Hydrogen cooling system intelligent inspection planning method and system based on harmony search algorithm

Abstract

The invention discloses a hydrogen cooling system intelligent inspection planning method and system based on a harmony search algorithm, and belongs to the technical field of electric power intelligent inspection. The problem that in the prior art, a traditional hydrogen cooling system intelligent inspection method lacks self-adaptive dynamic planning capacity is solved. The operation state data of each device in the hydrogen cooling system is acquired in real time through the data acquisition module; performing feature extraction through a state evaluation module to obtain processed data, establishing a multi-dimensional state evaluation function, and performing multi-target optimization calculation on the inspection path through a harmony search optimization module by adopting a harmony search algorithm to obtain an optimization result; an optimal inspection route is generated through a route planning module; the inspection equipment is driven by the control execution module; and through the man-machine interaction module, alarming is carried out on the calibration result, and alarming information and path planning are displayed. The method effectively improves the efficiency of intelligent routing inspection, and can be applied to routing inspection path planning of the hydrogen cooling system.

Description

TECHNICAL FIELD

[0001] The present application relates to a hydrogen cooling system intelligent inspection planning method and system, in particular to a hydrogen cooling system intelligent inspection planning method and system based on a harmony search algorithm, and belongs to the technical field of electric power intelligent inspection. BACKGROUND

[0002] The current inspection and maintenance of the hydrogen cooling system in the power system mainly relies on manual experience to formulate plans or fixed cycle inspection mode. The traditional manual planning method has problems of high redundancy of inspection path and low positioning efficiency of abnormal equipment, and is difficult to adapt to the dynamic changes of equipment state under complex working conditions. The existing automatic inspection system mainly uses genetic algorithm, particle swarm algorithm and other conventional optimization methods. When solving the inspection path planning under multiple constraint conditions, it is easy to fall into local optimal solution and the calculation convergence speed is insufficient, which leads to the difficulty in meeting the real-time requirements of industrial field. Therefore, the existing technology has the following defects: first, the fixed cycle inspection mode does not consider the real-time running state of the equipment, which may cause invalid inspection or missed inspection risk; second, the conventional optimization algorithm is difficult to realize global optimal path search when dealing with complex constraint conditions such as hydrogen concentration, temperature gradient and vibration frequency; third, the existing system lacks the ability to predict the degradation trend of the equipment, and cannot dynamically adjust the inspection priority. In addition, the traditional method has not established an effective multi-objective optimization model in the aspect of cooperative optimization of equipment state data and inspection resource allocation, which leads to the difficulty in balancing the inspection efficiency and cost control.

[0003] At present, there is no case of applying the harmony search algorithm to the intelligent inspection planning of the hydrogen cooling system. The existing harmony search algorithm is mainly limited to single parameter optimization in the application of power equipment maintenance, and has not realized the deep integration of multi-dimensional information such as equipment state monitoring data, environmental variables and operating load. Therefore, the existing inspection system lacks adaptive dynamic planning ability when facing the nonlinear equipment degradation process, and it is difficult to meet the intelligent demand of equipment state maintenance in modern power system.

[0004] In view of the above, a hydrogen cooling system intelligent inspection planning method and system based on a harmony search algorithm are needed. SUMMARY

[0005] In the following, a brief summary of the present application is given in order to provide a basic understanding of some aspects of the present application. It should be understood that this summary is not an exhaustive overview of the present application. It is not intended to identify key or important parts of the present application nor is it intended to limit the scope of the present application. Its purpose is merely to present some concepts in a simplified form as a prelude to the more detailed description to be discussed later.

[0006] In view of this, in order to solve the problem that the intelligent inspection method of the traditional hydrogen cooling system lacks adaptive dynamic planning capability in the prior art, the present application provides a hydrogen cooling system intelligent inspection planning method and system based on a harmony search algorithm.

[0007] The technical scheme one is as follows: a hydrogen cooling system intelligent inspection planning method based on a harmony search algorithm, comprising the following steps:

[0008] S1. Real-time acquisition of the running state data of each device in the hydrogen cooling system through a data acquisition module;

[0009] S2. Feature extraction of the running state data through a state evaluation module to obtain processed data, and establishment of a multi-dimensional state evaluation function;

[0010] S3. Multi-objective optimization calculation of the inspection path by the harmony search optimization module according to the processed data, using the harmony search algorithm, to obtain an optimization result;

[0011] S4. Generation of an optimal inspection route according to the optimization result by a path planning module;

[0012] S5. Driving of the inspection device by a control execution module according to the optimal inspection route to perform path planning, and output of a verification result;

[0013] S6. Storage of all current data by a man-machine interaction module, alarm of the calibration result according to the multi-dimensional state evaluation function, and display of the alarm information and the path planning.

