A temperature compensation method for electrolyte horizontal tilt sensor

Abstract

The present invention discloses a temperature compensation method for an electrolyte horizontal tilt angle sensor, belonging to the field of angle measurement. The method comprises: installing a plurality of temperature sensors on the sensor head of the electrolyte horizontal tilt angle sensor; collecting a training data set at consecutive angle points and temperature points, wherein each sample in the training data set includes a temperature value output by the temperature sensor, an inclination value output by the electrolyte horizontal tilt angle sensor, and a turntable rotation angle; evaluating the degree of outliers in the temperature data of each sample using a local outlier factor method, and removing abnormal samples; training a temperature error model using the training data set after removing abnormal samples, and compensating the electrolyte horizontal tilt angle sensor based on the trained temperature error model. The present invention reduces the influence of temperature on the angle of the electrolyte horizontal tilt angle sensor through the temperature error model, thereby further improving the accuracy of the electrolyte horizontal tilt angle sensor.

Description

Technical Field

[0001] The present invention relates to the field of angle measurement, and in particular to a temperature compensation method for an electrolyte horizontal tilt sensor. Background Art

[0002] The electrolyte horizontal tilt sensor places sensing electrodes inside a container made of glass, ceramic, or plastic, connected to the outside. A certain volume of electrolyte with gaps is added to the container. The flow of the electrolyte causes changes in the current and electric field passing through the electrodes, thereby measuring the tilt angle using these changes. This sensor has a wide range of applications, including bridge safety, dam safety, rail transit, and wind tunnel safety. The sensor integrates multiple signal processing units within the entire system for data acquisition and processing. It is a miniature sensor system that integrates signal processing and control functions. It can be stacked within a limited space, is more robust, can withstand greater impact and pressure, is more compact, and is easy to carry, providing more accurate readings to meet user needs.

[0003] Because the sensor requires power, changes in electrolyte temperature significantly impact the liquid's electric field and current. Liquid-based tilt sensors, which calculate the corresponding horizontal tilt angle based on the current and electric field changes in the energized liquid, are typically sensitive to temperature, causing their physical properties to be affected by temperature. Furthermore, after the entire system is designed, the sensor must be packaged, resulting in poor heat dissipation and further affecting its characteristics due to temperature changes. Therefore, eliminating the impact of temperature on the output angle of electrolyte-based tilt sensors is a pressing issue.

[0004] Currently, common solutions to temperature errors include developing sensors with reduced temperature sensitivity and adding internal temperature control devices to the sensor. While these two methods can help address sensor temperature errors, they also increase the sensor's size and cost, hindering engineering applications. Therefore, with engineering applications and cost savings as the starting point, a temperature compensation method for electrolyte level tilt sensors is needed that is applicable to most sensor structures. Summary of the Invention

[0005] In order to solve the shortcomings of the existing temperature effect on the electrolyte horizontal tilt sensor, the present invention provides a temperature compensation method for the electrolyte horizontal tilt sensor. By obtaining the tilt sensor angle data and temperature data, the relationship between the two is modeled, and the tilt sensor angle data is compensated in combination with the temperature data, so as to improve the accuracy of the electrolyte horizontal tilt sensor.

[0006] The present invention solves the above technical problems by the following technical solutions:

[0007] A temperature compensation method for an electrolyte horizontal tilt sensor comprises the following steps:

[0008] Step 1: Install several temperature sensors on the sensor head of the electrolyte level tilt sensor;

[0009] Step 2: collecting a training data set at consecutive angle points and temperature points, wherein each sample in the training data set includes a temperature value output by a temperature sensor, an inclination value output by an electrolyte level inclination sensor, and a rotation angle of a turntable;

[0010] Step 3: Use the local outlier factor method to evaluate the outlier degree of the temperature data in each sample and remove abnormal samples;

[0011] Step 4: Use the training data set after removing abnormal samples to train the temperature error model, and compensate the electrolyte horizontal tilt sensor according to the trained temperature error model.

