Ship cargo hold monitoring system and device based on optical fiber sensor

Abstract

The invention relates to the field of ship cargo hold monitoring, and discloses a ship cargo hold monitoring system and device based on an optical fiber sensor, and the system comprises an optical fiber sensing module, a cargo state discrimination module, and a decision control module. The multi-parameter optical fiber sensor obtains various target information including oxygen content, temperature, humidity, vibration and the like in a ship cabin and modulates the information into an electric signal; establishing a ship cargo information database to activate a corresponding measurement sub-module, performing model training and prediction by using machine learning algorithms including a convolutional neural network and the like, generating data prediction results including gas alarm limit value information, material form information and the like, and judging whether a threshold value is reached or not according to related index information; corresponding parameters reaching a threshold value are obtained and enter a decision control module; and the decision control module controls the corresponding indicating lamps to flicker and give an alarm according to the data calculation result, and generates protection operation strategy suggestions and misoperation matter attention according to the data calculation result.

Description

Technical Field

[0001] The present invention relates to the field of ship cargo hold monitoring, and in particular to a ship cargo hold monitoring system and device based on optical fiber sensors. Background Art

[0002] As the largest transportation vehicle, ships are responsible for the transportation of most bulk commodities. However, when sailing on the sea, they will encounter various weather and sea conditions, which are very likely to cause marine accidents and cause huge economic losses. At the same time, the types of cargo loaded on ships are complex, including flammable cargo, easy-flowing cargo, valuable cargo, etc., and ships need to be monitored in real time to prevent cargo from burning, leaking, etc. At the same time, the impact of different cargo conditions on ships and the solutions are also different, so a system that can provide correct safety measures is needed.

[0003] At present, the main problems of the ship cargo hold monitoring system are as follows: 1. The sensor sensing module is not sensitive enough and cannot accurately judge the impact of some goods that have strict requirements on temperature, humidity and other conditions. For example, when the water content exceeds the fluidization moisture limit of the cargo, the fluidized cargo will accelerate to transform into a liquid or fluid state, causing the ship to tilt or even flip, causing a large number of casualties and property losses, so it is necessary to strictly monitor the humidity changes in the cabin. 2. Due to the characteristics of the ship's navigation environment and the interference of factors such as vibration, accurate feature information cannot be extracted from the ship, so it is impossible to grasp the internal situation of the cabin in time and take measures. 3. Since the occurrence of accidents is uncertain and the handling measures for each accident are different, some on-duty personnel lack experience and are very likely to make wrong operations in the process of handling emergencies, causing the accident to expand.

[0004] Distributed fiber optic sensing systems can accurately measure a variety of parameters and effectively solve the problem of insufficient sensitivity of the sensing module. For example, application number 202310072639X discloses a local fire prevention method and device for collective dormitories based on fiber optic sensors. Distributed fiber optic sensing technology is used to collect multiple types of information and identify the mutation rate of target information to achieve hierarchical control of fire forecasts, warnings and alarms in collective dormitories. However, the above invention only uses the mutation rate as the basis for judgment, lacks in-depth exploration and analysis of the data, and cannot effectively control the impact of empirical values ​​on the actual environment. At the same time, this method is only applicable to fire protection and control work. The types of risks of cargo in the cabin are relatively complex and cannot be effectively monitored using this method. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a ship cargo hold monitoring system and device.

[0006] To achieve the above-mentioned purpose, the present invention adopts the following technical solution: a ship cargo hold monitoring system and device is composed of an optical fiber sensing module (1), a cargo state determination module (2), and a decision control module (3); The optical fiber sensing module (1) uses a multi-parameter optical fiber sensor to obtain multiple target information including oxygen content, temperature, humidity, vibration, etc. in the ship cabin, and modulates the information into an electrical signal; The cargo status identification module (2) establishes a ship cargo information database, automatically identifies the cargo type and related indicators to be detected by inputting the name and number of the cargo transported by the ship, and activates the corresponding measurement submodule. The submodule uses machine learning algorithms including convolutional neural network, random forest and support vector machine to perform model training and prediction based on the measured information data set, generates data prediction results such as gas alarm limit information and material form information, and determines whether the corresponding parameter reaches the threshold value based on the relevant indicator information output by the cargo information database and the various status information in the parameter measurement and prediction submodule. If there is a corresponding parameter that reaches the threshold value, the system obtains the corresponding parameter and enters the decision control module (3). Otherwise, the system does not make any further response. The decision control module (3) controls the corresponding indicator light of the parameter reaching the threshold value to flash and issue an alarm according to the data calculation result, and generates operation strategy suggestions and wrong operation precautions according to the data calculation result.

