CN111998885A - A kind of parameter calibration system for measurement - Google Patents

Abstract

The invention relates to the technical field of measurement, and discloses a parameter calibration system for measurement, which comprises a reference object mounting platform, a sensor mounting plate and a base platform. A first longitudinal sliding component and a transverse sliding component are also provided; the sensor mounting plate is arranged on the transverse sliding component, and the measuring sensor is arranged on the sensor mounting plate, which slides laterally with the transverse sliding component and the transverse sliding component follows. The first longitudinal sliding component slides longitudinally. A MSP430 single-chip detection unit is also set on the sensor mounting board, and a calibration algorithm is set on it to calibrate the measurement value of the measurement sensor. Compared with the prior art, the present invention realizes the lateral movement and longitudinal movement of the measuring sensor through the first longitudinal sliding component and the lateral sliding component, and calibrates the measured value of the measuring sensor through the monitoring unit, which can accurately calibrate the measuring sensor. Measurements.

Description

A kind of parameter calibration system for measurement

technical field

The invention relates to the technical field of measurement, in particular to a parameter calibration system for measurement.

Background technique

As the source of obtaining natural information, sensor is an indispensable and important element in industrial production, scientific research and other fields. There are many kinds of measurement sensors in the prior art, such as temperature and humidity sensors, infrared sensors, photoelectric sensors and so on.

For the various existing measurement sensors, in the measurement process, environmental factors have a great influence on the measurement, such as environmental pressure, environmental temperature, etc., have a certain impact on the measurement value of the measurement sensor, and how to accurately The actual parameter value of the reference object is measured. When using the measurement sensor to measure, how to calibrate and measure the value measured by the measurement sensor, so that the measured parameter value is the accurate parameter value of the reference object. The traditional measurement sensor measurement is not based on the reference object. The non-linearity of parameter changes, large hysteresis and complex changes in the displacement of the measurement sensor make it impossible to monitor and predict the measurement value of the measurement sensor.

SUMMARY OF THE INVENTION

Purpose of the invention: In view of the problems existing in the prior art, the present invention provides a parameter calibration system for measurement, which can quickly adjust the difference between the measurement sensor and the reference object by cooperating with the longitudinal sliding component and the transverse sliding component. Longitudinal distance, horizontal distance, and at the same time, the measurement value of the measurement sensor is calibrated through the MSP430 single-chip monitoring unit, which can accurately calibrate the measurement value of the measurement sensor.

Technical solution: The present invention provides a parameter calibration system for measurement, including a reference object mounting platform, a sensor mounting plate and a base platform, wherein the reference object mounting platform and the sensor mounting plate are arranged on the base platform, so the The base platform is also provided with a first longitudinal sliding assembly and a lateral sliding assembly; the sensor mounting plate is arranged on the lateral sliding assembly, and the sensor mounting plate slides laterally with the lateral sliding assembly And the lateral sliding component slides longitudinally along with the first longitudinal sliding component; a measurement sensor and an MSP430 single-chip monitoring unit are also arranged on the sensor mounting board, and the MSP430 single-chip monitoring unit includes a time-series DRNN neural network prediction model , empirical mode decomposition model, multiple Elman neural network models, multiple NARX neural network models, wavelet neural network fusion model and metabolic GM gray predictor; the output of the measurement sensor is used as the input of the time series DRNN neural network prediction model, The output of the time series DRNN neural network prediction model is used as the input of the empirical mode decomposition model, and the output of the empirical mode decomposition model outputs the low-frequency trend part and multiple high-frequency fluctuation parts of the measurement sensor as the input of the corresponding multiple Elman neural network models, respectively. The outputs of the multiple Elman neural network models are used as the input of the corresponding multiple NARX neural network models, the outputs of the multiple NARX neural network models are used as the input of the wavelet neural network fusion model, and the output of the wavelet neural network fusion model is used as the metabolic GM gray prediction. The input of the metabolic GM gray predictor is used as the measurement calibration value of the measurement sensor.

