Underwater wireless data transmission method of deepwater riser monitoring system

Abstract

The invention relates to the technical field of underwater wireless data transmission, in particular to an underwater wireless data transmission method of a deepwater riser monitoring system. The method comprises the following steps: carrying out sampling work according to a set sampling frequency by utilizing an acquisition device, and temporarily storing data obtained by sampling in an SRAM (Static Random Access Memory) of an MCU (Microprogrammed Control Unit); a reading device is utilized to approach the acquisition device, a Hall sensor on the acquisition device is triggered, and a 2.4 G wireless digital transmission module of the acquisition device is awakened to enter a half-duplex communication mode; the reading device sends an instruction through the wireless digital transmission module and specifies a time period and a transmission mode of data needing to be read, and the acquisition device extracts corresponding data according to the instruction and transmits the data to the reading device through the 2.4 G wireless module in a half-duplex communication mode. According to the invention, a timestamp synchronous storage mode is adopted, so that the collected data is prevented from time offset and displacement. And in the wireless data acquisition process, continuous data acquisition is still kept.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater wireless data transmission, in particular to an underwater wireless data transmission method of a deepwater riser monitoring system. Background Art

[0002] With increasing global energy demand and the depletion of onshore oil and gas resources, offshore oil production has become a crucial means of meeting this demand. Marine risers, crucial for transporting subsea oil and natural gas to land or other facilities, are crucial. Furthermore, marine risers play a crucial role in subsea oil and gas production, supporting subsea well facilities, and seabed expansion projects. Therefore, research on marine risers has become a crucial topic in the field of marine engineering.

[0003] Marine risers are a crucial component of modern marine engineering structures, yet they are also vulnerable components. They typically flow high-pressure oil or gas flows through them, while also being subjected to wave and current loads. Due to the complexity of the marine environment in which they operate, numerous factors influence their operation. Under the influence of internal fluids and external environmental loads, risers can be subject to collisions, fatigue damage, and fracture failure, resulting in significant damage, not only to the project itself but also to serious secondary hazards.

[0004] The main shortcomings of existing riser monitoring technology include: 1. Limited monitoring information; 2. Short underwater operation time; 3. Difficulty in signal reading; 4. Safety risks; and 5. Complex data processing and analysis. The difficulty in signal reading refers to the fact that in deepwater environments, sensor signal reading and data transmission are affected by factors such as water pressure and temperature, often leading to signal attenuation and transmission delays, which affect the accuracy and timeliness of monitoring data. Therefore, a method for underwater wireless data transmission in a deepwater riser monitoring system is provided. Summary of the Invention

[0005] The object of the present invention is to provide an underwater wireless data transmission method for a deepwater riser monitoring system, so as to solve the problems of difficulty in reading signals and low data transmission efficiency of the riser monitoring technology proposed in the above background technology.

[0006] To achieve the above object, the present invention provides an underwater wireless data transmission method for a deepwater riser monitoring system, comprising the following steps:

[0007] S1. Use the acquisition device to perform sampling according to the set sampling frequency, and temporarily store the sampled data in the SRAM of the MCU itself;

[0008] S2. Use the reading device to approach the acquisition device, trigger the Hall sensor on the acquisition device, wake up the acquisition device's 2.4G wireless digital transmission module and enter half-duplex communication mode, and introduce a Markov decision process combined with dynamic state transition probability estimation and multi-objective reward function to optimize the triggering process;

[0009] S3. The reading device sends an instruction through the wireless digital transmission module, specifying the time period and transmission method for the data to be read. The acquisition device extracts the corresponding data according to the instruction and transmits it to the reading device through the 2.4G wireless module in half-duplex communication mode;

[0010] S4. After the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the sleep standby state, completing the entire data transmission process.

[0011] As a further improvement of the present technical solution, in S1, the acquisition device is used to perform sampling according to a set sampling frequency, and the sampled data is temporarily stored in the SRAM of the MCU itself, including the following steps:

[0012] S1.1. Configure the acquisition device, including the MCU and sensors inside the acquisition device, and set the sampling frequency, data format, and timestamp parameters;

[0013] S1.2. According to the set sampling frequency, the MCU triggers the sensor to start collecting data at regular intervals;

[0014] S1.3. Each time a sample is taken, the sensor obtains the status information of the riser with a timestamp and sends the status information data to the MCU;

[0015] S1.4. After receiving the status information data of the sensor, the MCU temporarily stores the status information data in the SRAM integrated inside the MCU, using the timestamp synchronization storage method and introducing the Kalman filter algorithm for time calibration;

[0016] S1.5. When the accumulated state information data in the SRAM reaches a preset threshold value a, the accumulated state information data temporarily stored in the SRAM is transferred to a large-capacity memory for long-term storage;

[0017] S1.6. The acquisition device continues to sample at the set sampling frequency and continuously adds the newly collected data to the SRAM until the conditions are met for batch transfer to the large-capacity storage.

[0018] As a further improvement of the present technical solution, in S1.4, a Kalman filter algorithm is introduced to perform time calibration, including the following steps:

[0019] S1.41. Define the state vector x k=[t k , b k ] T , where t k Indicates real time, b k represents the clock deviation, T represents the transposition operation, and the state transfer matrix F is defined k , used to express the change law of the state from time k to time k+1; k represents the time;

[0020] S1.42, based on the state estimate at the previous moment State transition matrix F k , predict the current state:

[0021] S1.43. Calculate the error covariance matrix of the predicted state Among them, Q k represents the covariance matrix of process noise; P k-1 represents the state estimation error covariance matrix at time k-1;

[0022] S1.44, get the current measurement value z k , using the observation matrix to relate the actual measurements to the state vector: k =H k ·x k +v k , where v k represents the measurement noise; H k represents the observation matrix;

[0023] S1.45, calculate the Kalman gain K k , and update the state estimate based on the measurement value and update the error covariance matrix P k .

