Method for measuring conveying flow of multi-metal nodule in wide temperature range

Abstract

The invention relates to the technical field of deep-sea mining, and discloses a wide-temperature-range multi-metal nodule conveying flow measuring method. Coil temperature, ore temperature, voltage data of a sensor and detection data of ore flow related to metal nodule flow detection are collected based on triple circulation, characteristic variables and a correction model are constructed based on the Bloch's law, and a training set and a verification set are obtained according to the characteristic variables, the correction model and the detection data. Training a correction model based on the training set and the verification set; and the coil temperature of the detection coil, the ore temperature and the voltage data of the sensor in the actual mining process are obtained and input into the correction model, the corresponding ore flow is obtained, and measurement of the conveying flow of the multi-metal nodule is completed. The problem that the multi-metal nodule ore flow cannot be accurately measured in a wide temperature change range in existing deep sea mining is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of deep - sea mining, and particularly to a method for measuring the conveying flow rate of polymetallic nodules within a wide temperature range. Background Art

[0002] During the deep - sea mining process, after polymetallic nodule ores are collected by a mining vehicle, they are transported to a water - surface support platform through a lifting pipeline. Due to the requirements of evaluating mining efficiency and dynamically adjusting the traveling trajectory of the mining vehicle, it is necessary to measure in real - time the amount of nodule ore passing through the conveying pipeline per unit time (also known as flow rate), so as to form closed - loop feedback data. The polymetallic nodule flow detection sensor consists of a detection coil and a processing circuit. When the metal nodule ore passes through the conveying pipeline surrounded by the detection coil, it causes a change in magnetic resistance, and this change is processed by the processing circuit of the sensor into a voltage signal. The deep - sea mining control system obtains the magnitude of the flow rate of polymetallic nodules according to the numerical value of the voltage signal and the given mathematical expression of the sensor.

[0003] The above - mentioned measurement principle is based on the assumption that the change in magnetic resistance is only caused by the amount of polymetallic nodules passing through the pipeline per unit time, without considering other factors. The shortcoming of this assumption is that it does not take into account that the magnetic resistance of a magnetically permeable material is related to temperature changes, and the magnetic resistance changes of the coil itself and the polymetallic nodules with temperature will also be processed by the processing circuit of the sensor into voltage signals, forming false flow data, thus causing measurement errors and further affecting the decision - making of the deep - sea mining control system. Summary of the Invention

[0004] The present invention provides a method for measuring the conveying flow rate of polymetallic nodules within a wide temperature range to solve the problem that existing deep - sea mining cannot accurately measure the flow rate of polymetallic nodule ores within a relatively wide temperature change range.

[0005] To achieve the above object, the present invention is realized through the following technical solutions: In a first aspect, the present invention provides a method for measuring the conveying flow rate of polymetallic nodules within a wide temperature range, including the following steps: Step 1: Based on a triple - loop to collect coil temperature, ore temperature, voltage data of the sensor, and detection data of ore flow rate regarding the detection of metal nodule flow rate, construct characteristic variables and a calibration model based on Bloch's law, and obtain a training set and a validation set according to the characteristic variables, calibration model, and combined detection data, and train the calibration model based on the training set and the validation set; Step 2: Obtain the coil temperature, ore temperature of the detection coil, and voltage data of the sensor during the actual mining process, and input them into the calibration model to obtain the corresponding ore flow rate, thereby completing the measurement of the conveying flow rate of polymetallic nodules.

[0006] Further, the acquisition of coil temperature, ore temperature, voltage data of the sensor, and detection data of ore flow rate regarding metal nodule flow rate detection based on triple loops includes: setting corresponding loop initial values for coil temperature, ore temperature, and ore flow rate, constructing a nested loop process based on the loop initial values, executing the nested loop process, and recording the data of coil temperature, ore temperature, ore flow rate, and voltage data of the sensor in all loops as detection data.