[0014] Further, in S1, the following steps are included:

[0015] S11. Arrangement of a temperature sensor array on the surface of a heat exchanger of the hydrogen cooling system;

[0016] S12. Installation of a pressure transmitter array at key nodes of a hydrogen circulation pipeline;

[0017] S13. Setting of a vibration detection unit on a compressor and a pump body base;

[0018] S14. Configuration of a hydrogen purity sensor at the outlet end of a hydrogen recovery device;

[0019] S15. Preprocessing of the multi-source heterogeneous data obtained in steps S11-S14 by a data preprocessing unit to obtain processed multi-source data;

[0020] In S15, the preprocessing is performed by a FIR digital filter built in the data preprocessing unit according to a filter function, and the cutoff frequency is set to Hz;

[0021] The filter function is represented as:

[0022]

[0023] in, The coefficients are for the 128th order Hamming window. This is the original sampling sequence. This is the filtered output sequence;

[0024] S16. The processed data is fused using the industrial Ethernet protocol via a wireless communication unit to obtain the fused data;

[0025] In step S16, the wireless communication unit adopts LoRa modulation and has a transmission distance of ≥500m.

[0026] Furthermore, in step S2, the feature extraction unit uses wavelet packet transform to extract features from the fused data, thus decomposing the number of layers. The db4 wavelet basis function was selected to obtain the data features;

[0027] Establish a multidimensional state evaluation function based on a health scoring model. ;

[0028] Multidimensional state evaluation function Represented as:

[0029]

[0030] in, For the first The weighting coefficients of each parameter. For real-time measurements, and These are the upper and lower limits of the parameter's allowable range, respectively.

[0031] Multidimensional state evaluation function The weighting coefficients are dynamically updated using the entropy method.

[0032] The dynamic update of entropy is represented as:

[0033]

[0034] in, Let j be the information entropy value of the j-th index. For the first The indicator in the first The proportion in each sample The total number of samples.

[0035] Furthermore, step S3 includes the following steps:

[0036] S31. Construct a device status feature vector based on data characteristics. A comprehensive objective function is established through the fitness evaluation unit;

[0037] In step S31, the comprehensive objective function for evaluating the optimization objective of the inspection task is defined as follows:

[0038]

[0039] in, Total inspection time. Score the status of each node. For path redundancy, the weighting coefficient , , ;

[0040] Define local objective function ;

[0041] Local objective function Represented as:

[0042]

[0043] in, Indicates the inspection time. To reduce inspection costs, It's the safety factor. These are dynamically adjusted weighting parameters;

[0044] Safety factor in the objective function Represented as:

[0045]

[0046] in, , Indicates the first The characteristic values ​​of each monitoring point For the corresponding threshold, This is the sensitivity adjustment coefficient;

[0047] The collaborative optimization model is defined as follows:

[0048]

[0049] in, The function represents the length of the inspection path. The function representing the equipment risk index. Represents the energy consumption assessment function; For dynamic weighting coefficients, satisfying ,and ;

[0050] S32. Generate the initial harmony memory library ,in, The number of solution vectors, This represents the number of device nodes. The j-th component of the i-th solution vector in the harmony memory HM is stored when the harmony memory is initialized. N=50 sets of feasible solution vectors are stored, and the solution vector dimension d=8 corresponds to the inspection point parameters.

[0051] S33. An improved harmony search algorithm is executed through an iterative optimization unit;

[0052] In step S33, the update formula for the improved harmony search algorithm is:

[0053]

[0054] in, For pitch adjustment bandwidth, The random numbers are uniformly distributed in the range [-1, 1]. The neighborhood search step size, The existing solution vectors are randomly drawn from the harmony memory. The solution vector is generated completely randomly. The solution vector is adjusted based on the neighborhood.

[0055] S34. Application of probability adjustment mechanism ,in, For the number of iterations, To minimize the probability of memory adoption, The initial maximum memory adoption probability, The attenuation coefficient;

[0056] In S34, the optimal path cost function is defined as:

[0057]

[0058] in, Let be the length of the i-th path segment. m / s is the speed of movement. s represents the time consumed by a single point detection. This is the penalty coefficient for state degradation.