[0024] As a further improvement of this technical solution, in S1.45, the Kalman gain is:

[0025]

[0026] The updated status is estimated to be:

[0027]

[0028] The updated error covariance matrix is:

[0029]

[0030] Among them, R k represents the covariance matrix of the measurement noise; I represents the identity matrix.

[0031] As a further improvement of the present technical solution, in S2, the reading device is brought close to the acquisition device to trigger the Hall sensor on the acquisition device, thereby waking up the 2.4G wireless digital transmission module of the acquisition device and entering the half-duplex communication mode, including the following steps:

[0032] S2.1. When the reading device approaches the acquisition device, the strong magnet of the reading device triggers the Hall effect sensor on the acquisition device.

[0033] S2.2. After the Hall sensor is triggered, the logic circuit inside the acquisition device recognizes the trigger signal and wakes up the built-in 2.4G wireless digital transmission module according to the signal. The triggering process is optimized by introducing a Markov decision process combined with dynamic state transition probability estimation and a multi-objective reward function.

[0034] S2.3. After the wireless digital transmission module is awakened, the acquisition device enters the receiving state and is ready to receive a read instruction from the reading device. At this time, the reading device sends a read instruction to the acquisition device via the wireless digital transmission module.

[0035] S2.4. After the acquisition device successfully receives the read instruction, it establishes a half-duplex communication connection with the reading device.

[0036] As a further improvement of the present technical solution, in S2.2, a Markov decision process is introduced to combine dynamic state transition probability estimation and multi-objective reward function optimization triggering process, including the following steps:

[0037] S2.21. Define the state space S and the action space A;

[0038] S2.22. Based on historical data, estimate the probability of the system state changing after taking different actions a under different states s;

[0039] S2.23. Design a reward function to evaluate the quality of each state-action combination;

[0040] S2.24. Use the value iteration algorithm to find the triggering strategy that maximizes the long-term reward, choose whether to wake up in different states, and optimize the value iteration algorithm by combining dynamic state transition probability estimation and multi-objective reward function.

[0041] As a further improvement of this technical solution, in S2.24, the value iteration algorithm is:

[0042] V g+1 (s)=max a [R(s,a)+β∑ s′ D(s′|(s,a))V g (s′)];

[0043] Among them, V g+1 (s) represents the value function of state s at the g+1th iteration; R(s, a) represents the reward function; β represents the discount factor; D(s′|(s, a)) represents the state transition probability; s′ represents the next state; V g (s′) represents the gth iteration; g represents the number of iterations;

[0044] Combine historical transition probabilities and real-time environment parameters to calculate real-time transition probabilities and optimize state transition probabilities:

[0045]

[0046] Where D1(s′|(s, a, θ)) represents the dynamic state transition probability estimate after combining the historical transition probability and the real-time transition probability; Dhi represents the historical transition probability; Dre represents the real-time transition probability; θ represents the real-time environment parameter; δ represents the weight coefficient; NF represents the normalization factor;

[0047] In the reward function R(s, a), the three sub-goals of energy consumption, task success rate, and transmission delay are unified into a single comprehensive evaluation indicator through a dynamic weighting coefficient:

[0048] R1(s,a)=w1·R en +w2·R su +w3·R la ;

[0049] Among them, R1(s, a) represents the dynamically weighted multi-objective reward function; R en represents the energy consumption penalty term; R su Indicates the data transmission success rate; R la represents the delay penalty term; w1 represents the weight coefficient of energy consumption; w2 represents the weight coefficient of data transmission success rate; w3 represents the weight coefficient of delay;

[0050] In summary, combined with the dynamic state transition probability estimation and multi-objective reward function, the value iteration algorithm is optimized:

[0051] V1 g+1 (s)=max a [R1(s,a)+β∑ s′ D1(s′|(s,a,θ))V g (s′)];

[0052] Among them, V1 g+1 (s) represents the optimized value function.

[0053] As a further improvement of the present technical solution, in S3, the reading device sends an instruction through the wireless digital transmission module, specifying the time period and transmission mode for reading data, and the acquisition device extracts the corresponding data according to the instruction and transmits it to the reading device through the 2.4G wireless module in half-duplex communication mode, including the following steps:

[0054] S3.1. The reading device determines the data time period to be obtained from the acquisition device and selects the data transmission method;

[0055] S3.2. The reading device uses a 2.4G wireless digital transmission module to send instructions containing the data time period and data transmission method to the acquisition device;

[0056] S3.3. After the acquisition device receives the read instruction, the internal logic circuit parses the read instruction and identifies the required data time period and data transmission method;

[0057] S3.4. Based on the parsed information, the data collection device extracts data for the corresponding time period from the large-capacity storage device. If compressed transmission is specified, the data is compressed accordingly before transmission; if not, the raw data is directly prepared for transmission.

[0058] S3.5. After data analysis, the acquisition device begins transmitting data to the reading device via the 2.4G wireless module in half-duplex communication mode;

[0059] S3.6. After the transmission is completed, a notification of "data transmission completed" is sent to the reading device via the 2.4G wireless digital transmission module, notifying the reading device that the data transmission process has ended.

[0060] As a further improvement of the present technical solution, in S3.5, the acquisition device starts transmitting data to the reading device via the 2.4G wireless module in half-duplex communication mode, including the following steps:

[0061] S3.51. Before starting to send, the acquisition device checks whether the channel is idle;

[0062] S3.52. Establish a wireless communication link between the acquisition device and the reading device;

[0063] S3.53, dividing the data to be transmitted into packets of a fixed size m, and transmitting each packet in turn;

[0064] S3.54. After sending, the acquisition device switches to the receiving mode and waits for the ACK signal returned by the reading device.