[0012] Furthermore, the temperature sensor has the same number as the electrode sheets inside the sensor head and their positions correspond one to one.

[0013] Furthermore, the training data set at continuous angle points and temperature points is expressed as i=1,2,…T;k=1,2,…n;where, represents the kth group of temperature values ​​measured at the i-th temperature point, n represents the number of groups of temperature values ​​measured at each temperature point, A k 、A tk They respectively represent the inclination value output by the electrolyte horizontal inclination sensor and the turntable rotation angle value in the kth group of samples at a certain temperature point.

[0014] Furthermore, the data collection method of the training data set includes:

[0015] (2.1) Setting a designated temperature point according to the operating temperature range of the electrolyte level tilt sensor, wherein the designated temperature point covers the operating temperature range;

[0016] (2.2) Install the electrolyte horizontal tilt sensor on the turntable in the incubator, with the sensitive axis of the electrolyte horizontal tilt sensor parallel to the rotation axis of the turntable. In this case, the electrolyte horizontal tilt sensor is used to measure the rotation angle of the turntable.

[0017] (2.3) The initial temperature of the incubator is controlled at the lowest value of the specified temperature point. After keeping warm, the electrolyte horizontal tilt sensor is powered on. First, the turntable input rotation angle is initialized to 0°, and it rotates counterclockwise or clockwise at a fixed angular velocity. The temperature value of the temperature sensor and the tilt value output by the electrolyte horizontal tilt sensor are read at each specified rotation angle. When the turntable rotates one circle and returns to the initialization position, the incubator is controlled to heat up to the next specified temperature point and keep warm. The turntable rotation and sensor data reading process are repeated until the corresponding data at the last temperature point is collected.

[0018] Furthermore, the range of the designated temperature point is -40°C to 65°C.

[0019] Furthermore, the step 3 includes:

[0020] (3.1) Calculate the Euclidean distance between the mean of each group of temperature values ​​at the same temperature point and the mean of other groups of temperature values ​​at the same temperature point;

[0021] (3.2) Calculate the local reachability density of the mean of each group of temperature values ​​at the same temperature point based on the Euclidean distance value;

[0022] (3.3) The outlier factor of the mean of each group of temperature values ​​at the same temperature point is calculated based on the local reachable density. If the outlier factor is greater than 1, the samples containing the corresponding group of temperature values ​​are removed from the training data set.

[0023] Furthermore, the local reachable density calculation formula is:

[0024]

[0025] in, represents the local reachable density of the mean of the kth group of temperature values ​​at temperature point i, represents the kth group of temperature values ​​measured at the i-th temperature point, represents the mean value of the kth group of temperature values ​​at temperature point i, and n represents the number of groups of temperature values ​​measured at each temperature point. represents the mean temperature value of other groups except the kth group at temperature point i, It represents the Euclidean distance between the mean temperature value of the kth group at temperature point i and the mean temperature value of other groups.

[0026] Furthermore, the outlier factor calculation formula is:

[0027]

[0028] in, represents the mean value of the jth group of temperature values ​​at temperature point i, Indicates the outlier factor of the mean of the kth group of temperature values ​​at temperature point i.

[0029] Furthermore, the temperature error model adopts a least squares support vector machine, takes the mean of each group of temperature values ​​at the same temperature point obtained by eliminating outliers as temperature data, takes the inclination value output by the electrolyte horizontal inclination sensor corresponding to the temperature data as the inclination value to be compensated, and takes the turntable rotation angle value corresponding to the temperature data as the true inclination value to train the LSSVM model.

[0030] Furthermore, when compensating the electrolyte horizontal tilt sensor according to the trained temperature error model, the mean temperature value and the tilt value output by the electrolyte horizontal tilt sensor are input, and the trained LSSVM model generates the compensated tilt value as the final result.

