Comprehensive energy power supply coastal ecosystem environment data monitoring device and method

Abstract

The invention discloses a coastal ecosystem environmental data monitoring device and method powered by comprehensive energy, and the device mainly comprises two parts, namely a comprehensive energy collection system and an environmental data monitoring device. The comprehensive energy collection system is composed of a wind energy collection device, a solar energy collection device and a tidal energy collection device, different energy inputs are comprehensively regulated and controlled through an energy management board, dynamic regulation and storage of energy are achieved, and stable operation of the whole device is guaranteed. The environment data acquisition device mainly comprises a carbon dioxide sensor, a temperature and humidity sensor, an illumination sensor and other sensing units, and is responsible for collecting related data of the environment where the device is located, and calibrating and predicting sensor detection data in the time dimension and the space dimension by using a deep learning algorithm and a machine learning algorithm. And finally, the data is uploaded to the cloud by using the Internet of Things module, so that the data can be analyzed timely and measures are taken to solve abnormities.

Description

Technical Field

[0001] The present invention belongs to the field of integrated energy collection technology and data monitoring, and specifically relates to a device and method for monitoring environmental data of a coastal ecosystem powered by integrated energy. Background Art

[0002] Soil respiration includes the production, consumption and emission of carbon dioxide in soil and is a key component of the global carbon cycle. Soil CO2 flux is a key indicator of soil organic matter decomposition and stability. As atmospheric carbon dioxide levels continue to rise due to human industrial activities, understanding and tracking soil carbon dynamics is becoming increasingly important for climate change modeling and ecosystem management.

[0003] Comprehensive energy harvesting technology can be divided into two parts: energy collection and energy storage. Since the device terminal may be in a harsh environment, multiple energy harvesters are needed to convert multiple energy sources (such as solar energy, wind energy, etc.) into electrical energy to ensure the stability of system operation.

[0004] Regarding the current monitoring of carbon flux in coastal ecosystems, due to the unique geographical location of the equipment, the use of traditional monitoring methods and monitoring equipment will bring huge challenges to the continuous monitoring of soil CO2 flux at the field level due to the high spatiotemporal variability and cost limitations of manual measurements. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention proposes a coastal ecosystem environmental data monitoring device and method powered by comprehensive energy, which eliminates a series of disadvantages brought about by manual monitoring, and is powered by environmental energy collection, eliminating the process of repeated battery replacement. At the same time, multi-energy collection also ensures the continuity of energy supply and the stability of the entire monitoring system. On this basis, a reasonable device structure setting combined with deep learning and machine learning algorithms ensures the accuracy of data collected by the system.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A coastal ecosystem environmental data monitoring device powered by integrated energy includes a wind energy collection device, a solar energy collection device and a tidal energy collection device. The wind energy collection device includes a wind energy collection blade and is located at the top of a connecting rod, adopts a vertical axis structure design, and the blade of the wind energy collection blade is designed to be spiral; the solar energy collection device is a solar energy collection panel, a solar energy collection panel is placed at the upper left end of the connecting rod, and an environmental data monitoring device is placed at the upper right end of the connecting rod. The environmental data monitoring device is fixed by an electric pulley, and the electric pulley is used to switch between working and idle states to lift or lower the environmental data monitoring device. The tidal energy collection device includes a tidal energy collection float, which is located at the bottom end of the connecting rod. When the water level rises, the tidal energy collection float moves up and down at the connecting rod. The tidal energy collection float is designed as a three-quarter circle with a gap that does not affect the up and down movement of the environmental data monitoring device. The tidal energy collection float contains a position sensor, and when tide is detected, the monitoring device is not allowed to be lowered.

[0008] As a further improvement of the structure of the present invention, the coastal ecosystem environmental data monitoring device powered by integrated energy is provided with an energy management board. The energy management board is designed with two parts: a matching circuit and a distribution circuit. The matching circuit is composed of a shift register and a switch array. The shift register is continuously input and output. The single-chip microcomputer indicates the state of each switch, outputs instructions through a pin string, and controls each switch through parallel pins. The single-chip microcomputer controls the energy input of each energy source to a specific energy storage capacitor to achieve dynamic regulation of energy.