[0065] As a further improvement of the present technical solution, in S4, after the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the dormant standby state, completing the entire data transmission process, including the following steps:

[0066] S4.1. After receiving the notification that "data transmission is completed", the reader verifies all received data packets through CRC to ensure they are complete and correct.

[0067] S4.2. After confirming that all data has been correctly received, the reading device sends a confirmation response to the acquisition device;

[0068] S4.3. After receiving the successful confirmation from the reading device, the wireless data transmission module of the acquisition device performs the cleanup work and switches back to the sleep mode to complete the entire data transmission process.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] The underwater wireless data transmission method of this deepwater riser monitoring system utilizes timestamp synchronization to prevent time drift and cross-bits in the collected data. Continuous data collection is maintained during wireless data acquisition, preventing interruptions or loss of status data due to communication issues. Two reading modes are available: continuous reading and reading and writing within a specified time period (specific storage area), facilitating efficient business data analysis. Data reading also supports compression, improving data transmission efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The figure is a flow chart of the overall method of the present invention. DETAILED DESCRIPTION

[0072] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0073] Example: See Figure 1 As shown, this embodiment provides an underwater wireless data transmission method for a deepwater riser monitoring system, comprising the following steps:

[0074] S1. Use the acquisition device to perform sampling according to the set sampling frequency, and temporarily store the sampled data in the SRAM (static random access memory) of the MCU (microcontroller unit);

[0075] In this embodiment, a sampling device is used to perform sampling according to a set sampling frequency, and the sampled data is temporarily stored in the SRAM (static random access memory) of the MCU (microcontroller unit) itself, including the following steps:

[0076] S1.1. Configure the acquisition device, including the MCU and sensors inside the acquisition device, set the sampling frequency, data format, and timestamp parameters. The acquisition device must have a built-in large-capacity memory of no less than 16GB.

[0077] S1.2. The MCU periodically triggers the sensor to start collecting data based on the set sampling frequency. In this embodiment, if the sampling frequency is set to 10 times per second, a sampling operation is triggered every 100 milliseconds. When the data collection device is sending data, it maintains continuous data collection through timed interrupts, and does not cause interruption or loss of status data due to data communication;

[0078] S1.3. Each time a sample is taken, the sensor obtains the status information of the riser (including acceleration, inclination, etc.) with a time stamp and sends the status information data to the MCU;

[0079] S1.4. After receiving the sensor's status data, the MCU temporarily stores it in the SRAM integrated within the MCU. SRAM is used here mainly because of its fast read and write speed, which can efficiently handle large amounts of data cache requirements in a short period of time. Timestamp synchronization storage is used to prevent data time offset and cross-bit, and a Kalman filter algorithm is introduced for time calibration.

[0080] The Kalman filter algorithm is an efficient, recursive mathematical method for estimating the state of a dynamic system. The Kalman filter algorithm consists of two main steps: prediction and update (also known as measurement update). The prediction step involves using a state transition model to predict the current state based on the system's previous state. This includes a predicted state vector (containing all information about the system's state) and the uncertainty of the state (typically represented by a covariance matrix). The update step (measurement update) involves combining new measurements with the predicted values ​​to update the state estimate. By comparing the difference between the predicted and actual measurements, the Kalman filter can adjust its internal state estimate to achieve more accurate results. The update step also involves calculating the Kalman gain, a key parameter that determines how to adjust the prediction based on new measurements. By combining the prediction and measurement update steps, the Kalman filter can provide more accurate time synchronization than using either prediction or measurement alone. This is particularly important for data acquisition that requires high-precision timestamps, as it can reduce time errors caused by clock skew. In underwater environments, sensor data can be affected by various environmental factors, generating noise. The Kalman filter algorithm can effectively process these noises by estimating the uncertainty in the process, thereby obtaining smoother and more accurate state information (such as acceleration, inclination, etc.);

[0081] The Kalman filter algorithm is introduced for time calibration, which includes the following steps:

[0082] S1.41. Define the state vector x k =[t k , b k ] T , where t k Indicates real time, b k represents the clock deviation, T represents the transposition operation, and the state transfer matrix F is defined k , used to express the change law of the state from time k to time k+1; k represents the time;

[0083] S1.42, based on the state estimate at the previous moment State transition matrix F k , predict the current state:

[0084] S1.43. Calculate the error covariance matrix of the predicted state Among them, Q k represents the covariance matrix of process noise; P k-1 represents the state estimation error covariance matrix at time k-1;

[0085] S1.44, get the current measurement value z k, using the observation matrix to relate the actual measurements to the state vector: k =H k ·x k +v k , where v k represents the measurement noise; H k Represents the observation matrix, which defines how to get the current state vector x from the system k Calculate the expected measurement value z k ;

[0086] S1.45, calculate the Kalman gain K k , and update the state estimate based on the measurement value and update the error covariance matrix P k .

[0087] The Kalman gain is:

[0088]

[0089] The updated status is estimated to be:

[0090]

[0091] The updated error covariance matrix is:

[0092]

[0093] Among them, R k represents the covariance matrix of the measurement noise; I represents the identity matrix;

[0094] S1.5. When the accumulated state information data in the SRAM reaches a preset threshold a or after a certain time interval, the accumulated state information data temporarily stored in the SRAM is transferred to the large-capacity memory for long-term storage. This step is to avoid frequent wake-up of the large-capacity memory and reduce its state transition delay (the large-capacity memory requires a long initialization time from wake-up to operation), thereby improving the efficiency of the entire system;

[0095] S1.6. The acquisition device continues to sample at the set sampling frequency and continuously adds the newly collected data to the SRAM until the conditions are met for batch transfer to the large-capacity storage.