[0007] As a further solution of the present invention, the optical fiber sensing module (1) comprises a plurality of distributed optical fiber sensors arranged at equal intervals inside the cabin, dividing the cabin into a plurality of sub-units for measurement; The distributed optical fiber sensor is composed of a laser source (1-1), a circulator (1-2), a photoelectric detection module (1-3), a computer (1-4) and a modulator chain (1-5); The optical fiber sensor is composed of an incident optical fiber, a modulator, and an output optical fiber. The modulator is the main structure of the sensor, which modulates the target information of the measurement point into the optical signal emitted by the laser source; The plurality of modulators are connected in series to form a modulator chain (1-5), which is placed in a ship cabin, and different measurement points are set in sequence to measure and extract multiple parameter information; The photoelectric detection module includes a photoelectric detector, a rectifier and a filter, and is used to convert the extracted optical signal into an electrical signal and input it into a computer (1-4) for further processing.

[0008] As a further solution of the present invention, the cargo state identification module (2) comprises a data storage submodule (2-1), a cargo type identification submodule (2-2), a data preprocessing submodule (2-3), a parameter measurement and prediction submodule (2-4), and a safety state identification submodule (2-5); The data storage submodule (2-1) establishes a complete marine cargo data information database based on various marine cargo management regulations, including but not limited to gas explosion range, transportable moisture limit of fluid cargo, etc.; The cargo type identification submodule (2-2) identifies the corresponding cargo type information by inputting the transport name and category number of the transported cargo, obtains the corresponding index information, and activates the corresponding parameter measurement submodule; The data preprocessing submodule (2-3) uses Gaussian filtering to smooth and reduce noise on the signal, and uses principal component analysis to generate a key feature set of multiple measurement features; The parameter measurement and prediction submodule (2-4) includes a gas alarm limit prediction submodule (2-4-1), a cargo form prediction submodule (2-4-2), and a cargo stability prediction submodule (2-4-3); The gas alarm limit prediction submodule (2-4-1) uses a convolutional neural network based on characteristic information such as temperature, pressure and oxygen concentration, and predicts and corrects the gas alarm limit result; The cargo shape prediction submodule (2-4-2) predicts the shape of the cargo that is easy to flow in the cabin by using a random forest algorithm based on characteristic information such as vibration and humidity; The cargo stability prediction submodule (2-4-3) generates cargo stability results in the cabin using a support vector machine algorithm based on cargo shape and vibration characteristic information; The safety status determination submodule (2-5) determines whether the information of the corresponding parameter reaches a threshold value based on the relevant indicator information output by the cargo information database and the various status information in the parameter measurement and prediction submodule (2-4). If there is corresponding parameter information that reaches the threshold value, the system obtains the corresponding parameter information and enters the decision control module (3).

[0009] As a further solution of the present invention, the decision control module (3) includes an alarm submodule (3-1) and a decision management submodule (3-2); The alarm submodule (3-1) generates different alarm signals according to different discrimination information, controls the indicator light to flash and issues an alarm; The decision management submodule (3-2) obtains the identified cargo type (3-2-1), and generates operation strategy suggestions and incorrect operation precautions (3-2-2), including combustible operation strategy suggestions and combustible incorrect operation precautions.