[0059] Furthermore, step S4 includes the following steps:

[0060] S41. The evaluation function is generated using the Dijkstra algorithm through the Dijkstra algorithm unit;

[0061] In S41, the evaluation function Represented as:

[0062]

[0063] in, Representing a path For the The detection probability of defects in similar equipment. Assign weights to equipment importance;

[0064] S42. Construct an adjacency matrix of the device connection diagram by dynamically adjusting the units. ,in, , The node spacing, This represents the path risk value. These are risk impact factors.

[0065] Furthermore, step S5 includes the following steps:

[0066] S51. The optimized path is converted into motion control commands through the command issuing unit, and the inspection speed is dynamically adjusted according to real-time sensor data;

[0067] S52. Feedback verification unit, using the CRC-16 verification algorithm, generates the polynomial as follows: Verification of the polynomial This represents the polynomial variable corresponding to the shift register.

[0068] Furthermore, step S6 includes the following steps:

[0069] S61. Through the log management unit, record the equipment 3D model library, store the spatial coordinates and connection topology records of each device, the mapping relationship between equipment fault modes and detection locations, and save the optimal algorithm parameters under different working conditions to build an algorithm parameter knowledge base;

[0070] S62. Through the alarm unit, based on the multidimensional state evaluation function of the health scoring model... An alarm is triggered by comparing the value with a preset value.

[0071] In S62, the alarm unit 720 is equipped with a three-level early warning mechanism: when A yellow alert is triggered at any time. An orange alert is triggered at any time. A red alert is triggered when the log management unit 730 adopts a ring buffer structure, with a storage capacity of 8GB and a data retention period of ≥30 days.

[0072] S63. Display alarm information and optimal inspection route via touch screen.

[0073] Technical Solution 2: A smart inspection and planning system for a hydrogen cooling system based on a harmony search algorithm, used to execute the smart inspection and planning method for a hydrogen cooling system based on a harmony search algorithm described in Technical Solution 1, comprising a system architecture module, a data acquisition module, a status evaluation module, a harmony search optimization module, a path planning module, a control execution module, and a human-computer interaction module connected in sequence;

[0074] The system architecture module is connected to the data acquisition module, status evaluation module, harmony search optimization module, path planning module, control execution module, and human-computer interaction module, respectively.

[0075] Furthermore, the system architecture module includes a main controller, a data bus, and a power management unit;

[0076] The data acquisition module includes a sensor group, a data preprocessing unit, and a wireless communication unit connected in sequence. The sensor group includes a hydrogen purity sensor, a temperature sensor array, a vibration sensor, and a pressure transmitter array.

[0077] The status assessment module includes a feature extraction unit and a health scoring model connected in sequence;

[0078] The harmony search and optimization module includes a harmony memory library, an iterative optimization unit, and a fitness evaluation unit connected in sequence.

[0079] The path planning module includes a Dijkstra algorithm unit and a dynamic adjustment unit connected in sequence.

[0080] The control execution module includes an instruction issuing unit and a feedback verification unit connected in sequence.

[0081] The human-computer interaction module includes a touch screen, an alarm unit, and a log management unit connected in sequence.

[0082] The data bus is connected to the wireless communication unit, the wireless communication unit is connected to the feature extraction unit, the health scoring model is connected to the iterative optimization unit, the fitness evaluation unit is connected to the Dijkstra algorithm unit, the dynamic adjustment unit is connected to the instruction issuing unit, the feedback verification unit is connected to the log management unit, and the sensor group, feature extraction unit, harmony memory library, Dijkstra algorithm unit, instruction issuing unit, and touch screen are respectively connected to the power management unit.

[0083] The beneficial effects of this invention are as follows: This invention achieves dynamic optimization of inspection paths and detection strategies by constructing a mapping model between the state space of hydrogen cooling equipment and the solution space of the algorithm; This invention includes sensor network configuration, algorithm parameter setting and iterative optimization process, and has the characteristics of high detection accuracy and fast response speed, and is suitable for the full life cycle condition maintenance of hydrogen cooling equipment of generator sets;

[0084] The technical field of this invention covers hydrogen cooling system operation status monitoring technology, generator set cooling equipment maintenance strategy optimization method, and the application of artificial intelligence algorithms in industrial equipment inspection path planning. It focuses on solving the problems of low efficiency and unscientific path planning in traditional hydrogen cooling systems, and realizes intelligent decision support for equipment condition maintenance through the harmony search algorithm.