[0007] Further, the construction of the nested loop process based on the loop initial values includes: calculating the number of loops of the coil temperature loop, ore temperature loop, and ore flow rate loop respectively based on the loop initial values, and constructing a nested loop process based on the number of loops.

[0008] Through the above operations, the detection data collected based on triple loops comprehensively cover all data combinations, can more intuitively reflect the structural relationship of the data, and can independently control the parameters and conditions of each loop dimension, and can more effectively generate a dedicated data set for detecting ore flow rate at wide temperatures.

[0009] Further, the loop initial values include: initial coil temperature, final coil temperature, coil temperature increment step, initial ore temperature, final ore temperature, ore temperature increment step, initial ore flow rate, final ore flow rate, and ore flow rate increment step; The number of loops of the coil temperature loop is calculated by the following formula: ; where, represents the number of loops of the coil temperature loop, represents the final coil temperature, represents the initial coil temperature, represents the coil temperature increment step; The number of loops of the ore temperature loop is calculated by the following formula: ; where, represents the number of loops of the ore temperature loop, represents the final ore temperature, represents the initial ore temperature, represents the ore temperature increment step; The number of loops of the ore flow rate loop is calculated by the following formula: ; where, represents the number of loops of the ore flow rate loop, represents the final ore flow rate, represents the initial ore flow rate, Indicates the growth step of ore flow rate.

[0010] Furthermore, the construction of the loop nesting process executes the loop nesting process in the order of ore flow rate loop, coil temperature loop, and ore temperature loop in combination with the loop rules; The loop rules include: execute 1 coil temperature loop every k times of ore temperature loop execution, execute 1 ore flow rate loop every n times of coil temperature loop execution, execute a total of m times of ore flow rate loop, and adjust the coil temperature, ore temperature, and ore flow rate according to the corresponding coil temperature growth step, ore temperature growth step, and ore flow rate growth step respectively each time the ore temperature loop, coil temperature loop, and ore flow rate loop are executed; After executing one coil temperature loop, adjust the current ore temperature to the initial ore temperature. After executing one ore flow rate loop, adjust the current coil temperature to the initial coil temperature and the current ore temperature to the initial ore temperature.

[0011] Furthermore, the construction of the characteristic variables and calibration model based on Bloch's law includes: based on the linear relationship between the magnetoresistance and the 1.5th power of temperature in Bloch's law, the coil temperature and ore temperature have a common non-linear influence on the measurement output of the sensor, the magnetoresistance and flow rate are linear relationships, and they are linearly related to the square of the output voltage of the sensor to construct the characteristic variables and calibration model; The temperature Bloch's law describes the relationship between magnetoresistance and temperature in magnetic materials. Selecting the linear relationship between magnetoresistance and the 1.5th power of temperature conforms to the low to medium temperature range of deep-sea mining. Through the above operations, this non-linear change can be better fitted. Compared with a simple linear term, when the temperature changes greatly, it can more accurately reflect the influence of temperature on ore flow rate; The characteristic variables are expressed by the following formula: ; Where, represents the coil temperature, represents the ore temperature, represents the common influence factor, represents the voltage signal output by the sensor; The calibration model is expressed by the following formula: ; Where, represents the ore flow rate output by the calibration model, is the model coefficient.

[0012] Furthermore, the obtaining of the training set and validation set according to the characteristic variables, calibration model combined with the detection data includes: performing combined calculations based on the characteristic variables, calibration model combined with the detection data to obtain a structured data set With , after normalizing the structured data set and , a training set and a validation set are obtained.

[0013] Furthermore, after obtaining the training set and the validation set through normalization processing, a calibration model is trained by means of 5-fold cross-validation and support vector regression.

[0014] Furthermore, the normalization processing is a standardization method.

[0015] Furthermore, the hyperparameters of the calibration model are obtained through the GridSearchCV function during the training process.