[0010] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through distributed optical fiber sensing technology, the cargo information inside the cabin can be obtained more comprehensively, and the key data of the cargo can be accurately obtained, providing timely and reliable information support for subsequent decision-making; by combining with the sea freight data information database, other key safety evaluation indicators in addition to common indicators are obtained, and specific cargo information is provided to provide a basis for cargo status judgment and corresponding decision-making, reduce a large amount of useless measurements and data calculations, and improve system efficiency; use advanced machine learning technologies such as support vector machines to ensure the accuracy and robustness of the prediction; in addition, the system uses gas alarm limit results, cargo status results and stability results instead of traditional factors such as temperature, vibration, humidity, etc. as the basis for judging the cargo status, reducing the possibility of failure to identify in time due to environmental factors, and improving the accuracy of the system; use decision tree analysis to generate decision suggestions and wrong operation precautions to help personnel avoid risks caused by wrong operations, accurately and quickly operate, reduce cargo losses, and improve personnel safety and ship transportation revenue. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a system flow chart of the present invention; Figure 2 is a schematic diagram of an optical fiber sensing module of the present invention; Figure 3 This is a flow chart of the cargo status identification module of the present invention; Figure 4 This is a flow chart of the gas alarm limit prediction submodule of the present invention; Figure 5 This is a flow chart of the cargo form prediction submodule of the present invention; Figure 6 It is a schematic diagram of the decision control module of the present invention; Figure 7 This is a flow chart of the decision management submodule of the present invention. Specific implementation plan

[0012] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0013] The term "including" and other similar expressions in the specification or claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions, such as a process, method, system, or device including a series of steps or units is not limited to the listed steps or units. In addition, in the description of the present invention, the meaning of "etc." is two or more, unless otherwise clearly and specifically defined.

[0014] Example 1, please refer to Figure 1The present invention provides a technical solution: a ship cargo hold monitoring system and device, which is composed of an optical fiber sensing module (1), a cargo state determination module (2), and a decision control module (3); the optical fiber sensing module (1) uses a multi-parameter optical fiber sensor to obtain multiple target information including oxygen content, temperature, humidity, vibration, etc. in the ship cabin, and modulates the information into an optical signal; The cargo status identification module (2) establishes a ship cargo information database, automatically identifies the cargo type and related indicators to be detected by inputting the name and number of the cargo transported by the ship, and activates the corresponding measurement submodule. The model module uses machine learning algorithms including convolutional neural network, random forest and support vector machine to perform model training and prediction based on the measured information data set, and generates data prediction results such as gas alarm limit information and material form information. According to the relevant indicator information output by the cargo information database and the various state information in the parameter measurement and prediction submodule, it determines whether the corresponding parameter reaches the threshold. If there is a corresponding parameter that reaches the threshold, the system obtains the corresponding parameter and enters the decision control module (3). Otherwise, the system does not respond further; the decision control module (3) controls the corresponding indicator light of the corresponding parameter that reaches the threshold to flash and issue an alarm, and generates operation strategy suggestions and wrong operation matters according to the data calculation results.

[0015] First, the optical fiber sensing module (1) uses distributed optical fiber sensors to obtain spectral data of temperature, oxygen concentration, pressure, humidity, and vibration in the ship's cabin, and uses a demodulation submodule to convert the output signal into an electrical signal, thereby obtaining sensitive characteristic information, which is helpful for subsequent data processing.

[0016] The cargo status identification module (2) establishes a data information base for sea cargo based on various ship transportation cargo rules, and obtains all relevant index information of cargo in the cabin. These index information can help select the feature information that needs to be processed, reduce useless measurements and data calculations, and provide a basis for identifying the cargo status. After data processing, machine learning methods such as convolutional neural networks, random forests, and support vector machines are used to train and predict models based on temperature, oxygen concentration, air pressure, humidity, and vibration data to generate cabin gas alarm limits, easy-flow cargo morphology, and cargo stability prediction results. The generated results can more accurately describe the characteristic attributes of the cargo and help accurately identify the safety status of the cargo.

[0017] The decision control module (3) generates operation strategy suggestions and notes on erroneous operation based on the data calculation results. These decision suggestions can help personnel avoid the risks caused by erroneous operation, operate accurately and quickly, reduce cargo losses, and improve personnel safety and ship transportation revenue.

[0018] See also Figure 2, the optical fiber sensing module is composed of a number of distributed optical fiber sensors arranged at equal intervals inside the cabin, dividing the cabin into a number of sub-units for measurement; each distributed optical fiber sensor includes a laser light source (1-1), a circulator (1-2), a photoelectric detector module (1-3), a computer (1-4), a modulator chain (1-5) and a single-mode optical fiber for transmission; the laser light source uses an erbium-doped optical fiber laser to emit near-infrared wavelength light; the circulator sends a test pulse signal to the optical fiber, and guides the scattered signal reflected from the optical fiber to the detector, thereby measuring the characteristic changes along the optical fiber; the optical fiber sensor is composed of an incident optical fiber, a modulator, and an output optical fiber, the modulator is the main structure of the sensor, and a number of optical fiber sensors are connected in series and placed in the ship's cabin, and different measurement points are set in turn to measure and extract multiple parameter information; the optical detector module includes a photoelectric detector, a rectifier and a filter, which are used to convert the extracted optical signal into an electrical signal and input it into a computer for further processing.