[0009] As a further improvement of the structure of the present invention, the wind energy collecting blades are prepared and formed by three-dimensional printing, and the multi-directional wind energy collecting blades are prepared and formed by three-dimensional printing. The designed structure is conducive to collecting the strong vertical wind caused by the temperature difference between day and night at the seaside.

[0010] As a further improvement of the structure of the present invention, the environmental data monitoring device includes a carbon dioxide sensor, a temperature and humidity sensor, and a light sensor.

[0011] The monitoring method of the coastal ecosystem environmental data monitoring device powered by comprehensive energy according to the present invention comprises the following specific steps:

[0012] 1) For the collected sensor data, the gated recurrent unit (GRU) in the deep learning algorithm is used to capture the potential temporal patterns of carbon dioxide concentration, temperature and humidity, and light intensity, and to make temporal predictions of future carbon dioxide concentration, temperature and humidity, and light intensity. By weighted fusion of the predicted value and the measured value, the detection accuracy and stability can be improved;

[0013] 2) Using the Gaussian process regression (GPR) algorithm, based on the sensor data of the array arrangement at the known specific location, the detection data of all unobserved locations in the area are spatially predicted, thereby achieving the purpose of monitoring a large range of ecosystem environments with limited sensors;

[0014] The gated unit GRU used for temporal prediction contains two core gating mechanisms: update gate and reset gate. At time step t, the calculation formula of GRU is as follows:

[0015] Update gate: determines how much of the current hidden state comes from the previous hidden state and how much comes from the current input;

[0016] z t =σ(W z ·[h t-1 ,X t ]+b z )

[0017] Among them, W z and b z They are the weight matrix and bias term of the update gate respectively;

[0018] Remake gate: determines how to discard previous state information;

[0019] r t =σ(W r ·[h t-1 ,X t ]+b r )

[0020] Among them, W r and b r are the weight matrix and bias term of the reset gate, respectively, h t-1 is the hidden state of the previous moment, X t is the current input;

[0021] The optimizer uses Adam, and the loss function uses mean square error loss MSE, following the formula:

[0022]

[0023] Considering that there is an implicit relationship between carbon dioxide concentration, temperature and humidity, and light intensity, the input time series data also includes historical carbon dioxide concentration, temperature and humidity, and light intensity. The input of each time step is represented as a vector: X t =[CO2(t),Temp(t),Humidity(t),Light(t)] The target output is the predicted value at a certain moment in the future, for example, the data prediction after k moments:

[0024]

[0025] The vector x contains the four sensor information required for prediction;

[0026] Finally, the predicted value and the actual measured value are fused with a certain weight α to improve the prediction accuracy and system stability, following the formula:

[0027]

[0028] The Gaussian process regression algorithm used for spatial prediction performs non-parametric regression based on Bayesian inference and uses the radial basis kernel RBF as the kernel function;

[0029] Prior distribution of Gaussian process, first assume that the output data comes from Gaussian process, and the joint distribution follows the formula:

[0030]

[0031] where K(X,X) is the covariance matrix between known locations, σ 2 is the noise term, is the model error, and I is the identity matrix.

[0032] The covariance matrix uses the radial basis kernel RBF and follows the formula:

[0033]

[0034] For the new prediction point x * , infer its output value y * The joint distribution of and known data y. According to the properties of the Gaussian process, the joint distribution is:

[0035]

[0036] Among them, K(X,x * ) is the covariance vector between the known position and the predicted point, K(x * ,x * ) is the autocovariance of the forecast points;

[0037] According to the inference formula of Gaussian process, the posterior distribution is derived through conditional probability, that is, given the known data points X and y, the predicted point y is inferred * Distribution:

[0038]

[0039] Mean μ * : The most likely value of the prediction, following the formula:

[0040] μ * =K(x * ,X)(K(X,X)+σ 2 I) -1 y

[0041] variance The uncertainty of the forecast follows the formula:

[0042]

[0043] According to the above Gaussian process regression formula, the sensor detection values ​​with known coordinates are used to fit the constants in the formula, and Gaussian process regression models for different sensors are constructed. Finally, the coordinates of the unobserved points to be predicted are brought in, with the mean as the predicted value and the variance as the confidence indicator.