[0031] The present invention provides a method for temperature compensation of an electrolyte horizontal tilt sensor. The method studies the effect of temperature on the angle of the electrolyte horizontal tilt sensor and designs a training set experiment by collecting data at consecutive angles and temperatures. The method collects output data from the electrolyte horizontal tilt sensor at consecutive angles and temperatures. The temperature data is preprocessed and error modeled using a least squares support vector machine (LSSVM) to compensate for the tilt of the electrolyte horizontal tilt sensor. Experimental results demonstrate that the effect of temperature on the angle of the electrolyte horizontal tilt sensor is reduced, further improving the accuracy of the electrolyte horizontal tilt sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 4 is a flow chart of the temperature compensation method of the electrolyte horizontal tilt sensor shown in this embodiment.

[0033] Figure 2 Schematic diagram of the installation of the temperature sensor on the electrolyte level inclination sensor shown in this embodiment.

[0034] In the figure: 1-sensor head; 2-temperature sensor; 3-electrode sheet; 4-liquid level.

[0035] Figure 3 1 is a comparison result of the tilt angle data before and after temperature compensation shown in this embodiment. DETAILED DESCRIPTION

[0036] The technical solution of the present invention is further described in detail below with reference to the accompanying drawings.

[0037] like Figure 1 As shown, the present invention collects data at continuous angle points and temperature points to form a training set, and trains a temperature error model to compensate for the electrolyte level tilt sensor. Specifically, the following steps are included:

[0038] (1) A number of temperature sensors are installed on the sensor head 1 of the inclination sensor, and the installation position of each temperature sensor 2 corresponds to each electrode piece 3 inside the inclination sensor; since the temperature sensor is usually placed on the communication signal board, the data output by the temperature sensor at this time cannot fully reflect the real temperature information of each electrode piece 3 built into the sensor head 1. In addition, since there are many heat sources in the electrolyte level inclination sensor, each temperature sensor is installed relatively compactly, and the temperature information between the temperature sensors is different. When the ambient temperature field is complex, the temperature information of each electrode piece 3 built into the sensor head 1 cannot be measured in time to output an accurate temperature value. Therefore, the present invention compensates in the inclination range of 0° to ±180° within the full temperature range.

[0039] In a specific implementation of the present invention, Figure 2 As shown, six temperature sensors 2 are respectively adhered to the corresponding outer sides of the six electrode sheets 3 inside the sensor head through thermal conductive adhesive, and then the electrolyte horizontal tilt sensor is installed on the turntable in the temperature box. The sensitive axis of the electrolyte horizontal tilt sensor is parallel to the rotation axis of the turntable. At this time, the electrolyte horizontal tilt sensor is used to measure the rotation angle of the turntable.

[0040] (2) Collect data at continuous angle points and temperature points to form a training set D.

[0041] Can be abbreviated as i=1,2,…T;k=1,2,…n. represents the kth group of temperature values ​​measured at the i-th temperature point; n represents the number of groups of temperature values ​​measured at each temperature point, that is, n groups of samples are measured at each temperature point, and each group of samples corresponds to a different turntable rotation angle; A k 、A tk They respectively represent the inclination value output by the electrolyte horizontal inclination sensor and the rotation angle value output by the turntable in the kth group of samples at a certain temperature point.

[0042] In a specific embodiment of the present invention, a total of 11 temperature points, namely -40°C, -30°C, -20°C, -10°C, 0°C, 10°C, 20°C, 30°C, 40°C, 50°C, and 65°C, and a total of 7 rotation angle points, namely 0°, ±60°, ±120°, and ±180°, are designated as experimental objects. The steps are as follows:

[0043] (2-1) The initial temperature of the incubator is controlled at -40°C and maintained for one hour, and the electrolyte level tilt sensor is powered on.

[0044] (2-2) One hour after the electrolyte level tilt sensor is powered on, the turntable provides a changing input angular position at each specified temperature point.