[0016] Beneficial effects: A method for measuring the conveying flow rate of polymetallic nodules in a wide temperature range provided by the present invention makes significant adjustments to the use of data from sensors for detecting the conveying flow rate of polymetallic nodules, overcomes the problem that the output data of the original device has large errors when the temperature of the ore in the conveying pipeline or the sensor's own coil changes, and can accurately detect the flow rate of polymetallic nodule ore within a relatively wide temperature change range, providing reliable data and an effective detection model for the control system of deep-sea mining. Description of the drawings

[0017] Figure 1 is a flowchart of a method for measuring the conveying flow rate of polymetallic nodules in a wide temperature range according to an embodiment of the present invention; Figure 2 is a flowchart of a loop-nested process in an embodiment of the present invention; Figure 3 is a schematic diagram of the training and use of a calibration model in an embodiment of the present invention. Specific implementation manners

[0018] The technical solutions of the present invention will be described clearly and completely below. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without any creative efforts shall fall within the protection scope of the present invention.

[0019] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention pertains. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "a" or "one" do not denote a quantity limitation, but mean that there is at least one. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0020] Please refer to Figure 1 , the embodiment of the present application provides a method for measuring the conveying flow rate of polymetallic nodules in a wide temperature range, including the following steps: Step 1: Based on a triple loop, collect the coil temperature, ore temperature, voltage data of the sensor, and detection data of the ore flow rate for the detection of the metal nodule flow rate. Construct a characteristic variable and a calibration model based on Bloch's law, and obtain a training set and a validation set according to the characteristic variable, the calibration model, and the detection data. Train the calibration model based on the training set and the validation set; Please refer to Figure 2 , specifically, set corresponding loop initial values for the coil temperature, ore temperature, and ore flow rate. Based on the loop initial values, construct a nested loop process, execute the nested loop process, and record the data of the coil temperature, ore temperature, ore flow rate, and voltage data of the sensor in all loops as the detection data.

[0021] The loop initial values include: the initial coil temperature, the final coil temperature, the coil temperature increment step, the initial ore temperature, the final ore temperature, the ore temperature increment step, the initial ore flow rate, the final ore flow rate, and the ore flow rate increment step; In this embodiment, the initial coil temperature is set to 20 °C, the final coil temperature is set to 40 °C, the coil temperature increment step is 1 °C, the initial ore temperature is 20 °C, the final ore temperature is 60 °C, the ore temperature increment step is 2 °C, the initial ore flow rate is 600 kg / min, the final ore flow rate is 1200 kg / min, and the ore flow rate increment step is 50 kg / min.

[0022] Based on the loop initial values, calculate the number of loops for the coil temperature loop, the ore temperature loop, and the ore flow rate loop respectively, and construct a nested loop process based on the number of loops.

[0023] Among them, the nested loop process is executed in the order of the ore flow rate loop, the coil temperature loop, and the ore temperature loop in combination with the loop rules; The cycling rules include: performing one coil temperature cycle for every k ore temperature cycles, performing one ore flow cycle for every n coil temperature cycles, and performing a total of m ore flow cycles. Each time an ore temperature cycle, a coil temperature cycle, and an ore flow cycle are performed, the coil temperature, the ore temperature, and the ore flow are adjusted according to the corresponding coil temperature increase step, ore temperature increase step, and ore flow increase step respectively. After performing one coil temperature cycle, the current ore temperature is adjusted to the initial ore temperature. After performing one ore flow cycle, the current coil temperature is adjusted to the initial coil temperature, and the current ore temperature is adjusted to the initial ore temperature.

[0024] The number of cycles of the ore temperature cycle is calculated by the following formula: ; where represents the number of cycles of the ore temperature cycle, represents the final ore temperature, represents the initial ore temperature, represents the ore temperature increase step; The number of cycles of the coil temperature cycle is calculated by the following formula: ; where represents the number of cycles of the coil temperature cycle, represents the final coil temperature, represents the initial coil temperature, represents the coil temperature increase step; The number of cycles of the ore flow cycle is calculated by the following formula: ; where represents the number of cycles of the ore flow cycle, represents the final ore flow, represents the initial ore flow, represents the ore flow increase step.