[0019] See also Figure 3 The cargo status identification module includes a data storage submodule (2-1), a cargo type identification submodule (2-2), a data preprocessing submodule (2-3), a parameter measurement and prediction submodule (2-4), and a safety status identification submodule (2-5); The data storage submodule (2-1) establishes a complete marine cargo data information database based on various marine cargo management regulations, including but not limited to the explosion range of flammable gases, the transportable moisture limit of fluid cargoes, etc.; The cargo type identification submodule (2-2) inputs the name and category number of the transported cargo, identifies the corresponding cargo type information and the corresponding index information, and activates the corresponding parameter measurement submodule; The data preprocessing submodule (2-3) uses Gaussian filtering to smooth and reduce noise on the signal, and uses principal component analysis to generate a key feature set of multiple measurement features; The parameter measurement and prediction submodule (2-4) includes a gas alarm limit prediction submodule (2-4-1), a cargo shape prediction submodule (2-4-2), and a cargo stability prediction submodule (2-4-3); The gas alarm limit prediction submodule (2-4-1) uses a convolutional neural network based on characteristic information such as temperature, pressure and oxygen concentration to predict and correct the gas alarm limit results and provide a basis for judging flammable materials; The cargo shape prediction submodule (2-4-2) uses the random forest algorithm based on characteristic information such as vibration and humidity to predict the moisture content of the easy-flowing cargo in the cabin, generate the easy-flowing cargo shape results, and provide a judgment basis for easy-flowing solid cargo; The cargo stability prediction submodule (2-4-3) generates cargo stability results in the cabin based on cargo shape and vibration characteristic information using the support vector machine algorithm. Cargo stability is an important evaluation indicator of the cargo safety status in the cabin and can provide a basis for judging the safety status of all cargo. The safety status determination submodule (2-5) determines whether the corresponding parameter information reaches the threshold value according to the relevant indicators output by the cargo information database. If there is corresponding parameter information that reaches the threshold value, the system obtains the corresponding parameter information and enters the decision control module (3).

[0020] The data storage submodule (2-1) establishes a complete marine cargo data information database, which stores information including the name and category number of the cargo transported by the ship, the type of cargo and index parameters.

[0021] The data storage submodule (2-1) establishes a maritime cargo data information database according to the cargo categories classified in the International Solid Bulk Cargo Code, the International Maritime Dangerous Goods Code, the International Convention for the Prevention of Pollution from Ships and the International Convention for the Safety of Life at Sea.

[0022] The cargo type identification submodule (2-2) identifies the corresponding cargo type information by inputting the transport name and category number of the transported cargo, obtains key safety evaluation index information based on the specific classification information of the cargo, and activates the corresponding parameter measurement submodule.

[0023] It needs to be explained that according to the provisions of the International Solid Bulk Cargo Code, the International Maritime Dangerous Goods Code, the International Convention for the Prevention of Pollution from Ships and the International Convention for the Safety of Life at Sea, the types of ship cargo mainly include: solid bulk cargo, dangerous goods, liquid bulk cargo, and non-dangerous goods. Dangerous goods include some solid bulk cargoes and liquid bulk cargoes, and non-dangerous goods include some solid bulk cargoes. When obtaining cargo indicators, these goods are not reclassified and obtained. The specific classification and indicator information of the goods are as follows:

[0024] ①Dangerous goods Specific categories include: explosives, gases, flammable liquids, flammable solids, oxidizers and organic peroxides, toxic and infectious substances, radioactive materials, corrosive substances, and miscellaneous dangerous goods; Index information: temperature, oxygen concentration, pressure.

[0025] ②Solid bulk cargo The specific categories include: Class A cargo, Class B cargo and Class C cargo. Class A cargo refers to cargo that is easy to flow, Class B cargo refers to cargo that has chemical hazards during transportation, and Class C cargo refers to cargo that is not easy to liquefy and does not have chemical hazards. Index information: humidity, pressure.