Further, the first longitudinal sliding assembly includes a first longitudinal sliding guide rail and a first longitudinal sliding platform, the first longitudinal sliding guide rail is fixed on the base platform through a bracket, and its lower surface is along its length. A first screw rod is connected in the direction of rotation, a first sliding block is threadedly connected to the first screw rod, and the first longitudinal sliding platform is sleeved on the first longitudinal sliding guide rail and is connected with the first longitudinal sliding rail. A corresponding position of a rod is fixedly connected with the first sliding block.

Further, the lateral sliding assembly includes a second screw rod that is perpendicular to the first longitudinal sliding guide rail and is rotatably connected to the upper surface of the first longitudinal sliding platform, and the second screw rod is threadedly connected with a second screw rod. A second sliding block, a lateral sliding platform is fixed on the second sliding block, and the sensor mounting plate is fixed on the lateral sliding platform.

Further, the two sides of the first longitudinal sliding guide rail are set as inwardly recessed arc guide rails, and the first longitudinal sliding platform and the arc guide rail are matched at corresponding positions.

Further, a plurality of guide rail balls are rolled and arranged on the first longitudinal sliding platform at the matching position of the arc guide rail.

Further, the upper surface of the first longitudinal sliding guide rail is also provided with a pair of longitudinal linear guide rails along the sliding direction of the first longitudinal sliding platform, and the first longitudinal sliding platform and the longitudinal linear guide rails have corresponding positions. A pair of bar-shaped protrusions are provided, and each of the bar-shaped protrusions is matched with each of the longitudinal linear guide rails.

Further, the lateral sliding assembly is also provided with a second longitudinal sliding assembly, and the second longitudinal sliding assembly includes a third longitudinal sliding assembly that is rotatably connected to the lateral sliding platform and is perpendicular to the second screw rod. A lead screw, a third sliding block is threadedly connected to the third lead screw, the pitch of the third lead screw is smaller than that of the first lead screw, and the sensor mounting plate is fixed on the lateral sliding platform superior.

Further, the first longitudinal sliding assembly, the lateral sliding assembly and the second longitudinal sliding assembly are all provided with driving mechanisms with the same structure, and the driving mechanisms are the first stepping motor and the second stepping motor respectively. and a third stepper motor, one end of the first screw rod, the second screw rod and the third screw rod are all provided with a main bevel gear and a slave bevel gear with the same structure, and the output shafts of the three stepper motors are all connected with the same structure. The center of the corresponding main bevel gear is fixedly connected, the main bevel gear meshes with the corresponding secondary bevel gear, and the secondary bevel gear is sleeved and fixed on its corresponding screw rod.

Further, one end of the first screw rod, the second screw rod and the third screw rod is respectively fixed with a rotating handle.

Further, the reference object mounting platform also includes a pad height fixing platform fixed on the base platform, and a plurality of vertically arranged electric push rods fixed on the top of the pad height fixing platform, and the top of the electric push rod is set. There is the reference mount.

Beneficial effects:

1. The time series DRNN neural network prediction model used in the present invention is a dynamic regression neural network with feedback and the ability to adapt to time-varying characteristics. The network can more directly and vividly reflect the dynamic changing performance of the measurement value of the measurement sensor, and can be more accurate. To calibrate the actual value of the measurement sensor, the time series DRNN neural network model is a 10-21-1 3-layer network structure, the hidden layer is the regression layer, and the output layer is the measurement calibration value of the measurement sensor.

2. The input of the NARX neural network model used in the present invention includes the output of the corresponding Elman neural network model for a period of time and the output history feedback of the NARX neural network model. This part of the feedback input can be considered to include the corresponding Elman neural network for a period of time. The historical information of the model output and the corresponding NARX neural network model output participates in the calibration of the measurement output value of the measurement sensor. For a suitable feedback time length, the NARX neural network model output has a good effect on the sensor calibration prediction. The empirical mode decomposition (EMD) model of this patent, the Elman neural network model and the NARX neural network model provide an effective The method for calibrating the measured value of the measuring sensor.