[0085] To address the shortcomings of existing hydrogen cooling systems, such as low efficiency and unscientific path planning during manual inspections, this system employs a three-tiered architecture comprising a data acquisition module, a path optimization module, and an execution control module. The data acquisition module acquires real-time temperature field distribution data and equipment operating parameters from the cooling system. The path optimization module utilizes an improved acoustic search algorithm to establish a multi-objective optimization model constrained by equipment failure probability, detection time, and path cost. The execution control module generates inspection routes based on the optimization results and dynamically adjusts the detection frequency. Through an acoustic memory update mechanism and a probabilistic perturbation strategy, this system achieves adaptive optimization of the inspection path, significantly improving hydrogen leak detection efficiency while reducing the risk of equipment downtime. Attached Figure Description

[0086] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0087] Figure 1 This is a flowchart illustrating an intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm.

[0088] Figure 2 This is a schematic diagram of the structure of an intelligent inspection and planning system for a hydrogen cooling system based on a harmony search algorithm.

[0089] Figure 3 This is a schematic diagram of the sensor array.

[0090] Figure Descriptions: 100, System Architecture Module; 200, Data Acquisition Module; 300, Status Assessment Module; 400, Harmony Search Optimization Module; 500, Path Planning Module; 600, Control Execution Module; 700, Human-Machine Interaction Module; 210, Sensor Group; 220, Data Preprocessing Unit; 230, Wireless Communication Unit; 211, Hydrogen Purity Sensor; 212, Temperature Sensor Array; 213, Vibration Sensor; 214, Pressure Transmitter Array; 310, Feature Extraction Unit; 320, Health Scoring Model; 410, Harmony Memory Library; 420, Iterative Optimization Unit; 430, Fitness Assessment Unit; 510, Dijkstra Algorithm Unit; 520, Dynamic Adjustment Unit; 610, Command Issuance Unit; 620, Feedback Verification Unit; 710, Touch Screen; 720, Alarm Unit; 730, Log Management Unit. Detailed Implementation

[0091] To make the technical solutions and advantages of the embodiments of the present invention clearer, the exemplary embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not an exhaustive list of all embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0092] Example 1: Reference Figure 1 This embodiment describes an intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm, specifically including the following steps:

[0093] S1. Real-time acquisition of operating status data of each device in the hydrogen cooling system through the data acquisition module;

[0094] S2. The operating status data is feature extracted through the status assessment module to obtain the processed data, and a multi-dimensional status evaluation function is established.

[0095] S3. Using the harmony search optimization module, the inspection path is optimized for multiple objectives based on the processed data using the harmony search algorithm to obtain the optimization results;

[0096] S4. The optimal inspection route is generated based on the optimization results through the path planning module;

[0097] S5. Drive the inspection equipment through the control execution module, perform path planning according to the optimal inspection route, and output the verification results;

[0098] S6. Through the human-computer interaction module, store all current data, issue alarms for calibration results based on the multi-dimensional state evaluation function, and display alarm information and path planning.

[0099] Specifically, (1) the system architecture of the present invention consists of a hydrogen cooling system body, a sensor network, a sound search optimization, and a decision execution mechanism;

[0100] (2) The collaborative mechanism establishes a mapping between equipment operating parameters and the algorithm search space;

[0101] (3) The optimization model adopts an improved multi-objective fitness function;

[0102] (4) The execution process includes four stages: parameter acquisition, status assessment, path planning, and instruction issuance;

[0103] (5) The algorithm improvements include a neighborhood search strategy and a weight adaptive mechanism;

[0104] (6) The dynamic adjustment mechanism updates the search parameters based on real-time operating data;

[0105] (7) The auxiliary function module integrates a visual interface and an anomaly alarm unit.

[0106] Furthermore, step S1 includes the following steps:

[0107] S11. Arrange the temperature sensor array on the surface of the heat exchanger of the hydrogen cooling system;

[0108] S12. Install the pressure transmitter array at key nodes in the hydrogen circulation pipeline;

[0109] S13. The vibration detection unit is installed on the compressor and pump body base;

[0110] S14. Place the hydrogen purity sensor at the outlet of the hydrogen recovery device;

[0111] S15. The multi-source heterogeneous data obtained in steps S11-S14 is preprocessed by the data preprocessing unit to obtain the processed multi-source data.

[0112] In step S15, preprocessing is performed using the FIR digital filter built into the data preprocessing unit according to the filtering function, and the cutoff frequency is set to... Hz;

[0113] The filter function is expressed as:

[0114]

[0115] in, The coefficients are for the 128th order Hamming window. This is the original sampling sequence. This is the filtered output sequence;

[0116] S16. The processed data is fused using the industrial Ethernet protocol via a wireless communication unit to obtain the fused data;

[0117] In step S16, the wireless communication unit adopts LoRa modulation and has a transmission distance of ≥500m.

