perform status warning based on the operation status; The data acquisition module includes: an acquisition unit, a display unit, and an abnormal alarm unit; The acquisition unit is used to acquire the actual operation data of the ship and the simulated operation data; The display unit is used to display the simulation operation data in real time; The abnormal alarm unit is used to alarm based on the simulated operation data and the preset limit value; The analysis module includes: a main analysis unit and a plurality of sub-analysis units; A plurality of said sub-analysis units are connected to said acquisition unit, and respectively process and analyze the data acquired by said power system simulation module to obtain sub-unit analysis results; 4 sub-analysis units: the first sub-analysis unit, the second sub-analysis unit, the third sub-analysis unit and the fourth sub-analysis unit, which are used to analyze the operation data collected by the main propulsion unit operation sub-unit, the diesel engine operation sub-unit, the generator unit operation sub-unit and the boiler operation sub-unit respectively; The working process of the first sub-analysis unit includes: First, perform health monitoring on the main thrusters: Statistical distribution test of equipment health characterization parameters: For distribution test, select the Kermogorov test method. When more than one distribution form has been tested, select the distribution likelihood ratio test method to select the most suitable distribution form; Determine the health control limits of health characterization parameters under various working conditions: When it has been determined that the equipment is in a stable working state, determine the health characterization parameter control limits under the corresponding working conditions based on the health characterization parameter statistical distribution test results and operating condition information; Analyze based on health control limits and health monitoring data: conduct over-limit analysis, deviation analysis and trend analysis respectively; Then, the health of the main thruster is evaluated: the health level is described by the fluctuation of the health characteristic value of the equipment. The calculation method is as follows: A=6σ In the formula, σ represents the standard deviation of the health characteristic value, and A represents the health degree of the main thruster; Calculate health based on health: In the formula, LMAX represents the upper limit of health control, LMIN represents the lower limit of health control, and C represents the health degree; The second sub-analysis unit is used to calculate and analyze the operation data of the diesel engine operation sub-unit, and the method includes: The acquisition unit acquires the vibration signal, instantaneous speed signal, temperature signal and pressure signal of the diesel engine when it fails, and uniformly converts them into time domain signals; Get the characteristic vector of the above signal, specifically: Perform EMD decomposition on the above signal to obtain the first three order IMF components; Based on the IMF components obtained in the first three orders, the skewness and kurtosis values ​​are calculated; Calculate the marginal spectra of the first three-order IMF components, divide the frequency band at 10 kHz, calculate the normalized energy, and construct the eigenvector; The characteristic vector is constructed by calculating the centroid frequency value and the centroid amplitude, wherein the centroid frequency calculation method includes: In the formula, f i Indicates the frequency point, G(f i ) represents the energy of the frequency point; The center of gravity amplitude calculation method includes: The BP neural network is used to identify, diagnose and analyze diesel engine faults: the number of input layer nodes is set to 27, the number of hidden layer nodes is set to 28, and the number of output layer nodes is set to 4; the BP neural network is trained and optimized by selecting training samples; the fault is diagnosed based on the optimized BP neural network; The third sub-analysis unit is used to calculate and analyze the collected operation data of the generator set operation sub-unit; The fault diagnosis and analysis is carried out by using the method of optimizing and training neural network based on ant algorithm; The specific methods of ant algorithm optimization training BP neural network include: Set initialization conditions: set time t = 0 and number of iterations NC = 0, set the maximum number of iterations NC max , the number of ants is K, let the set I pi The pheromone τ of element j in (1≤i≤M) j (I pi )(0)=C, and Δτ j (I pi )=0, place all ants in the ant nest; Start all ants, each ant starts from the first set and selects an element from each set in turn according to the path selection rule; Path selection rule: For set I pi , any ant randomly selects its jth element according to the probability calculated by the following formula; Until all the ants have reached the food source; When all ants have selected an element in each set and returned to the ant nest along the original path, the process takes n time units, and the pheromone of the selected element is adjusted accordingly: t j (I pi )(t+n)=ρτ j (I pi )(t)+Δτ j (I pi ) Where ρ represents the persistence of pheromone, represents the pheromone left by the k-th ant on the j-th element of the set Ipi in this cycle, where In the formula, Q is a constant, e k is the output error of the weights and thresholds selected by the kth ant as the weights and thresholds of the neural network, e k =|On-Op|; On and Op represent the actual output and expected output of the neural network respectively; Repeat the above steps until all ants converge to one path or the maximum number of iterations NC is reached. max , output the optimal solution, and the loop ends; Finally, the faults are detected based on the optimized neural network; The number of neurons in the input layer of the neural network is set to 6, the number of neurons in the output layer is set to 2, and the number of neurons in the hidden layer is set to 8; The fourth sub-analysis unit is used to calculate and analyze the collected operation data of the boiler operation sub-unit; by constructing a hybrid neural network, the boiler system is fault detected and analyzed; the hybrid neural network is constructed based on the SOM neural network and the BP neural network; First, the collected data is input into the SOM neural network, and the collected data is preliminarily classified. The preliminary classification results are used as the input of the BP neural network, and the network is trained and optimized until the hybrid neural network obtains the optimal accuracy. It is used as the final hybrid neural network model to diagnose the fault and obtain the diagnosis result; The hybrid neural network includes input layer, competition layer, hidden layer and output layer; the number of input layer nodes is set to 7; the number of competition layer nodes is set to 25, and the topological structure of the competition layer adopts a hexagonal topological structure; the hidden layer is designed as a single hidden layer; the number of output layer neuron nodes is set to 9; the hidden layer activation function is set to the Logsig function, and the output layer activation function is set to the Perelin function; The main analysis unit is used to comprehensively evaluate the analysis results of the sub-units, obtain the operating status of the power system simulation module, and issue a status warning based on the operating status.

2. The fault simulation and analysis system for a ship power system according to claim 1, It is characterized in that The acquisition unit includes: a rotation speed sensor, a temperature sensor, a pressure sensor, a vibration sensor, and an acoustic emission sensor.

3. A fault simulation analysis method applied to a ship power system, the fault simulation analysis method applying the fault simulation analysis system according to any one of claims 1 to 2, It is characterized in that include: Collect the actual operation data of the ship as the historical operation data of the simulation system, and collect the simulated operation data of the simulation system in real time; Storing and displaying the simulation operation data in real time; Constructing an analysis model, and performing calculation and analysis on the simulated operation data based on the analysis model to obtain a simulated operation status; Based on the operating conditions, a fault analysis is performed on the simulation system to obtain an analysis result.