[0025] Then the constructed nested loop process includes: Maintain the initial ore flow in the conveying pipeline at 600 kg / min, the initial coil temperature of the coil at 20 °C, then set the initial ore temperature of the ore at 20 °C, and then perform 20 data acquisition operations. After each data acquisition, each time the temperature of the ore is adjusted according to the ore temperature increase step until the ore temperature reaches the final ore temperature of 60 °C. During this process, always maintain the initial ore flow at 600 kg / min and the initial coil temperature of the coil at 20 °C.

[0026] After 20 data acquisitions, the initial coil temperature is adjusted according to the coil temperature increase step, that is, the coil temperature is adjusted from 20°C to 21°C, and the ore temperature at this time is adjusted from 60°C to the initial ore temperature of 20°C. Then, 20 data acquisition operations are performed again. Repeat the above process, that is, for each coil temperature adjustment, 20 ore temperature adjustments are made until the coil temperature ranges from the initial coil temperature to the final coil temperature. Each time, the coil temperature is adjusted according to the coil temperature increase step. During this process, the ore flow rate is maintained at 600 kg / min all the time. Then, a total of 400 consecutive data are collected. The collected data include ore temperature, coil temperature, ore flow rate, and voltage signal data output by the sensor.

[0027] After collecting 400 data, the ore flow rate is adjusted from 600 kg / min to 650 kg / min according to the ore flow rate increase step. Then, the coil temperature is adjusted from 40°C to 20°C, and the ore temperature is adjusted from 60°C to 20°C. Then, for each adjustment of the ore flow rate, 20 coil temperature adjustments are made to the coil temperature. For each adjustment of the coil temperature, 20 ore temperature adjustments are made. Each time the coil temperature is adjusted, the ore temperature is adjusted to 20°C. Each time the ore flow rate is adjusted, the ore temperature is adjusted to 20°C, and the coil temperature is adjusted to 20°C. Then, a total of 4800 data can be collected (including the 400 consecutive data collected previously).

[0028] Constructing the characteristic variables and calibration model based on Bloch's law includes: based on the linear relationship between the magnetoresistance and the 1.5th power of temperature in Bloch's law, the coil temperature and the ore temperature have a common non-linear influence on the measurement output of the sensor. The magnetoresistance and the flow rate are in a linear relationship and are linearly related to the square of the output voltage of the sensor to construct the characteristic variables and calibration model; The characteristic variables are expressed by the following formula: ; where, represents the coil temperature, represents the ore temperature, represents the common influence factor, represents the voltage signal output by the sensor; The calibration model is expressed by the following formula: ; where, represents the ore flow rate output by the calibration model, is the model coefficient.

[0029] Bloch's law describes the relationship between magnetoresistance and temperature in magnetic materials. The linear relationship between magnetoresistance and the 1.5th power of temperature is selected to fit the low to medium temperature range of deep-sea mining, which can better fit this non-linear variation. Magnetoresistance usually decreases as temperature increases. is the Curie temperature. In the multi-metal nodule flow detection sensor, the magnetoresistance is proportional to the flow rate of the ore. It can be seen that there is a linear relationship between the flow rate of the ore and the 1.5th power of temperature. Selecting as the characteristic variable can better fit this non-linear variation. Compared with the simple linear terms and and , especially when the temperature changes greatly, it can more accurately reflect the influence of temperature on the ore flow rate; In addition, the coil temperature and the ore temperature do not independently affect the ore flow rate. There is an interaction between them. When both change simultaneously, the thermal effects can be superimposed, resulting in a more significant change in the ore flow rate than when considering them separately. Therefore, the product term is added to capture this coupling effect; As for and as the first-order terms, although the influence of temperature on the ore flow rate may be non-linear as a whole, the first-order terms are the basic part. Adding the first-order terms ensures that the model includes temperature effects from simple to complex, avoiding missing the basic linear contributions. Based on the above, a calibration model for ore flow detection can be constructed more accurately; Please refer to Figure 3 . According to the characteristic variables and the calibration model, combined with the detection data, the training set and the validation set are obtained, including: based on the characteristic variables, the calibration model, and combined with the detection data for combined calculation to obtain a structured data set and . After normalizing the structured data set and , the training set and the validation set are obtained.