[0096] S2. Utilize a reading device close to the acquisition device to trigger the Hall effect sensor on the acquisition device, waking up the acquisition device's 2.4G wireless digital transmission module to enter half-duplex communication mode. A Markov decision process (MDP) is then introduced to optimize the triggering process, combining dynamic state transition probability estimation with a multi-objective reward function.

[0097] In this embodiment, the reading device is brought close to the acquisition device to trigger the Hall sensor on the acquisition device, waking up the 2.4G wireless digital transmission module of the acquisition device and entering the half-duplex communication mode, including the following steps:

[0098] S2.1. Bring the reader device close to the acquisition device. This process requires the operator to carefully move the reader device close to the end face of the acquisition device. The reader device is equipped with a strong magnet. When the reader device approaches the acquisition device, the strong magnet triggers the Hall effect sensor on the acquisition device. The Hall effect sensor is a magnetic sensor based on the Hall effect that can detect changes in the magnetic field. In this scenario, it is used to sense the proximity of the reader device and serve as a signal to initiate data transmission.

[0099] S2.2. After the Hall sensor is triggered, the logic circuit inside the acquisition device recognizes the trigger signal and wakes up the built-in 2.4G wireless digital transmission module according to the signal. When data is not being transmitted, in order to save power, the wireless transmission module is usually in a dormant state and will only be activated after receiving a specific signal (Hall sensor trigger). If activation is not successful, it will continue to sleep. A Markov decision process (MDP) is introduced to combine dynamic state transition probability estimation and multi-objective reward function to optimize the triggering process and reduce the number of invalid wake-ups.

[0100] Among them, the Markov decision process (MDP) is introduced to combine the dynamic state transition probability estimation and multi-objective reward function optimization trigger process, which includes the following steps:

[0101] Markov Decision Process (MDP) is a mathematical framework used to simulate the process of making decisions in a sequential manner in a specific environment. An MDP consists of the following basic elements: state space, action space, transition probability matrix, reward function, and discount factor. By using MDP, the system can make more intelligent decisions based on the current state and historical data to determine whether the 2.4G wireless digital transmission module should be woken up according to the trigger signal of the Hall sensor. This decision is not based on simple thresholds or rules, but takes into account the combined influence of multiple factors. Invalid wake-ups will lead to unnecessary power consumption, which is an important consideration, especially in battery-powered devices. By optimizing the triggering process, these invalid wake-up events can be significantly reduced, extending the working time of the equipment and reducing maintenance costs. Since the number of unnecessary wake-ups is reduced, the hardware wear of the system will also be reduced accordingly, which helps to improve the reliability and life of the entire system. In addition, more precise control can also reduce the risk of misoperation.

[0102] S2.21. Define the state space S, which contains the current Hall sensor measurement values ​​and other environment variables:

[0103] s∈S, s=(B, Δt);

[0104] Where s represents the state, B represents the magnetic field strength currently measured by the Hall sensor, and Δt represents the time interval from the last valid wake-up to the current state.

[0105] Define the action space A:

[0106] a∈A={0,1};

[0107] Wherein, a represents an action, a=1: select "wake up" (activate the 2.4G wireless module), a=0: select "do not wake up" (remain in sleep state);

[0108] S2.22. Based on historical data (including historical sleep and wake states, Hall effect sensor trigger events, etc.), estimate the probability of the system state changing after taking different actions a in different states s. If the current state is sleep and a trigger signal is received, calculate the probability of waking up in this case; and vice versa.

[0109] S2.23. Design a reward function to evaluate the quality of each state-action combination;

[0110] The reward function R(s, a) is:

[0111]

[0112] Among them, rv represents the positive benefits brought by effective wake-up (timely data reception, energy saving, etc.); c represents the energy consumption cost of the wake-up process;

[0113] S2.24. Use the value iteration algorithm to find the triggering strategy that maximizes long-term rewards, choose whether to wake up in different states, and reduce the number of invalid wake-ups. Combined with dynamic state transition probability estimation and multi-objective reward function, the value iteration algorithm is optimized.

[0114] Furthermore, the value iteration algorithm is a dynamic programming method for solving Markov decision processes (MDPs), which aims to find the optimal strategy. It approximates the optimal value function by iteratively updating the value function of each state, and then determines the optimal strategy. Through the value iteration algorithm, the system can decide the optimal action (such as whether to wake up the wireless transmission module) based on the current state and historical data, thereby effectively reducing unnecessary energy consumption. This is especially important in battery-powered devices because it can directly affect the working life of the equipment. In complex and changeable deepwater riser monitoring systems, traditional static models are difficult to accurately describe the state transition rules. The value iteration algorithm, combined with dynamic state transition probability estimation and multi-objective reward functions, can better meet these challenges and provide more accurate and effective decision support for the system.