3. The NARX neural network model used in the present invention is a dynamic neural network model that can effectively perform dynamic calibration and prediction on the nonlinear and non-stationary time series of the measured values of the measurement sensor. Improves the accuracy of calibration predictions for time series measurements of measurement sensors. Compared with the traditional forecasting model method, this method has the advantages of good effect in dealing with non-stationary time series, fast calculation speed and high accuracy. This patent verifies that the NARX neural network model can improve the accuracy and reliability of time series calibration prediction by comparing the actual data of the non-stationary measurement sensor measurement values. At the same time, the experimental results also prove that the NARX neural network model performs better than the traditional calibration model in the non-stationary time series calibration.

4. The present invention adopts the dynamic recursive network established by introducing the delay module and output feedback into the NARX neural network model, and takes the output of the Elman neural network model as the input and the output vector delay feedback of the NARX neural network model into the network training, forming a The new input vector has good nonlinear mapping ability. The input of the NARX neural network model includes not only the input data of the original Elman neural network model, but also the output data of the trained NARX neural network model. The generalization of the NARX neural network model network The ability is improved, so that it has better calibration accuracy and adaptive ability than the traditional static neural network control in the time series calibration of nonlinear measurement sensors.

5. The present invention uses the metabolic GM(1,1) gray predictor to predict the measurement value of the measurement sensor with a long time span. Using the metabolic GM(1,1) gray predictor, the measurement values of the measurement sensors in the future can be predicted according to the historical parameter values of the wavelet neural network fusion model. After the measurement values of the measurement sensors predicted by the above method, they are added to the wavelet neural network. In the original sequence output by the fusion model, correspondingly remove a data modeling at the beginning of the sequence, and then calibrate the measurement value of the measurement sensor. And so on, the measurement value of the measuring sensor is calibrated out. This method, called the metabolic GM(1,1) grey predictor, enables calibrated predictions over longer periods of time. Users can more accurately grasp the change trend of the measurement value of the measurement sensor and prepare for accurate measurement of the measurement value.

Sixth, the present invention sets the longitudinal sliding component and the transverse sliding component to cooperate with each other, so as to facilitate the adjustment of the longitudinal and transverse distances between the measurement sensor and the reference object, which can be realized by either a stepping motor or a hand rotating handle. Use in specific situations.

7. In the present invention, the two sides of the first longitudinal sliding guide rail of the first longitudinal sliding assembly are set as inwardly recessed arc guide rails, and the first longitudinal sliding platform and the arc guide rail are matched at the corresponding positions, so that when the When a longitudinal sliding platform slides on the first longitudinal sliding guide rail, there will be no shaking and tilting, and the first longitudinal sliding platform is relatively fixed to slide on the first longitudinal sliding guide rail, which is more stable. The friction force is reduced by the guide balls and the resistance during the sliding process is reduced. A longitudinal linear guide rail is also arranged on the upper surface of the first longitudinal sliding guide rail, and the first longitudinal sliding platform is further limited relative to the first longitudinal sliding guide rail.

8. The present invention is provided with a second longitudinal sliding assembly on the lateral sliding assembly, and the screw pitch of the lead screw on the second longitudinal sliding assembly is smaller than that of the first longitudinal sliding assembly, which plays a fine-tuning role and makes the adjustment more precise. Precise, can adjust the distance before and after the reference in a small range.

9. An electric push rod is arranged under the reference object mounting platform of the present invention, which is convenient to adjust the height of the reference object and facilitate the measurement of the measurement sensor.

Description of drawings

Fig. 1 is the overall structure schematic diagram of the present invention;

Fig. 2 is the system block diagram of MSP430 single-chip microcomputer monitoring unit of the present invention;

3 is a schematic diagram of the overall structure of the fine-tuning positioning of the measurement sensor of the present invention;

4 is a schematic diagram of the connection structure between the first longitudinal sliding platform and the first screw rod of the present invention;

FIG. 5 is a schematic structural diagram of the first longitudinal sliding platform of the present invention.