[0118] Furthermore, in step S2, the feature extraction unit uses wavelet packet transform to extract features from the fused data, thus decomposing the number of layers. The db4 wavelet basis function was selected to obtain the data features;

[0119] Establish a multidimensional state evaluation function based on a health scoring model. ;

[0120] Multidimensional state evaluation function Represented as:

[0121]

[0122] in, For the first The weighting coefficients of each parameter. For real-time measurements, and These are the upper and lower limits of the parameter's allowable range, respectively.

[0123] Multidimensional state evaluation function The weighting coefficients are dynamically updated using the entropy method.

[0124] The dynamic update of entropy is represented as:

[0125]

[0126] in, Let j be the information entropy value of the j-th index. For the first The indicator in the first The proportion in each sample The total number of samples.

[0127] Furthermore, step S3 includes the following steps:

[0128] S31. Construct a device status feature vector based on data characteristics. A comprehensive objective function is established through the fitness evaluation unit;

[0129] In step S31, the comprehensive objective used to evaluate the optimization goal of the inspection task is defined as follows:

[0130]

[0131] in, Total inspection time. Score the status of each node. For path redundancy, the weighting coefficient , , ;

[0132] To reflect the dynamic balance between inspection time, cost, and safety, a local objective function is defined. ;

[0133] Local objective function Represented as:

[0134]

[0135] in, Indicates the inspection time. To reduce inspection costs, Safety factor, These are dynamically adjusted weighting parameters;

[0136] Safety factor in the objective function Represented as:

[0137]

[0138] in, , Indicates the first The characteristic values ​​of each monitoring point For the corresponding threshold, This is the sensitivity adjustment coefficient;

[0139] The collaborative optimization model is defined as follows:

[0140]

[0141] in, The function represents the length of the inspection path. The function representing the equipment risk index. Represents the energy consumption assessment function; For dynamic weighting coefficients, satisfying ,and ;

[0142] S32. Generate the initial harmony memory library ,in, The number of solution vectors, This represents the number of device nodes. The j-th component of the i-th solution vector in the harmony memory HM is stored when the harmony memory is initialized. N=50 sets of feasible solution vectors are stored, and the solution vector dimension d=8 corresponds to the inspection point parameters.

[0143] S33. An improved harmony search algorithm is executed through an iterative optimization unit;

[0144] In step S33, the update formula for the improved harmony search algorithm is:

[0145]

[0146] in, For pitch adjustment bandwidth, The random numbers are uniformly distributed in the range [-1, 1]. The neighborhood search step size, The existing solution vectors are randomly drawn from the harmony memory. The solution vector is generated completely randomly. The solution vector is adjusted based on the neighborhood.

[0147] S34. Application of probability adjustment mechanism ,in, For the number of iterations, To minimize the probability of memory adoption, The initial maximum memory adoption probability, The attenuation coefficient;

[0148] In S34, the optimal path cost function is defined as:

[0149]

[0150] in, Let be the length of the i-th path segment. m / s is the speed of movement. s represents the time consumed by a single point detection. This is the penalty coefficient for state degradation.

[0151] Specifically, the weight adaptive mechanism adjusts the weights using an exponential function:

[0152]

[0153] in, The attenuation coefficient is... For real-time risk levels, Set a maximum risk value for the system;

[0154] The dynamic adjustment mechanism is based on equipment operating status data. Real-time update of search step size :

[0155]

[0156] In the formula, This is the standard parameter vector for the equipment. The maximum allowable deviation norm.

[0157] This invention constructs a device state feature vector. The matching mechanism with the harmony memory library enables dynamic optimization of the inspection strategy. The feature vector dimension includes key parameters such as temperature gradient, hydrogen purity, and sealing pressure, and its update frequency is positively correlated with the rate of change of equipment operating conditions.

[0158] Furthermore, step S4 includes the following steps:

[0159] S41. The evaluation function is generated using the Dijkstra algorithm through the Dijkstra algorithm unit;

[0160] In S41, the evaluation function Represented as:

[0161]

[0162] in, Representing a path For the The detection probability of defects in similar equipment. Assign weights to equipment importance;

[0163] S42. Construct an adjacency matrix of the device connection diagram by dynamically adjusting the units. ,in, , The node spacing, This represents the path risk value. These are risk impact factors.