[0044] As a further improvement of the present invention, when it is necessary to monitor environmental data, the surrounding environment is monitored by controlling the electric pulley to place a monitoring device. At the same time, when monitoring is not needed, the electric pulley is controlled again to lift the monitoring device, and the single-chip microcomputer main control is used to cut off the energy distribution to the data monitoring device, thereby reducing unnecessary energy loss.

[0045] As a further improvement of the present invention, the environmental data monitoring device also has an Internet of Things module, which synchronously uploads the data to the cloud, allowing technicians to read relevant environmental parameters in real time and take timely measures for abnormal parameters.

[0046] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0047] 1. The multi-energy powered carbon flux monitoring device collects and utilizes the energy in the environment to provide power for the entire system, thus solving the energy limitations brought about by the device setting environment and eliminating a series of disadvantages brought about by manual monitoring.

[0048] 2. The environmental parameter monitoring device of the invention is fixed on the connecting rod through a pulley at the mechanical structure level, and can be switched between working and idle states, and the monitoring device can be raised or lowered, which is convenient for energy saving of the system and enhances the environmental protection of the device. When the tide is detected, the monitoring device is not allowed to be lowered, which avoids the damage of the monitoring device by the tide. At the environmental detection level, deep learning and machine learning related algorithms are used to predict and calibrate the environmental detection data in the time dimension and the space dimension, which improves the accuracy and stability of environmental detection.

[0049] 3. The energy management board designed by the invention solves the problem of energy imbalance caused by the constantly changing environment of the device terminal. It uses a single-chip microcomputer main control to achieve dynamic adjustment of the three energies, thereby enhancing the stability of the system.

[0050] 4. According to the geographical environment factors where the equipment is located, three different energy collection devices are set up, and the three devices are organically integrated through a reasonable mechanical structure. The three devices do not interfere with each other, and cooperate with each other to solve the energy problem of the device.

[0051] 5. The sensor modules configured in the monitoring device can be increased or decreased according to specific application scenarios. For example, (oxygen content sensors) can be added in certain environments, which facilitates customized preparation according to usage conditions and improves the applicability of the monitoring device. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 The figure is a mechanical structure diagram of the device of the present invention, wherein 1 is a wind energy collecting blade, 2 is an electric pulley, 3 is an environmental data monitoring device, 4 is a solar energy collecting panel, and 5 is a tidal energy collecting float;

[0053] Figure 2 Assigning management diagrams to energy management board capacitors;

[0054] Figure 3 It is a logic diagram of integrated energy management;

[0055] Figure 4 This is the circuit logic diagram of the environmental data monitoring device. DETAILED DESCRIPTION

[0056] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments:

[0057] Structural aspects such as Figure 1 As shown, the designed wind energy collection device is located at the top of the connecting rod, adopts a vertical axis structure design, and the blades are designed to be spiral, so that it has the function of collecting horizontal wind and vertical wind energy; the solar energy collection panel is placed on the upper left end of the connecting rod, and the environmental data monitoring device is placed on the upper right end. At the same time, the device can be switched between working and idle states through a pulley fixation, and the monitoring device can be raised or lowered. The float of the tidal energy collection device is located at the bottom of the connecting rod. When the water level rises, the float moves up and down at the connecting rod. At the same time, the float is designed to be a three-quarter circle. When floating up and down with the tide, the gap left does not affect the use of the environmental monitoring device. The float part contains a position sensor. When the tide is detected, the monitoring device is not allowed to be lowered.

[0058] The integrated energy harvesting system mainly consists of three parts: energy collection, energy storage, and energy distribution. The energy collection part is responsible for collecting other energy from the environment and converting it into electrical energy. Most system architectures use solar collectors as energy collection devices because solar energy is a very abundant energy source. On this basis, according to the particularity of the coastal environment, wind energy collection devices and tidal energy collection devices are also selected. The energy storage part is responsible for storing the electrical energy obtained from the energy harvester. Commonly used storage devices are rechargeable batteries and capacitors. The energy distribution part is responsible for distributing the energy stored in the energy storage device to the microcontroller and peripherals for use. Due to changes in the environment in which the monitoring device is located, traditional energy collection methods will lead to a series of problems in the efficiency of energy collection, so an energy management board is required. Capacitor distribution management such as Figure 2 As shown in Figure 2, the integrated energy management logic is as follows: Figure 3 As shown, the energy management board dynamically adjusts energy to maintain the stability of the entire system operation.