[0045] In this embodiment, the temperature and the rotation angle of the turntable change rules are as follows:

[0046] The incubator is initially set to -40°C. The turntable is initialized to a rotation angle of 0° and rotated counterclockwise or clockwise at an angular velocity of 10° / s. The temperature sensor and the electrolyte level tilt sensor output are read at each specified rotation angle. After the turntable completes one rotation and returns to its initialization position, the incubator is heated to the next specified temperature point and held there for one hour. The turntable rotation and sensor data reading process is repeated until the data corresponding to the last temperature point is collected. In this embodiment, the temperature sensor output value is a set of temperature data consisting of the output values ​​of six temperature sensors.

[0047] (3) Temperature error data preprocessing.

[0048] The present invention uses the local outlier factor method to evaluate the outlier degree of temperature data in each sample, removes abnormal samples, and preprocesses the sensor data.

[0049] In one embodiment of the present invention, the mean of a set of temperature values ​​is first calculated. The average temperature value of other groups at this temperature point The Euclidean distance of :

[0050]

[0051] in, represents the mean value of the kth group of temperature values ​​at temperature point i, represents the mean value of the jth group of temperature values ​​at temperature point i, represents the mean temperature value of other groups except the kth group at temperature point i, represents the Euclidean distance between the mean of the kth temperature value and the mean of the temperature values ​​of other groups at temperature point i. In this embodiment, It represents the mean value of the six sensors in the kth group at temperature point i.

[0052] Calculate the local reachable density of the mean temperature value at the same temperature point. The calculation formula is:

[0053]

[0054] in, Represents the local reachable density of the mean of the kth group of temperature values ​​at temperature point i.

[0055] The outlier factor of the mean temperature value at the same temperature point is calculated using the formula:

[0056]

[0057] in, Indicates the outlier factor of the mean temperature value of the kth group at temperature point i. The mean temperature value at this temperature point is significantly different from that of its neighborhood, so it is considered an outlier. The sample of the item is removed from the collected data; if The mean temperature value at this temperature point is slightly different from that of its neighborhood, so it is considered a normal point and no processing is required.

[0058] (4) A temperature error model is constructed using the data set after removing outliers to compensate for the electrolyte horizontal tilt sensor.

[0059] In a specific implementation of the present invention, the temperature error model uses a least squares support vector machine (LSSVM) to remove the outliers and obtain the mean of each group of temperature values ​​at the same temperature point. As temperature data, the corresponding The electrolyte level inclination sensor outputs the inclination value A k As the inclination value to be compensated, the corresponding The turntable rotation angle value A tk As the true value of the inclination, A k As input value, take A tk As the output value, the LSSVM model is trained.

[0060] (5) The temperature error model is used to compensate the electrolyte horizontal tilt sensor.

[0061] In a specific implementation of the present invention, the average value of the temperature sensor installed on the electrolyte horizontal tilt sensor and the tilt value output by the electrolyte horizontal tilt sensor are collected as inputs of the LSSVM model, and the LSSVM model directly outputs the compensated tilt value as the final result.

[0062] like Figure 3 As shown, the horizontal axis is the nominal angle value, and the vertical axis is the error angle value, that is, the measurement error between the inclination value output by the electrolyte horizontal inclination sensor and the nominal angle value. It can be seen that after compensation by the method of the present invention, the measurement error is controlled at ±0.1°. Compared with the data before compensation, the measurement error is significantly reduced.

[0063] In the embodiments of the present invention, those skilled in the art will understand that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention to other forms. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can still modify the technical contents described in the above embodiments or replace some of the technical features therein with equivalents. Any modifications, equivalent replacements, etc. made within the technical essence of the present invention and the above embodiments shall be included in the scope of protection of the present invention.