[0164] Furthermore, step S5 includes the following steps:

[0165] S51. The optimized path is converted into motion control commands through the command issuing unit, and the inspection speed is dynamically adjusted according to real-time sensor data;

[0166] S52. Feedback verification unit, using the CRC-16 verification algorithm, generates the polynomial as follows: , This represents the check decision polynomial. This represents the polynomial variable corresponding to the shift register.

[0167] Specifically, the actuator interface supports protocol adaptation with tracked robots, drones, and fixed inspection devices.

[0168] Furthermore, step S6 includes the following steps:

[0169] S61. Through the log management unit, record the equipment 3D model library, store the spatial coordinates and connection topology records of each device, the mapping relationship between equipment fault modes and detection locations, and save the optimal algorithm parameters under different working conditions to build an algorithm parameter knowledge base;

[0170] S62. Through the alarm unit, based on the multidimensional state evaluation function of the health scoring model... An alarm is triggered by comparing the value with a preset value.

[0171] In S62, the alarm unit 720 is equipped with a three-level early warning mechanism: when A yellow alert is triggered at any time. An orange alert is triggered at any time. A red alert is triggered when the log management unit 730 adopts a ring buffer structure, with a storage capacity of 8GB and a data retention period of ≥30 days.

[0172] S63. Display alarm information and optimal inspection route via touch screen.

[0173] Specifically, the system executes the following process during runtime: after receiving the device status trigger signal, it retrieves the corresponding device model data from the database; it generates no fewer than three candidate paths based on the harmony search algorithm; it evaluates the fault detection coverage of each path through Monte Carlo simulation; it selects the path with the highest comprehensive score and sends it to the actuator, while simultaneously updating the algorithm parameter knowledge base.

[0174] Example 2: Reference Figure 2 This embodiment describes an intelligent inspection and planning system for a hydrogen cooling system based on a harmony search algorithm. This system is used to execute the intelligent inspection and planning method for a hydrogen cooling system based on a harmony search algorithm described in Embodiment 1. It includes a system architecture module 100, a data acquisition module 200, a status evaluation module 300, a harmony search optimization module 400, a path planning module 500, a control execution module 600, and a human-computer interaction module 700, connected sequentially.

[0175] The system architecture module 100 is connected to the data acquisition module 200, the status evaluation module 300, the harmony search optimization module 400, the path planning module 500, the control execution module 600, and the human-computer interaction module 700, respectively.

[0176] Furthermore, the system architecture module 100 includes a main controller 110, a data bus 120, and a power management unit 130;

[0177] The data acquisition module 200 includes a sensor group 210, a data preprocessing unit 220 and a wireless communication unit 230 connected in sequence. The sensor group 210 includes a hydrogen purity sensor 211, a temperature sensor array 212, a vibration sensor 213 and a pressure transmitter array 214.

[0178] The status assessment module 300 includes a feature extraction unit 310 and a health scoring model 320 connected in sequence.

[0179] The harmony search optimization module 400 includes a harmony memory bank 410, an iterative optimization unit 420, and a fitness evaluation unit 430 connected in sequence.

[0180] The path planning module 500 includes a Dijkstra algorithm unit 510 and a dynamic adjustment unit 520 connected in sequence.

[0181] The control execution module 600 includes an instruction issuing unit 610 and a feedback verification unit 620 connected in sequence.

[0182] The human-computer interaction module 700 includes a touch screen 710, an alarm unit 720, and a log management unit 730 connected in sequence.

[0183] The data bus 120 is connected to the wireless communication unit 230, the wireless communication unit 230 is connected to the feature extraction unit 310, the health scoring model 320 is connected to the iterative optimization unit 420, the fitness evaluation unit 430 is connected to the Dijkstra algorithm unit 510, the dynamic adjustment unit 520 is connected to the instruction issuing unit 610, the feedback verification unit 620 is connected to the log management unit 730, and the sensor group 210, the feature extraction unit 310, the harmony memory bank 410, the Dijkstra algorithm unit 510, the instruction issuing unit 610, and the touch screen 710 are respectively connected to the power management unit 130.

[0184] Specifically, the data bus 120 uses the CAN2.0B protocol to realize data interaction between modules; the power management unit 130 integrates multiple DC-DC conversion circuits to output a stable 12V / 5V dual voltage.

[0185] The hydrogen purity sensor 211 has a measurement range of 85%-99% vol and an accuracy of ±0.5% FS; the temperature sensor array 212 contains 8 PT100 platinum resistance temperature measurement points; the vibration sensor 213 adopts a MEMS triaxial accelerometer with a frequency response range of 10Hz-1kHz.