[0030] After obtaining the training set and the validation set through normalization, the calibration model is also trained by 5-fold cross-validation and support vector regression. The normalization is the standardization method.

[0031] Specifically, with the characteristic variables as the input matrix and the model output as the output matrix, 4800 pieces of collected data are input in sequence for calculation and processing to generate the structured data set required for training and , for 4800 pieces of data, every 20 pieces in sequence are taken as a sampling unit, and 16 pieces are sampled as training data and 4 pieces as validation data. After one division, the dataset is decomposed into a training set consisting of 3840 pieces of data and a validation set consisting of 960 pieces of data. This division operation is carried out 5 times, and a total of 5 training sets and 5 validation sets are obtained; Data pipeline processing, including data standardization and SVR model specification: First, unify the scale of feature data. Use the Z-score standardization method to standardize the training data and transform it into a standard normal distribution with a mean of 0 and a standard deviation of 1 to prevent the adverse effects of different data scales on the performance of SVR. In the Python code, it is implemented by calling the StandardScaler function; Since the feature input of the calibration model is non-linear, the Gaussian radial basis kernel (RBF) is selected to map the data to a high-dimensional space, and then linear regression is realized in the high-dimensional space. Use Pipeline to create a pipeline containing StandardScaler and SVR; Use the GridSearchCV function to specify hyperparameter search and 5-fold cross-validation. The object constructed by this function uses the fit function to fit the training set to find the best hyperparameter combination, and then retrain the model on the entire training set using the best parameters to obtain the final calibration model; Among them, the best hyperparameter combination means that in the 5-fold cross-validation, for the calibration model with this set of hyperparameters, it alternately uses 4 parts for training and 1 part for validation, calculates the accuracy and F1 score 5 times, and takes their average. The set of hyperparameters with a higher average is the best hyperparameter combination.

[0032] Step 2: Obtain the coil temperature and ore temperature of the detection coil and the voltage data of the sensor during the actual mining process, and input them into the calibration model to obtain the corresponding ore flow rate, and complete the measurement of the multi-metal nodule conveying flow rate.

[0033] During the actual ore flow rate measurement process, collect the temperature TS of the sensor coil, the temperature TM of the multi-metal nodule ore, and the output voltage of the sensor, perform data calculation and processing according to the requirements of the feature input, and use the predict function of the calibration model to calculate the feature input. The output value of the model is the calculated multi-metal nodule ore flow rate.

[0034] In the experimental data, under the condition of maintaining a given standard conveying flow rate of 800 kg / min, when the ore temperature exceeds the initial experimental temperature by 10 °C, the flow rate calculated according to the original method of the polymetallic nodule detection sensor has an offset of about 14%, resulting in a large error. However, the flow rate error output by the model of the present invention is about 5%. Even when the ore temperature exceeds the initial temperature by 20 °C, the maximum deviation between the flow rate data calculated by the present invention and the standard flow rate is still within 10%, which proves that the data stability of the method proposed by the present invention in a relatively wide temperature range is significantly improved.

[0035] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations according to the concept of the present invention without creative work. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field based on the concept of the present invention through logical analysis, reasoning or limited experiments on the basis of the prior art should fall within the protection scope determined by the claims.

Claims

1. A method for measuring the transport flow of polymetallic nodules in a wide temperature range, characterized in that: The steps include: Step 1: Based on the triple cycle, the coil temperature, ore temperature, sensor voltage data and ore flow detection data of metal nodule flow detection are collected, and the characteristic variables and correction model are constructed based on Bloch's law. The training set and validation set are obtained according to the characteristic variables, correction model and detection data, and the correction model is trained based on the training set and validation set; Step 2: Obtain the coil temperature of the detection coil and the ore temperature as well as the voltage data of the sensor during the actual mining process, and input them into the calibration model to obtain the corresponding ore flow rate and complete the measurement of the polymetallic nodule transport flow rate.

2. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 1, characterized in that: The triple-cycle-based collection of coil temperature, ore temperature, sensor voltage data and ore flow detection data for metal nodule flow detection includes: setting corresponding cycle initial values ​​for the coil temperature, ore temperature and ore flow, constructing a loop nesting process based on the cycle initial values, executing the loop nesting process and recording the data of the coil temperature, ore temperature, ore flow and sensor voltage data in all cycles as detection data.

3. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 2, characterized in that: The constructing of the loop nesting process based on the loop initial value includes: calculating the number of cycles of the coil temperature cycle, the ore temperature cycle and the ore flow cycle respectively based on the loop initial value, and constructing the loop nesting process based on the number of cycles.

4. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 3, characterized in that: The cycle initial values ​​include: coil initial temperature, coil termination temperature, coil temperature growth step, ore initial temperature, ore termination temperature, ore temperature growth step, ore initial flow, ore termination flow and ore flow growth step; The number of cycles of the coil temperature cycle is calculated by the following formula: ； in, Indicates the number of cycles of coil temperature cycle, Indicates the coil termination temperature, represents the initial temperature of the coil, Indicates the coil temperature growth step; The number of cycles of the ore temperature cycle is calculated by the following formula: ； in, Indicates the number of cycles of the ore temperature cycle, Indicates the ore termination temperature, represents the initial temperature of the ore, Indicates the step length of ore temperature growth; The number of cycles of the ore flow cycle is calculated by the following formula: ； in, Indicates the number of cycles of ore flow cycle, Indicates the end of ore flow, represents the initial flow rate of ore, Indicates the growth step of ore flow.

5. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 4, characterized in that: The loop nesting process is constructed by executing the loop nesting process in the order of ore flow cycle, coil temperature cycle, and ore temperature cycle in combination with loop rules; The cycle rule includes: executing one coil temperature cycle for every k ore temperature cycles, executing one ore flow cycle for every n coil temperature cycles, and executing m ore flow cycles in total, and adjusting the coil temperature, ore temperature and ore flow according to the corresponding coil temperature growth step, ore temperature growth step and ore flow growth step for each execution of the ore temperature cycle, the coil temperature cycle and the ore flow cycle; After executing a coil temperature cycle, the current ore temperature is adjusted to the ore initial temperature. After executing an ore flow cycle, the current coil temperature is adjusted to the coil initial temperature, and the current ore temperature is adjusted to the ore initial temperature.

6. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to any one of claims 1 to 5, characterized in that: The method of constructing the characteristic variable and the correction model based on Bloch's law includes: based on the linear relationship between magnetic resistance and the 1.5th power of temperature in Bloch's law, the coil temperature and the ore temperature have a common nonlinear influence on the measurement output of the sensor, the magnetic resistance is linearly related to the flow rate, and is linearly related to the square of the output voltage of the sensor to construct the characteristic variable and the correction model; The characteristic variable is expressed by the following formula: ； in, Indicates the coil temperature, Indicates the ore temperature, represents the common impact factor, Indicates the voltage signal output by the sensor; The correction model is expressed by the following formula: ； in, represents the ore flow rate output by the calibration model, is the model coefficient.

7. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 6, characterized in that: The method of obtaining a training set and a validation set based on the characteristic variables, the correction model and the detection data includes: performing combined calculation based on the characteristic variables, the correction model and the detection data to obtain a structured data set. and , for structured data sets and After normalization, the training set and validation set are obtained.

8. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 7, characterized in that: After normalization to obtain the training set and validation set, the model was calibrated using the 5-fold crossover method and support vector regression training.

9. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 7, characterized in that: The normalization process is Standardized methods.

10. The method for measuring the transport flow of polymetallic nodules in a wide temperature range according to claim 1, characterized in that: The calibration model obtains hyperparameters through the GridSearchCV function during the training process.

Images