Claims

1. A temperature compensation method for an electrolyte horizontal tilt sensor, characterized in that: The following steps are involved: Step 1: Install several temperature sensors on the sensor head of the electrolyte level tilt sensor; Step 2: collecting a training data set at continuous angle points and temperature points, wherein each sample in the training data set includes temperature data output by a temperature sensor, an inclination value output by an electrolyte level inclination sensor, and a rotation angle of a turntable; The data collection method of the training data set includes: (2.1) Setting a designated temperature point according to the operating temperature range of the electrolyte level tilt sensor, wherein the designated temperature point covers the operating temperature range; (2.2) Install the electrolyte horizontal tilt sensor on the turntable in the incubator, with the sensitive axis of the electrolyte horizontal tilt sensor parallel to the rotation axis of the turntable. In this case, the electrolyte horizontal tilt sensor is used to measure the rotation angle of the turntable. (2.3) The initial temperature of the incubator is controlled at the lowest value of the specified temperature point. After keeping warm, the electrolyte horizontal tilt sensor is powered on. First, the turntable input rotation angle is initialized to 0°, and it rotates counterclockwise or clockwise at a fixed angular velocity. The temperature value of the temperature sensor and the tilt value output by the electrolyte horizontal tilt sensor are read at each specified rotation angle. When the turntable rotates one circle and returns to the initialization position, the incubator is controlled to heat up to the next specified temperature point and keep warm. The turntable rotation and temperature sensor and electrolyte horizontal tilt sensor data reading process are repeated until the data corresponding to the last temperature point is collected. Step 3: Use the local outlier factor method to evaluate the outlier degree of the temperature data in each sample and remove abnormal samples; Step 4: using the training data set after removing abnormal samples to train the temperature error model, and compensating the electrolyte horizontal tilt sensor according to the trained temperature error model; The temperature error model adopts the least squares support vector machine, and the mean of each group of temperature values ​​at the same temperature point obtained by removing abnormal samples is used as the temperature data, the inclination value output by the electrolyte level inclination sensor corresponding to the temperature data is used as the inclination value to be compensated, and the turntable rotation angle value corresponding to the temperature data is used as the true inclination value to train the LSSVM model; When compensating the electrolyte horizontal tilt sensor according to the trained temperature error model, the mean temperature value and the tilt value output by the electrolyte horizontal tilt sensor are used as input, and the trained LSSVM model generates the compensated tilt value as the final result.

2. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 1, characterized in that: The temperature sensors are the same in number as the electrode sheets inside the sensor head and their positions correspond one to one.

3. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 1, characterized in that: The training data set at continuous angle points and temperature points is represented as in, represents the kth group of temperature values ​​measured at the i-th temperature point, n represents the number of groups of temperature values ​​measured at each temperature point, A k 、A tk They represent the inclination value output by the electrolyte horizontal inclination sensor and the turntable rotation angle value in the kth group of samples at a certain temperature point.

4. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 1, characterized in that: The range of the designated temperature points is from -40°C to 65°C.

5. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 1, characterized in that: The step 3 includes: (3.1) Calculate the Euclidean distance between the mean of each group of temperature values ​​at the same temperature point and the mean of other groups of temperature values ​​at the same temperature point; (3.2) Calculate the local reachability density of the mean of each group of temperature values ​​at the same temperature point based on the Euclidean distance; (3.3) The outlier factor of the mean of each group of temperature values ​​at the same temperature point is calculated based on the local reachable density. If the outlier factor is greater than 1, the samples containing the corresponding group of temperature values ​​are removed from the training data set.

6. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 5, characterized in that: The local reachable density calculation formula is: in, represents the local reachable density of the mean of the kth group of temperature values ​​at temperature point i, represents the kth group of temperature values ​​measured at the i-th temperature point, represents the mean value of the kth group of temperature values ​​at temperature point i, and n represents the number of groups of temperature values ​​measured at each temperature point. represents the mean temperature value of other groups except the kth group at temperature point i, It represents the Euclidean distance between the mean temperature value of the kth group at temperature point i and the mean temperature value of other groups.

7. The temperature compensation method of the electrolyte horizontal tilt sensor according to claim 6, characterized in that: The outlier factor calculation formula is: in, represents the mean value of the jth group of temperature values ​​at temperature point i, Indicates the outlier factor of the mean of the kth group of temperature values ​​at temperature point i.

Images