[0186] The instruction issuing unit 610 is connected to the actuator via an RS485 interface, with the baud rate set to 115200bps.

[0187] Although the invention has been described with reference to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and instructional purposes, and not for the purpose of interpreting or limiting the subject matter of the invention. Therefore, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of the invention is illustrative and not restrictive, and the scope of the invention is defined by the appended claims.

Claims

1. A method for intelligent inspection planning of a hydrogen cooling system based on a harmony search algorithm, characterized in that, Includes the following steps: S1. Real-time acquisition of operating status data of each device in the hydrogen cooling system through the data acquisition module; S2. The operating status data is feature extracted through the status assessment module to obtain the processed data, and a multi-dimensional status evaluation function is established. S3. Using the harmony search optimization module, the inspection path is optimized for multiple objectives based on the processed data using the harmony search algorithm to obtain the optimization results; S4. The optimal inspection route is generated based on the optimization results through the path planning module; S5. Drive the inspection equipment through the control execution module, perform path planning according to the optimal inspection route, and output the verification results; S6. Through the human-computer interaction module, store all current data, issue alarms for calibration results based on the multi-dimensional state evaluation function, and display alarm information and path planning.

2. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 1, characterized in that, S1 includes the following steps: S11. Arrange the temperature sensor array on the surface of the heat exchanger of the hydrogen cooling system; S12. Install the pressure transmitter array at key nodes in the hydrogen circulation pipeline; S13. The vibration detection unit is installed on the compressor and pump body base; S14. Place the hydrogen purity sensor at the outlet of the hydrogen recovery device; S15. The multi-source heterogeneous data obtained in steps S11-S14 is preprocessed by the data preprocessing unit to obtain the processed multi-source data. In step S15, preprocessing is performed using the FIR digital filter built into the data preprocessing unit according to the filtering function, and the cutoff frequency is set to... Hz; The filter function is expressed as: in, The coefficients are for the 128th order Hamming window. This is the original sampling sequence. This is the filtered output sequence; S16. The processed data is fused using the industrial Ethernet protocol via a wireless communication unit to obtain the fused data; In step S16, the wireless communication unit adopts LoRa modulation and has a transmission distance of ≥500m.

3. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 2, characterized in that, In step S2, the feature extraction unit uses wavelet packet transform to extract features from the fused data and decomposes the number of layers. The db4 wavelet basis function was selected to obtain the data features; Establish a multidimensional state evaluation function based on a health scoring model. ; Multidimensional state evaluation function Represented as: in, For the first The weighting coefficients of each parameter, For real-time measurements, and These are the upper and lower limits of the parameter's allowable range, respectively. Multidimensional state evaluation function The weighting coefficients are dynamically updated using the entropy method. The dynamic update of entropy is represented as: in, Let j be the information entropy value of the j-th index. For the first The indicator in the first The proportion in each sample The total number of samples.

4. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 3, characterized in that, S3 includes the following steps: S31. Construct a device status feature vector based on data characteristics. A comprehensive objective function is established through the fitness evaluation unit; In step S31, the comprehensive objective function for evaluating the optimization objective of the inspection task is defined as follows: in, Total inspection time. Score the status of each node. For path redundancy, the weighting coefficient , , ; To reflect the dynamic balance between inspection time, cost, and safety, a local objective function is defined. ; Local objective function Represented as: in, Indicates the inspection time. To reduce inspection costs, It's the safety factor. These are dynamically adjusted weighting parameters; Safety factor in local objective function Represented as: in, , Indicates the first The characteristic values ​​of each monitoring point For the corresponding threshold, This is the sensitivity adjustment coefficient; The collaborative optimization model is defined as follows: in, The function represents the length of the inspection path. The function representing the equipment risk index. Represents the energy consumption assessment function; For dynamic weighting coefficients, satisfying ,and ; S32. Generate the initial harmony memory library ,in, The number of solution vectors, This represents the number of device nodes. The j-th component of the i-th solution vector in the harmony memory HM is stored when the harmony memory is initialized. N=50 sets of feasible solution vectors are stored, and the solution vector dimension d=8 corresponds to the inspection point parameters. S33. An improved harmony search algorithm is executed through an iterative optimization unit; In step S33, the update formula for the improved harmony search algorithm is: in, For pitch adjustment bandwidth, The random numbers are uniformly distributed in the range [-1, 1]. The neighborhood search step size, The existing solution vectors are randomly drawn from the harmony memory. The solution vector is generated completely randomly. The solution vector is adjusted based on the neighborhood. S34. Application of probability adjustment mechanism ,in, For the number of iterations, To minimize the probability of memory adoption, The initial maximum memory adoption probability, The attenuation coefficient; In S34, the optimal path cost function is defined as: in, Let be the length of the i-th path segment. m / s is the speed of movement. s represents the time consumed by a single point detection. This is the penalty coefficient for state degradation.

5. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 4, characterized in that, S4 includes the following steps: S41. The evaluation function is generated using the Dijkstra algorithm through the Dijkstra algorithm unit; In S41, the evaluation function Represented as: in, Representing a path For the The detection probability of defects in similar equipment. Assign weights to equipment importance; S42. Construct an adjacency matrix of the device connection diagram by dynamically adjusting the units. ,in, , The node spacing, This represents the path risk value. These are risk impact factors.

6. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 5, characterized in that, S5 includes the following steps: S51. The optimized path is converted into motion control commands through the command issuing unit, and the inspection speed is dynamically adjusted according to real-time sensor data; S52. Feedback verification unit, using the CRC-16 verification algorithm, generates the polynomial as follows: , This represents the check decision polynomial. This represents the polynomial variable corresponding to the shift register.

7. The intelligent inspection planning method for a hydrogen cooling system based on a harmony search algorithm according to claim 6, characterized in that, S6 includes the following steps: S61. Through the log management unit, record the equipment 3D model library, store the spatial coordinates and connection topology records of each device, the mapping relationship between equipment fault modes and detection locations, and save the optimal algorithm parameters under different working conditions to build an algorithm parameter knowledge base; S62. Through the alarm unit, based on the multidimensional state evaluation function of the health scoring model... An alarm is triggered by comparing the value with a preset value. In S62, the alarm unit 720 is equipped with a three-level early warning mechanism: when A yellow alert is triggered at any time. An orange alert is triggered at any time. A red alert is triggered when the log management unit 730 adopts a ring buffer structure, with a storage capacity of 8GB and a data retention period of ≥30 days. S63. Display alarm information and optimal inspection route via touch screen.

8. A smart inspection and planning system for a hydrogen cooling system based on a harmony search algorithm, characterized in that, The method for executing the intelligent inspection planning method for a hydrogen cooling system based on the harmony search algorithm as described in any one of claims 1-7 includes a system architecture module (100), a data acquisition module (200), a status assessment module (300), a harmony search optimization module (400), a path planning module (500), a control execution module (600), and a human-computer interaction module (700) connected in sequence. The system architecture module (100) is connected to the data acquisition module (200), the status evaluation module (300), the harmony search optimization module (400), the path planning module (500), the control execution module (600), and the human-computer interaction module (700), respectively.

9. The intelligent inspection and planning system for a hydrogen cooling system based on a harmony search algorithm according to claim 8, characterized in that, The system architecture module (100) includes a main controller (110), a data bus (120), and a power management unit (130) connected in sequence. The data acquisition module (200) includes a sensor group (210), a data preprocessing unit (220) and a wireless communication unit (230) connected in sequence. The sensor group (210) includes a hydrogen purity sensor (211), a temperature sensor array (212), a vibration sensor (213) and a pressure transmitter array (214). The status assessment module (300) includes a feature extraction unit (310) and a health scoring model (320) connected in sequence. The harmony search optimization module (400) includes a harmony memory bank (410), an iterative optimization unit (420), and a fitness evaluation unit (430) connected in sequence. The path planning module (500) includes a Dijkstra algorithm unit (510) and a dynamic adjustment unit (520) connected in sequence. The control execution module (600) includes an instruction issuing unit (610) and a feedback verification unit (620) connected in sequence. The human-computer interaction module (700) includes a touch screen (710), an alarm unit (720), and a log management unit (730) connected in sequence. The data bus (120) is connected to the wireless communication unit (230), the wireless communication unit (230) is connected to the feature extraction unit (310), the health scoring model (320) is connected to the iterative optimization unit (420), the fitness evaluation unit (430) is connected to the Dijkstra algorithm unit (510), the dynamic adjustment unit (520) is connected to the instruction issuing unit (610), the feedback verification unit (620) is connected to the log management unit (730), and the sensor group (210), the feature extraction unit (310), the harmony memory bank (410), the Dijkstra algorithm unit (510), the instruction issuing unit (610), and the touch screen (710) are respectively connected to the power management unit (130).

Images