[0115] The value iteration algorithm is:

[0116] V g+1 (s)=max a [R(s,a)+β∑ s′ D(s′|(s,a))V g (s′)];

[0117] Among them, V g+1 (s) represents the value function of state s at the g+1th iteration. Here, “value” refers to the expected total reward that can be obtained by following a certain strategy starting from the current state; R(s, a) represents the reward function, which represents the reward obtained immediately after the system is in state s and performs action a; β represents the discount factor, which is used to calculate the importance of future rewards relative to immediate rewards. It is a value between 0 and 1 that determines the degree of influence of future rewards on the current decision; D(s′|(s, a)) represents the state transition probability, that is, the probability of reaching the next state s′ after performing action a from state s; s′ represents the next state; V g (s′) represents the estimated value of the value function of state s′ at the gth iteration, which is calculated by summing the values ​​of all possible next states s′, taking into account all possible outcomes and their probabilities; g represents the number of iterations;

[0118] In the deepwater riser monitoring system, the underwater environment's magnetic field strength, water flow velocity and other parameters have non-steady-state characteristics (sudden water flow disturbances, sudden changes in communication delays), and the traditional static transition probability model cannot accurately describe the state transition law; in the Markov decision process (MDP), the state transition probability is a key factor in determining how the system transitions from one state to another. Accurate state transition probability helps the system make more intelligent decisions based on the current state and historical data. By introducing real-time environmental parameters to optimize the state transition probability, the system can more accurately predict future states in complex and changing underwater environments, thereby providing more effective support in decisions such as whether to wake up the 2.4G wireless digital transmission module, reducing the number of invalid wake-ups, and improving the operating efficiency of the entire system; if the Hall sensor causes instantaneous magnetic field anomalies (B value mutations) due to water flow disturbances, real-time parameters can suppress invalid wake-ups triggered by noise to avoid power waste;

[0119] The real-time transition probability is calculated by combining the historical transition probability (identifying the periodic characteristics of the number of wake-up times in a specific time period) and the real-time environmental parameters (specifically, the real-time environmental parameters include: magnetic field strength, communication delay, and water flow speed) to optimize the state transition probability:

[0120]

[0121] Where D1(s′|(s, a, θ)) represents the dynamic state transition probability estimate after combining the historical transition probability and the real-time transition probability; Dhi represents the historical transition probability; Dre represents the real-time transition probability; θ represents the real-time environment parameter; δ represents the weight coefficient; NF represents the normalization factor;

[0122] Underwater wireless data transmission needs to meet the requirements of energy saving, reliability, and real-time performance simultaneously. However, there is a natural contradiction between the three. Frequently waking up the module can improve the mission success rate, but it will increase energy consumption and hardware loss. Delaying transmission may save energy, but it will lead to a decrease in data timeliness. Therefore, a multi-objective reward function is used to balance these objectives so that the entire system can achieve optimal performance within given resource constraints. When the device has sufficient power (w1 is small), it can prioritize ensuring the mission success rate (w2 is large) or reducing latency (w3 is large). When the power is insufficient (w1 is large), the system automatically reduces the weight of other objectives and prioritizes energy saving to extend the device's operating time.

[0123] In the reward function R(s, a), the three sub-goals of energy consumption, task success rate, and transmission delay are unified into a single comprehensive evaluation indicator through a dynamic weighting coefficient. The weight coefficient w1 is adaptively adjusted based on the remaining battery power of the device (w1 increases when the battery power is low, prioritizing energy conservation and avoiding monitoring interruptions due to battery exhaustion), achieving multi-objective collaborative optimization:

[0124] R1(s,a)=w1·R en +w2·R su +w3·R la ;

[0125] Among them, R1(s, a) represents the dynamically weighted multi-objective reward function; R en represents the energy consumption penalty term, R en =-k·Number of wake-up times; R su Indicates the data transmission success rate; R la represents the delay penalty term; w1 represents the weight coefficient of energy consumption; w2 represents the weight coefficient of data transmission success rate; w3 represents the weight coefficient of delay;

[0126] In summary, combined with the dynamic state transition probability estimation and multi-objective reward function, the value iteration algorithm is optimized:

[0127] V1 g+1 (s)=max a [R1(s,a)+β∑ s′ D1(s′|(s,a,θ))V g (s′)];

[0128] Among them, V1 g+1 (s) represents the optimized value function.

[0129] In this embodiment, in the deepwater riser monitoring system, the acquisition device is deployed on the seabed at a depth of 1000 meters. The wake-up strategy of its 2.4G wireless module needs to be optimized. The current environmental parameters are as follows:

[0130] Magnetic field strength B: 120mT (affected by water flow disturbances, the normal trigger threshold is 100mT); Time interval Δt: 2 hours since the last effective wake-up; Real-time environmental parameters θ: Water flow velocity 1.2m / s, communication delay 30ms; Device remaining battery: 25% (low battery state);

[0131] The state space and action space are defined as:

[0132] Current state: s = (B = 120 mT, Δt = 2 h), action space: A = {0 (not awakened), 1 (awakened)};

[0133] The dynamic state transition probability is calculated as:

[0134] The historical transfer probability (Dhi) is: Based on the log data statistics of the past 3 months: when B ≥ 100mT and Δt ≥ 1h, the probability of successful transmission after wakeup is 80% (that is, the probability of transitioning to the valid state s′ is 0.8); the probability of invalid wakeup (false triggering) is 20%;

[0135] Therefore, the historical transition probability is: Dhi(s′|s, a=1)=0.8 (effective awakening), 0.2 (ineffective awakening);

[0136] The real-time transition probability (Dre) is:

[0137] A water flow rate of 1.2 m / s may cause magnetic field noise, reducing trigger reliability. Through real-time sensor calibration: the effective wake-up probability is corrected to 0.8 × (1-water flow rate impact factor) = 0.8 × 0.7 = 0.56; the invalid wake-up probability increases to 0.44;

[0138] Therefore, the real-time transfer probability is: Dre(s′|s, a=1, θ)=0.56(valid), 0.44(invalid);

[0139] The final dynamic transition probability is:

[0140] Effective awakening probability = (0.6 × 0.8 + 0.4 × 0.56) / 0.704 ≈ 0.727;