[0026] ③Liquid bulk cargo Specific categories include: petroleum and petroleum products, hazardous liquid substances, liquefied gases, acids and alkalis, liquid foods and edible oils, waste oils and waste liquids; Indicator information: temperature.

[0027] ④Non-dangerous goods Specific categories include: general cargo, container cargo, bulk cargo, ro-ro cargo, and special cargo; Indicator information: vibration.

[0028] It is worth noting that, according to the above cargo types, the cargo types in the cabin are divided into categories: Category 1, combustibles, including explosives, flammable gases, flammable liquids and flammable solids, activating the gas alarm limit prediction submodule; Category 2, easy-flowing goods, activate the goods shape prediction submodule; Category 3, toxic gas, the measurement indicator information is the gas concentration in the cabin; Category 4, oxidants or organic peroxides, the measurement indicator information is the oxygen concentration in the cabin; Category 5, general cargo, no special indicator information.

[0029] It should be explained that there are empirical values ​​for the indicator information in category 1 and category 2, and there are errors in the process of judging the safety status. The special parameter measurement and prediction module is activated to generate indicator information for judgment; the safety evaluation indicator information of all goods has stability information. Therefore, for any category of goods, the cargo stability prediction submodule (2-4-3) in the parameter measurement and prediction module should be activated for identification.

[0030] The data preprocessing submodule (2-3) performs smoothing and noise reduction processing on the signal and generates a key feature set of multiple measurement features. The specific steps are as follows: The optical fiber sensing module (1) transmits the converted electrical signal to the data preprocessing submodule (2-3), generates a Gaussian kernel for convolution operation, and the noise in the signal is smoothed to generate a more stable signal; Select appropriate wavelet basis and decomposition layer number, decompose the signal by wavelet to get wavelet coefficients, remove noise in the signal by threshold denoising, and then perform wavelet reconstruction. The reconstructed signal further removes noise and becomes clearer. The principal component analysis method extracts the main features from the signal by identifying and retaining the main components. Using the principal component analysis method, a key feature set of multiple measurement features is generated.

[0031] After receiving the multiple parameter feature information output by the data preprocessing module (2-3), the parameter measurement and prediction submodule (2-4) measures the index information identified in the cargo type discrimination module (2-2). It is worth noting that the present invention abandons the discrimination and control of direct index information, but first generates key index information, and uses machine learning methods such as convolutional neural networks, random forests and support vector machines to predict factors that directly affect cargo safety, such as cabin gas alarm limits, easy-flow cargo forms and cargo stability, thereby reducing errors caused by ideal formula conversion and empirical value substitution, making the measurement results more accurate and providing a reliable basis for safety status discrimination and decision-making control.

[0032] See also Figure 3 , the parameter measurement and prediction submodule (2-4) includes a gas alarm limit prediction submodule (2-4-1), a cargo shape prediction submodule (2-4-2), and a cargo stability prediction submodule (2-4-3); The gas alarm limit prediction submodule (2-4-1) is applicable to the prediction of combustibles. Based on characteristic information such as temperature, pressure and oxygen concentration, it uses a convolutional neural network to generate the measurement results of the gas composition in the cabin, and predict and correct the gas alarm limit results; The cargo shape prediction submodule (2-4-2) generates the shape results of easy-flow cargo based on the characteristic information such as vibration and humidity using the random forest algorithm; The cargo stability prediction submodule (2-4-3) generates cargo stability results in the cabin based on cargo shape and vibration characteristic information using a support vector machine algorithm.

[0033] See also Figure 4 The gas alarm prediction submodule (2-4-2) collects temperature, oxygen content, and combustible gas content over a period of time, and uses a convolutional neural network to predict the gas alarm threshold under different conditions. The steps for predicting the alarm threshold are as follows: Step S1, receiving temperature, oxygen content, and combustible gas content data transmitted by multiple distributed optical fiber sensors over a period of time; Step S2, obtaining the finite element combustible gas proportion and the rate of change of the gas concentration in the finite element over time according to the calculation; Step S3, taking the temperature, the finite element combustible gas proportion and the rate of change of the gas concentration in the finite element over time as fixed input tensors, and performing a convolution operation, wherein the convolution operation can be expressed as: Among them, i and j are the output positions, m and n are the indices of the convolution kernel; Step S4, introducing nonlinear characteristics to help the network learn complex features; Step S5, perform pooling, select the maximum value in a specific area (such as a 2x2 window), and reduce the size of the feature map. The formula for the pooling operation is as follows: MaxPooling(R) = max(R); Where |R| is the number of elements in the window; Step S6, flattening the convolution and pooling feature map into a one-dimensional vector, multiplying the flattened feature vector with a set of weight matrices, and generating output through an activation function; Step S7, using the training set to train the model, the output layer is the alarm limit of the gas in the cabin.