[0059] In terms of energy collection, multiple energy harvesters are used to collect multiple environmental energies respectively, and another super-large capacitor is used to store the energy collected by all energy harvesters. The energy output by multiple energy harvesting modules is input into the matching circuit through a booster, and then transmitted to the corresponding energy storage capacitor. Finally, the energy in the capacitor is transmitted to the corresponding device through a distribution circuit, which solves the energy loss caused by the different energy levels collected by different energy harvesters due to the constant changes in the external environment. At the same time, the microcontroller main control module is used to further regulate the energy distribution. The microcontroller runs two processes, one for matching and the other for distribution. The matching process is responsible for controlling the energy flow from multiple sources, that is, which capacitor each is input to. The distribution process is responsible for controlling the energy flow in multiple capacitors, that is, which devices the capacitors are allocated to.

[0060] Environmental data monitoring devices such as Figure 4 The sensors shown mainly include carbon dioxide sensors, temperature and humidity sensors, light sensors, etc. When it is necessary to monitor environmental data, the monitoring device is placed by controlling the electric pulley to monitor the surrounding environment, thereby improving the real-time nature of environmental monitoring data. At the same time, when monitoring is not required, the electric pulley is controlled to lift the monitoring device again, and the energy distribution to the data monitoring device is cut off by the single-chip main control, thereby reducing unnecessary energy loss. In addition, there is an Internet of Things module in the environmental data monitoring device, which can upload data to the cloud synchronously, so that technicians can read relevant environmental parameters in real time and take timely measures for abnormal parameters.

[0061] For the collected sensor data, on the one hand, the gated recurrent unit GRU in the deep learning algorithm is used to capture the potential temporal patterns of carbon dioxide concentration, temperature and humidity, and light intensity, and to make temporal predictions of future carbon dioxide concentration, temperature and humidity, and light intensity. By weighted fusion of predicted values ​​and measured values, the detection accuracy and stability are improved. On the other hand, the Gaussian process regression GPR algorithm is used to spatially predict the detection data of all unobserved locations in the area based on the sensor data of the array arrangement at known specific locations, thereby achieving the purpose of monitoring a large range of ecosystem environments with limited sensors.

[0062] The gated unit GRU used for temporal prediction contains two core gating mechanisms: update gate and reset gate. At time step t, the calculation formula of GRU is as follows:

[0063] Update gate: determines how much of the current hidden state comes from the previous hidden state and how much comes from the current input.

[0064] z t =σ(W z ·[h t-1 ,X t ]+b z )

[0065] Among them, W z and b z are the weight matrix and bias term of the update gate respectively.

[0066] Remake Gate: Determines how to discard previous state information.

[0067] r t =σ(W r ·[h t-1 ,X t ]+b r )

[0068] Among them, W r and b r are the weight matrix and bias term of the reset gate, respectively, h t-1 is the hidden state of the previous moment, X t is the current input.

[0069] The optimizer uses Adam, and the loss function uses mean square error loss MSE, following the formula:

[0070]

[0071] Considering the implicit relationship between carbon dioxide concentration, temperature and humidity, and light intensity, the input time series data also includes historical carbon dioxide concentration, temperature and humidity, and light intensity. The input of each time step can be expressed as a vector:

[0072] X t =[CO2(t),Temp(t),Humidity(t),Light(t)] The target output is the predicted value at a certain moment in the future, for example, the data prediction after k moments:

[0073]

[0074] The vector x contains the four sensor information that need to be predicted.

[0075] Finally, the predicted value and the actual measured value will be fused with a certain weight α to improve the prediction accuracy and system stability. Follow the formula:

[0076]

[0077] The Gaussian process regression algorithm used for spatial prediction performs non-parametric regression based on Bayesian inference and uses the radial basis kernel RBF as the kernel function.