[0141] Invalid wake-up probability = (0.6 × 0.2 + 0.4 × 0.44) / 0.704 ≈ 0.273;

[0142] Design a multi-objective reward function:

[0143] The sub-goal is defined as: Energy penalty term (R en ):R en = -k × wake-up times, k = 5; success rate reward R su ): If the transmission is successful after waking up, the reward is +20; if it fails, -10; delay penalty (R la ): If Δt ≥ 3h, penalty -15 (to avoid data backlog);

[0144] Dynamic weight adjustment (low battery priority energy saving): Remaining battery 25% → Weight coefficient: w1 = 0.7 (energy saving), w2 = 0.2 (success rate), w3 = 0.1 (delay);

[0145] Calculate the reward function: If action a = 1 (wake up): Successful transmission: R1 = 0.7 × (-5) + 0.2 × 20 + 0.1 × 0 = -3.5 + 4 = +0.5; Invalid wake up: R1 = 0.7 × (-5) + 0.2 × (-10) = -3.5 - 2 = -5.5;

[0146] If action a=0 (no wake-up): delay penalty: R1=0.7×0+0.1×(-15)=-1.5;

[0147] The value iteration process is:

[0148] Initialization value function: the initial value of all states V0(s′) = 0;

[0149] First iteration (g=0):

[0150] For state s = (120 mT, 2 h), calculate the expected value of the two actions:

[0151] Action a = 1 (arousal): expected value = 0.727 × 0.5 + 0.273 × (-5.5) + β × [0.727 × 0 + 0.273 × 0] = 0.3635 - 1.5015 ≈ -1.138;

[0152] Action a=0 (no wake-up):

[0153] Expected value = -1.5 + β × [future state value (if the penalty is higher after the delay is increased)] ≈ -1.5 (simplified calculation);

[0154] Update value: V1(s) = max(-1.138, -1.5) = -1.138;

[0155] Second iteration (g=1):

[0156] If the value of the future state (e.g., Δt = 3h) is updated to V1(s′) = -2 (delay penalty is increased);

[0157] Recalculate action a = 1: expected value = 0.727 × 0.5 + 0.273 × (-5.5) + 0.9 × [0.727 × (-2) + 0.273 × (-2)] = -1.138 + 0.9 × (-2) = -1.138 - 1.8 = -2.938;

[0158] The value of action a=0 is still -1.5;

[0159] Update value: V2(s) = max(-2.938, -1.5) = -1.5;

[0160] The decision results are:

[0161] After multiple iterations, the system selects a = 0 (no wake-up) in low-battery state to avoid invalid wake-up due to magnetic field disturbances and prioritize energy saving;

[0162] Traditional threshold method: B ≥ 100mT, that is, wake-up → this trigger is invalid, wasting 5 units of energy; optimized strategy: save 5 units of energy, delay penalty -1.5, but extend the device operation time by 2 hours;

[0163] By dynamically fusing environmental parameters with historical data, the value iteration algorithm chooses a conservative strategy in low-battery scenarios, balancing energy consumption and data timeliness. In this example, the system reduced one invalid wake-up, achieving significant energy-saving effects and verifying the algorithm's practicality in complex underwater environments.

[0164] S2.3. After the wireless digital transmission module is awakened, the acquisition device enters the receiving state and is ready to receive read instructions from the reading device. At this time, the reading device sends instructions to the acquisition device through the wireless digital transmission module, including the start and end times of the data to be read and the transmission method. There are two reading modes: continuous reading and reading and writing in a specified time period (specific storage area), facilitating efficient business data analysis.

[0165] S2.4. After the acquisition device successfully receives the read instruction, it establishes a half-duplex communication connection with the reading device to ensure that the subsequent data transmission can proceed smoothly.

[0166] S3. The reading device sends an instruction through the wireless digital transmission module, specifying the time period (continuous or specific interval) and transmission mode (compressed or uncompressed) for reading data. The acquisition device extracts the corresponding data from the large-capacity storage according to the instruction and transmits it to the reading device through the 2.4G wireless module in half-duplex communication mode.

[0167] In this embodiment, the reading device sends instructions via the wireless digital transmission module, specifying the time period (continuous or specific interval) and transmission mode (compressed or uncompressed) for data to be read. The acquisition device extracts the corresponding data from the large-capacity storage according to the instructions and transmits it to the reading device via the 2.4G wireless module in half-duplex communication mode. The steps include:

[0168] S3.1. The reading device determines the data time period to be obtained from the acquisition device (which can be a continuous time period or a specific interval) and selects the data transmission method (compressed or uncompressed);

[0169] S3.2. The reading device uses a 2.4G wireless digital transmission module to send a command containing the data time period and data transmission method to the acquisition device. The command will clearly indicate the start and end timestamps (for data requests for a specific time period) or the starting point of a continuous data stream (for continuous reading requests), and whether to request data compression for more efficient bandwidth utilization.

[0170] S3.3. After the acquisition device receives the read instruction, its internal logic circuit parses the instruction and identifies the required data time period and data transmission method (compressed or uncompressed). This step ensures that the acquisition device can accurately understand the requirements from the reading device.

[0171] S3.4. Based on the parsed information, the data collection device extracts data for the corresponding time period from the large-capacity storage device. If compressed transmission is specified, the data is compressed accordingly before transmission; if not, the raw data is directly prepared for transmission.

[0172] S3.5. After data analysis, the acquisition device begins transmitting data to the reader via the 2.4G wireless module in half-duplex communication mode. In this mode, the communication channel can only operate in one direction at a time, that is, either from the acquisition device to the reader or in the opposite direction, but not in both directions at the same time.