[0034] See also Figure 5 The cargo shape prediction submodule is suitable for the detection of easy-flow cargo. Based on the characteristic information such as vibration and humidity, the random forest algorithm is used to predict the shape information of easy-flow cargo in the cabin. The steps are as follows: Step S1, receiving humidity and vibration data information within a period of time transmitted by multiple distributed optical fiber sensors; Step S2, data fusion of liquid material flow model characteristic information, humidity and single vibration material displacement to form a data set with unified dimension, and divide the training set and the test set into a ratio of 8:2. The training set is used to build a random forest model, and the test set is used to evaluate the performance of the model. Step S3, randomly extract multiple sample subsets from the original data set with replacement, and train a decision tree for each sample subset. The selected features are used to split the tree during construction until the specified depth is reached or the number of samples at the leaf node is less than the threshold. The classification process can be expressed by the formula: Where Entropy represents entropy, |D i | is the number of samples of the i-th category, |D| is the total number of samples; Step S4: For new samples, prediction is performed through each decision tree, and the final prediction result is the majority voting result of the prediction results of all trees to predict and classify the cargo form information.

[0035] The cargo stability prediction submodule (2-4-3) is suitable for conventional cargo detection. Based on the finite element pressure information provided by each sensor, the support vector machine uses the radial basis function as the kernel function to map the input features to a high-dimensional space and establish an optimal hyperplane for classification or regression. The training of these models can build a cargo stability prediction model, and by learning the patterns and rules related to the feature set, accurate prediction and classification of stability can be achieved.

[0036] The safety status determination submodule (2-5) determines whether the corresponding parameter reaches the threshold value according to the relevant indicator information output by the cargo information database and the various status information in the parameter measurement and prediction submodule (2-4). If there is corresponding parameter information that reaches the threshold value, the system obtains the corresponding parameter information and enters the decision control module (3), wherein the combustible state is determined based on the combustible gas stability parameter generated by the modified gas alarm limit.

[0037] It is worth noting that the safety status judgment module (2-5) not only judges the three parameters of combustible gas alarm limit, cargo form, and cargo stability, but also judges the parameters of toxic gas stability, oxidant and organic peroxide. The main judgment parameter for the stability of the two substances is the gas concentration, so there is no need to establish a new parameter prediction module for prediction.

[0038] See also Figure 6 The decision control module (3) comprises an alarm submodule (3-1) and a decision management submodule (3-2), wherein the alarm submodule (3-1) generates different alarm signals according to different discrimination information, controls the indicator light to flash and issues an alarm; The decision management submodule (3-2) generates relevant operation information to assist and standardize decision-making based on different cargo category information.

[0039] See also Figure 7 , the decision management submodule obtains the identified cargo type (3-2-1), and generates operation strategy suggestions and wrong operation precautions (3-2-2); When the gas stability parameter exceeds the threshold, it generates combustible operation strategy suggestions and combustible incorrect operation precautions; When the category is toxic gas and the gas concentration in the cabin exceeds the threshold, generate toxic gas operation strategy suggestions and toxic gas incorrect operation precautions; When the category is oxidant or organic peroxide, and the oxygen concentration in the cabin exceeds the threshold, generate oxidant organic peroxide operation strategy recommendations and oxidant vortex organic peroxide incorrect operation precautions; When the category is easy-flowing goods and the shape parameters of the goods exceed the threshold, easy-flowing material operation strategy suggestions and easy-flowing material incorrect operation precautions are generated; For the types of goods used, when the stability parameters of the goods exceed the threshold, a fixed operation strategy recommendation for the goods is generated. At the same time, according to the types of goods in the type identification submodule, relevant incorrect operation precautions for the goods are generated.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A ship cargo hold monitoring system and device based on optical fiber sensors, comprising an optical fiber sensing module (1), a cargo state determination module (2), and a decision control module (3); The optical fiber sensing module (1) uses a multi-parameter optical fiber sensor to obtain multiple target information including oxygen content, temperature, humidity, vibration, etc. in the ship cabin, and modulates the information into an electrical signal; The cargo status identification module (2) establishes a ship cargo information database, automatically identifies the cargo type and related indicators to be detected by inputting the name and number of the cargo transported by the ship, and activates the corresponding measurement submodule. The submodule uses machine learning algorithms including convolutional neural networks, random forests and support vector machines to perform model training and prediction based on the measured information data set, generates data prediction results such as gas alarm limit information and material form information, and determines whether the corresponding parameters reach the threshold value based on the relevant indicator information output by the cargo information database and the various status information in the parameter measurement and prediction submodule. If there are corresponding parameters that reach the threshold value, the system obtains the corresponding parameters and enters the decision control module. Otherwise, the system does not respond further. The decision control module (3) controls the corresponding indicator light of the parameter that reaches the threshold value to flash and issue an alarm based on the data calculation result, and generates operation strategy suggestions and wrong operation matters to be noted based on the data calculation result.