[0078] Prior distribution of Gaussian process, first assume that the output data comes from Gaussian process, and the joint distribution follows the formula:

[0079]

[0080] where K(X,X) is the covariance matrix between known locations, σ 2 is the noise term (usually the model error), and I is the identity matrix.

[0081] The covariance matrix uses the radial basis kernel RBF and follows the formula:

[0082]

[0083] For the new prediction point x * , infer its output value y * The joint distribution of and known data y. According to the properties of the Gaussian process, the joint distribution is:

[0084]

[0085] Among them, K(X,x * ) is the covariance vector between the known position and the predicted point, K(x * ,x * ) is the autocovariance of the forecast points.

[0086] According to the inference formula of Gaussian process, the posterior distribution can be derived through conditional probability, that is, given the known data points X and y, the predicted point y can be inferred * Distribution:

[0087]

[0088] Mean μ * : The most likely value of the prediction, following the formula:

[0089] μ * =K(x * ,X)(K(X,X)+σ 2 I) -1 y

[0090] variance The uncertainty of the forecast follows the formula:

[0091]

[0092] According to the above Gaussian process regression formula, the sensor detection values ​​with known coordinates are used to fit the constants in the formula, and Gaussian process regression models for different sensors are constructed. Finally, the coordinates of the unobserved points to be predicted are brought in, with the mean as the predicted value and the variance as the confidence indicator.

[0093] Working principle: According to the geographical environment characteristics of the monitoring device, select the corresponding multi-energy collection device such as (solar energy, wind energy, tidal energy, etc.). The three processes of energy collection, energy storage and energy distribution are intelligently adjusted by the single-chip master control to maintain the stable operation of the system. Through the reasonable mechanical structure design, the normal operation of each part of the entire device is guaranteed. At the same time, the switching of the use and idle state of the monitoring device is designed, which saves energy and avoids the problem of damage to the monitoring equipment during high tide.

[0094] The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.

Claims

1. A coastal ecosystem environmental data monitoring device powered by integrated energy, including a wind energy collection device, a solar energy collection device and a tidal energy collection device, characterized in that: The wind energy collection device comprises a wind energy collection blade (1) which is located at the top of the connecting rod and adopts a vertical axis structure design. The blade of the wind energy collection blade (1) is designed to be spiral. The solar energy collection device is a solar energy collection panel (4). The solar energy collection panel (4) is placed at the upper left end of the connecting rod. The environmental data monitoring device (3) is placed at the upper right end of the connecting rod. The environmental data monitoring device (3) is fixed by an electric pulley (2) and is switched between working and idle states by the electric pulley (2). , lift or lower the environmental data monitoring device (3), the tidal energy collection device includes a tidal energy collection float (5), the tidal energy collection float (5) is located at the bottom end of the connecting rod, when the water level rises, the tidal energy collection float (5) moves up and down at the connecting rod, the tidal energy collection float (5) is designed as a three-quarter circle and has a gap that does not affect the up and down movement of the environmental data monitoring device (3), the tidal energy collection float (5) contains a position sensor, when the tide is detected, the monitoring device is not allowed to be lowered.

2. The device and method for monitoring coastal ecosystem environmental data with integrated energy power supply according to claim 1, characterized in that: The coastal ecosystem environmental data monitoring device powered by integrated energy is provided with an energy management board. The energy management board is designed with two parts: a matching circuit and a distribution circuit. The matching circuit is composed of a shift register and a switch array. The shift register is continuously input and output. The single-chip microcomputer indicates the state of each switch, outputs instructions through a pin string, and controls each switch through parallel pins. The single-chip microcomputer controls the energy input of each energy source to a specific energy storage capacitor to achieve dynamic regulation of energy.

3. The device and method for monitoring coastal ecosystem environmental data with integrated energy supply according to claim 1, characterized in that: The wind energy collecting blade (1) is formed by three-dimensional printing.

4. The device and method for monitoring coastal ecosystem environmental data with integrated energy supply according to claim 1, characterized in that: The environmental data monitoring device (3) comprises a carbon dioxide sensor, a temperature and humidity sensor, and a light sensor.