[0173] The acquisition device starts transmitting data to the reading device via a 2.4G wireless module in half-duplex communication mode, including the following steps:

[0174] S3.51. Before starting to send, the acquisition device detects whether the channel is idle and ensures that the current channel state is in the receiving or idle state to avoid conflicts;

[0175] S3.52. Establishing a wireless communication link between the acquisition device and the reading device, which involves a handshake protocol between the two devices to ensure that both parties are ready to exchange data;

[0176] S3.53, dividing the data to be transmitted into packets of a fixed size m, and transmitting each packet in turn;

[0177] S3.54. After sending, the acquisition device switches to receiving mode and waits for the ACK signal returned by the reading device. If an ACK is received within T, the data packet is considered to have been sent successfully and the next packet is sent. If no ACK is received, the data packet is retransmitted until the retransmission limit is reached or success is achieved.

[0178] S3.6. After the transmission is completed, a "data transmission completed" notification is sent to the reader via the 2.4G wireless digital transmission module, notifying the reader that the data transmission process has ended. This message contains some form of confirmation code or flag bit, indicating that all expected data packets have been successfully sent.

[0179] S4. After the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the sleep standby state, completing the entire data transmission process;

[0180] In this embodiment, after the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the sleep standby state, completing the entire data transmission process, including the following steps:

[0181] S4.1. After receiving the "data transmission completed" notification, the reader device verifies the integrity of all received data packets using CRC. If the reader device finds any missing data packets or detects data corruption (CRC verification failure), it requests the acquisition device to resend the missing or corrupted data. This retransmission mechanism helps ensure that the final received data is completely consistent with the original data.

[0182] S4.2. After confirming that all data has been correctly received, the reader sends a confirmation response to the acquisition device. This response is a brief message informing the acquisition device that it can safely shut down its wireless transmission module and return to sleep mode to save power.

[0183] S4.3. After receiving the successful confirmation from the reading device, the wireless data transmission module of the acquisition device performs cleanup work (such as disconnecting, releasing resources, etc.), switches back to sleep mode, and completes the entire data transmission process. In the sleep state, the module consumes very little power until it is triggered and awakened by the Hall sensor next time. The entire acquisition device returns to the initial standby state and waits for the next trigger event. During this period, in addition to the wireless transmission module, other components (such as MCU, sensors) may also remain active or enter their respective energy-saving modes according to design choices to further reduce energy consumption.

[0184] The basic principles, main features, and advantages of the present invention are shown and described above. It should be understood by those skilled in the art that the present invention is not limited to the above-described embodiments. The above-described embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention, and such changes and modifications fall within the scope of the invention claimed.

Claims

1. An underwater wireless data transmission method for a deepwater riser monitoring system, characterized in that: The following steps are involved: S1. Use the acquisition device to perform sampling according to the set sampling frequency, and temporarily store the sampled data in the SRAM of the MCU itself; S2. Use the reading device to approach the acquisition device, trigger the Hall sensor on the acquisition device, wake up the acquisition device's 2.4G wireless digital transmission module and enter half-duplex communication mode, and introduce a Markov decision process combined with dynamic state transition probability estimation and multi-objective reward function to optimize the triggering process; S3. The reading device sends an instruction through the wireless digital transmission module, specifying the time period and transmission method for the data to be read. The acquisition device extracts the corresponding data according to the instruction and transmits it to the reading device through the 2.4G wireless module in half-duplex communication mode; S4. After the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the sleep standby state, completing the entire data transmission process.

2. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 1, characterized in that: In S1, the acquisition device is used to perform sampling according to the set sampling frequency, and the sampled data is temporarily stored in the SRAM of the MCU itself, including the following steps: S1.

1. Configure the acquisition device, including the MCU and sensors inside the acquisition device, and set the sampling frequency, data format, and timestamp parameters; S1.

2. According to the set sampling frequency, the MCU triggers the sensor to start collecting data at regular intervals; S1.

3. Each time a sample is taken, the sensor obtains the status information of the riser with a timestamp and sends the status information data to the MCU; S1.

4. After receiving the status information data of the sensor, the MCU temporarily stores the status information data in the SRAM integrated inside the MCU, using the timestamp synchronization storage method and introducing the Kalman filter algorithm for time calibration; S1.

5. When the accumulated state information data in the SRAM reaches a preset threshold value a, the accumulated state information data temporarily stored in the SRAM is transferred to a large-capacity memory for long-term storage; S1.

6. The acquisition device continues to sample at the set sampling frequency and continuously adds the newly collected data to the SRAM until the conditions are met for batch transfer to the large-capacity storage.

3. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 2, characterized in that: In S1.4, a Kalman filter algorithm is introduced to perform time calibration, which includes the following steps: S1.

41. Define the state vector x k =[t k , b k ] T , where t k Indicates real time, b k represents the clock deviation, T represents the transposition operation, and the state transfer matrix F is defined k , used to express the change law of the state from time k to time k+1; k represents the time; S1.42, based on the state estimate at the previous moment State transition matrix F k , predict the current state: S1.

43. Calculate the error covariance matrix of the predicted state Among them, Q k represents the covariance matrix of process noise; P k-1 represents the state estimation error covariance matrix at time k-1; S1.44, get the current measurement value z k , using the observation matrix to relate the actual measurements to the state vector: k =H k ·x k +v k , where v k represents the measurement noise; H k represents the observation matrix; S1.45, calculate the Kalman gain K k , and update the state estimate based on the measurement value and update the error covariance matrix P k .

4. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 3, characterized in that: In S1.45, the Kalman gain is: The updated status is estimated to be: The updated error covariance matrix is: Among them, R k represents the covariance matrix of the measurement noise; I represents the identity matrix.

5. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 4, characterized in that: In S2, the reading device is brought close to the acquisition device to trigger the Hall sensor on the acquisition device, thereby waking up the 2.4G wireless digital transmission module of the acquisition device and entering the half-duplex communication mode, including the following steps: S2.

1. When the reading device approaches the acquisition device, the strong magnet of the reading device triggers the Hall effect sensor on the acquisition device. S2.

2. After the Hall sensor is triggered, the logic circuit inside the acquisition device recognizes the trigger signal and wakes up the built-in 2.4G wireless digital transmission module according to the signal. The triggering process is optimized by introducing a Markov decision process combined with dynamic state transition probability estimation and a multi-objective reward function. S2.

3. After the wireless digital transmission module is awakened, the acquisition device enters the receiving state and is ready to receive a read instruction from the reading device. At this time, the reading device sends a read instruction to the acquisition device via the wireless digital transmission module. S2.

4. After the acquisition device successfully receives the read instruction, it establishes a half-duplex communication connection with the reading device.

6. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 5, characterized in that: In S2.2, the Markov decision process is introduced to combine the dynamic state transition probability estimation and the multi-objective reward function optimization trigger process, including the following steps: S2.

21. Define the state space S and the action space A; S2.

22. Based on historical data, estimate the probability of the system state changing after taking different actions a under different states s; S2.

23. Design a reward function to evaluate the quality of each state-action combination; S2.

24. Use the value iteration algorithm to find the triggering strategy that maximizes the long-term reward, choose whether to wake up in different states, and optimize the value iteration algorithm by combining dynamic state transition probability estimation and multi-objective reward function.

7. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 6, characterized in that: In S2.24, the value iteration algorithm is: V g+1 (s)=max a [R(s,a)+β∑ s′ D(s′|(s,a))V g (s′)]; Among them, V g+1 (s) represents the value function of state s at the g+1th iteration; R(s, a) represents the reward function; β represents the discount factor; D(s′|(s, a)) represents the state transition probability; s′ represents the next state; V g (s′) represents the gth iteration; g represents the number of iterations; Combine historical transition probabilities and real-time environment parameters to calculate real-time transition probabilities and optimize state transition probabilities: Where D1(s′|(s, a, θ)) represents the dynamic state transition probability estimate after combining the historical transition probability and the real-time transition probability; Dhi represents the historical transition probability; Dre represents the real-time transition probability; θ represents the real-time environment parameter; δ represents the weight coefficient; NF represents the normalization factor; In the reward function R(s, a), the three sub-goals of energy consumption, task success rate, and transmission delay are unified into a single comprehensive evaluation indicator through a dynamic weighting coefficient: R1(s,a)=w1·R en +w2·R su +w3·R la ; Among them, R1(s, a) represents the dynamically weighted multi-objective reward function; R en represents the energy consumption penalty term; R su Indicates the data transmission success rate; R la represents the delay penalty term; w1 represents the weight coefficient of energy consumption; w2 represents the weight coefficient of data transmission success rate; w3 represents the weight coefficient of delay; In summary, combined with the dynamic state transition probability estimation and multi-objective reward function, the value iteration algorithm is optimized: V1 g+1 (s)=max a [R1(s,a)+β∑ s′ D1(s′|(s,a,θ))V g (s′)]; Among them, V1 g+1 (s) represents the optimized value function.

8. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 7, characterized in that: In S3, the reading device sends an instruction through the wireless digital transmission module, specifying the time period and transmission method for reading data. The acquisition device extracts the corresponding data according to the instruction and transmits it to the reading device through the 2.4G wireless module in half-duplex communication mode, including the following steps: S3.

1. The reading device determines the data time period to be obtained from the acquisition device and selects the data transmission method; S3.

2. The reading device uses a 2.4G wireless digital transmission module to send instructions containing the data time period and data transmission method to the acquisition device; S3.

3. After the acquisition device receives the read instruction, the internal logic circuit parses the read instruction and identifies the required data time period and data transmission method; S3.

4. Based on the parsed information, the data collection device extracts data for the corresponding time period from the large-capacity storage device. If compressed transmission is specified, the data is compressed accordingly before transmission; if not, the raw data is directly prepared for transmission. S3.

5. After data analysis, the acquisition device begins transmitting data to the reading device via the 2.4G wireless module in half-duplex communication mode; S3.

6. After the transmission is completed, a "data transmission completed" notification is sent to the reading device via the 2.4G wireless digital transmission module, notifying the reading device that the data transmission process has ended.

9. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 8, characterized in that: In S3.5, the acquisition device starts transmitting data to the reading device via the 2.4G wireless module in half-duplex communication mode, including the following steps: S3.

51. Before starting to send, the acquisition device checks whether the channel is idle; S3.

52. Establish a wireless communication link between the acquisition device and the reading device; S3.53, dividing the data to be transmitted into packets of a fixed size m, and transmitting each packet in turn; S3.

54. After sending, the acquisition device switches to the receiving mode and waits for the ACK signal returned by the reading device.

10. The underwater wireless data transmission method for a deepwater riser monitoring system according to claim 9, characterized in that: In S4, after the data transmission is completed, the reading device responds to the acquisition device, and the wireless data transmission module of the acquisition device re-enters the dormant standby state, completing the entire data transmission process, including the following steps: S4.

1. After receiving the "data transmission completed" notification, the reader verifies the integrity of all received data packets using CRC. S4.

2. After confirming that all data has been correctly received, the reading device sends a confirmation response to the acquisition device; S4.

3. After receiving the successful confirmation from the reading device, the wireless data transmission module of the acquisition device performs the cleanup work and switches back to the sleep mode to complete the entire data transmission process.

Images