2. According to claim 1, the optical fiber sensing module (1) is composed of a plurality of distributed optical fiber sensors arranged at equal intervals inside the cabin, dividing the cabin into a plurality of sub-units for measurement; The distributed optical fiber sensor is composed of a laser source (1-1), a circulator (1-2), a photoelectric detection module (1-3), a computer (1-4) and a modulator chain (1-5); The optical fiber sensor is composed of an incident optical fiber, a modulator, and an output optical fiber. The modulator is the main structure of the sensor, which modulates the target information of the measurement point into the optical signal emitted by the laser source; The plurality of modulators are connected in series to form a modulator chain (1-5), which is placed in a ship cabin, and different measurement points are set in sequence to measure and extract multiple parameter information; The photoelectric detection module includes a photoelectric detector, a rectifier and a filter, and is used to convert the extracted optical signal into an electrical signal and input it into a computer (1-4) for further processing.

3. According to claim 1, the cargo status identification module (2) comprises a data storage submodule (2-1), a cargo type identification submodule (2-2), a data preprocessing submodule (2-3), a parameter measurement and prediction submodule (2-4), and a safety status identification submodule (2-5); The data storage submodule (2-1) establishes a complete marine cargo data information database based on various marine cargo management regulations, including but not limited to gas explosion range, transportable moisture limit of fluid cargo, etc.; The cargo type identification submodule (2-2) identifies the corresponding cargo type information by inputting the transport name and category number of the transported cargo, obtains the corresponding index information, and activates the corresponding parameter measurement submodule; The data preprocessing submodule (2-3) uses Gaussian filtering to smooth and reduce noise on the signal, and uses principal component analysis to generate a key feature set of multiple measurement features; The parameter measurement and prediction submodule (2-4) includes a gas alarm limit prediction submodule (2-4-1), a cargo form prediction submodule (2-4-2), and a cargo stability prediction submodule (2-4-3); The gas alarm limit prediction submodule (2-4-1) uses a convolutional neural network based on characteristic information such as temperature, pressure and oxygen concentration, and predicts and corrects the gas alarm limit result; The cargo shape prediction submodule (2-4-2) predicts the shape of the cargo that is easy to flow in the cabin by using a random forest algorithm based on characteristic information such as vibration and humidity; The cargo stability prediction submodule (2-4-3) generates cargo stability results in the cabin using a support vector machine algorithm based on cargo shape and vibration characteristic information; The safety status determination submodule (2-5) determines whether the corresponding parameter information reaches a threshold value based on the relevant indicator information output by the cargo information database and the various status information in the parameter measurement and prediction submodule (2-4). If there is corresponding parameter information that reaches the threshold value, the system obtains the corresponding parameter information and enters the decision control module.

4. According to claim 1, the decision control module comprises an alarm submodule (3-1) and a decision management submodule (3-2); The alarm submodule (3-1) generates different alarm signals according to different discrimination information, controls the indicator light to flash and issues an alarm; The decision management submodule (3-2) obtains the identified cargo type (3-2-1), and generates operation strategy suggestions and incorrect operation precautions (3-2-2), including combustible operation strategy suggestions and combustible incorrect operation precautions.

Images