5. The monitoring method of the coastal ecosystem environmental data monitoring device powered by comprehensive energy according to claims 1-4 is characterized in that: The specific steps are as follows: 1) For the collected sensor data, the gated recurrent unit (GRU) in the deep learning algorithm is used to capture the potential temporal patterns of carbon dioxide concentration, temperature and humidity, and light intensity, and to make temporal predictions of future carbon dioxide concentration, temperature and humidity, and light intensity. By weighted fusion of the predicted value and the measured value, the detection accuracy and stability can be improved; 2) Using the Gaussian process regression (GPR) algorithm, based on the sensor data of the array arrangement at the known specific location, the detection data of all unobserved locations in the area are spatially predicted, thereby achieving the purpose of monitoring a large range of ecosystem environments with limited sensors; The gated unit GRU used for temporal prediction contains two core gating mechanisms: update gate and reset gate. At time step t, the calculation formula of GRU is as follows: Update gate: determines how much of the current hidden state comes from the previous hidden state and how much comes from the current input; z t =σ(W z ·[h t-1 ,X t ]+b z ) Among them, W z and b z They are the weight matrix and bias term of the update gate respectively; Remake gate: determines how to discard previous state information; r t =σ(W r ·[h t-1 ,X t ]+b r ) Among them, W r and b r are the weight matrix and bias term of the reset gate, respectively, h t-1 is the hidden state of the previous moment, X t is the current input; The optimizer uses Adam, and the loss function uses mean square error loss MSE, following the formula: Considering that there is an implicit relationship between carbon dioxide concentration, temperature and humidity, and light intensity, the input time series data also includes historical carbon dioxide concentration, temperature and humidity, and light intensity. The input of each time step is represented as a vector: X t =[CO2(t),Temp(t),Humidity(t),Light(t)] The target output is the predicted value at a certain moment in the future, for example, the data prediction after k moments: The vector x contains the four sensor information required for prediction; Finally, the predicted value and the actual measured value are fused with a certain weight α to improve the prediction accuracy and system stability, following the formula: The Gaussian process regression algorithm used for spatial prediction performs non-parametric regression based on Bayesian inference and uses the radial basis kernel RBF as the kernel function; Prior distribution of Gaussian process, first assume that the output data comes from Gaussian process, and the joint distribution follows the formula: where K(X,X) is the covariance matrix between known locations, σ 2 is the noise term, is the model error, and I is the identity matrix. The covariance matrix uses the radial basis kernel RBF and follows the formula: For the new prediction point x * , infer its output value y * The joint distribution of and known data y. According to the properties of the Gaussian process, the joint distribution is: Among them, K(X,x * ) is the covariance vector between the known position and the predicted point, K(x * ,x * ) is the autocovariance of the forecast points; According to the inference formula of Gaussian process, the posterior distribution is derived through conditional probability, that is, given the known data points X and y, the predicted point y is inferred * Distribution: Mean μ * : The most likely value of the prediction, following the formula: μ * =K(x * ,X)(K(X,X)+σ 2 I) -1 y variance The uncertainty of the forecast follows the formula: According to the above Gaussian process regression formula, the sensor detection values ​​with known coordinates are used to fit the constants in the formula, and Gaussian process regression models for different sensors are constructed. Finally, the coordinates of the unobserved points to be predicted are brought in, with the mean as the predicted value and the variance as the confidence indicator.

6. The monitoring method of the coastal ecosystem environmental data monitoring device powered by comprehensive energy according to claim 5 is characterized in that: When it is necessary to monitor environmental data, the monitoring device is placed to monitor the surrounding environment by controlling the electric pulley. When monitoring is not needed, the electric pulley is controlled to lift the monitoring device again, and the single-chip main control is used to cut off the energy distribution to the data monitoring device, thereby reducing unnecessary energy loss.

7. The monitoring method of the coastal ecosystem environmental data monitoring device powered by comprehensive energy according to claim 5 is characterized in that: The environmental data monitoring device (3) also has an Internet of Things module, which uploads data synchronously to the cloud, allowing technicians to read relevant environmental parameters in real time and take timely measures for abnormal parameters.

Images