Among them, 1- base platform, 2- sensor mounting plate, 3- reference object mounting platform, 101- pad height fixed platform, 102- electric push rod, 201- first longitudinal sliding guide rail, 202- first longitudinal sliding Platform, 203-bracket, 204-first screw, 205-first slider, 206-second screw, 207-second slider, 208-transverse sliding platform, 209-arc guide, 210-guide Ball, 211-longitudinal linear guide, 212-stripe protrusion, 213-third screw, 214-third slider, 215-first stepper motor, 216-second stepper motor, 217-third step Incoming motor, 218-main bevel gear, 219-slave bevel gear, 220-rotating handle, 221-foot slot.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solutions of the present invention more clearly, and cannot be used to limit the protection scope of the present invention.

The invention discloses a parameter calibration system for measurement. The structure of the calibration system mainly includes a reference object mounting platform 3, a sensor mounting plate 2 and a base layer platform 1. The reference object mounting platform 3 and the sensor mounting plate 2 are arranged on the base layer. On the platform 1, the base platform 1 is also provided with a first longitudinal sliding assembly and a lateral sliding assembly, the sensor mounting plate 2 is arranged on the lateral sliding assembly, and the sensor mounting plate 2 slides laterally with the lateral sliding assembly. And the lateral sliding component slides longitudinally along with the first longitudinal sliding component. A foot slot 221 is provided on the sensor mounting plate 2, which is convenient for fixing the measuring sensor by a tripod or the like.

The reference object mounting platform 3 also includes a pad height fixing platform 101 fixed on the base platform 1, and a plurality of vertically arranged electric push rods 102 fixed on the top of the pad height fixing platform 101. The top of the electric push rod 102 is provided with a reference object mounting platform 3.

The calibration part of the calibration system mainly includes an MSP430 single-chip microcomputer monitoring unit. The MSP430 single-chip microcomputer monitoring unit is arranged on the sensor mounting plate 2. The MSP430 single-chip microcomputer monitoring unit is arranged on the sensor mounting plate 2 along with the first longitudinal sliding component and lateral sliding. The vertical and horizontal distances are fine-tuned by the assembly and the second longitudinal sliding assembly.

The MSP430 microcontroller monitoring unit includes a time series DRNN neural network prediction model, an empirical mode decomposition (EMD) model, multiple Elman neural network models, multiple NARX neural network models, a wavelet neural network fusion model and a metabolic GM (1,1) Grey predictor, a parameter calibration system is designed in the MSP430 single-chip monitoring unit to realize the calibration and prediction of the measurement value of the measurement sensor. The block diagram of the parameter calibration system is shown in Figure 2, and the design process is as follows:

1. Design of time series DRNN neural network prediction model

The time series DRNN neural network prediction model is a dynamic regression neural network with feedback and the ability to adapt to time-varying characteristics. The time series DRNN neural network prediction model can more directly and vividly reflect the dynamic change performance of the measurement sensor measurement value, and can predict more accurately The measurement value of the sensor is measured. The time series DRNN neural network prediction model is a 10-21-1 three-layer network structure, and its hidden layer is a regression layer. In the time series DRNN neural network prediction model of the present invention, the time series measurement value of the output of the measurement sensor is the input of the time series DRNN neural network prediction model, and I=[I ₁ (t), I ₂ (t),..., I _n (t)] is the input vector of the time series DRNN neural network prediction model, where I _i (t) is the input of the ith neuron in the input layer of the time series DRNN neural network prediction model at time t, and the jth neuron in the regression layer The output is X _j (t), S _j (t) is the sum of the input of the jth regression neuron, and f( ) is a function of S, then O(t) is the output of the time series DRNN neural network prediction model. Then the output of the time series DRNN neural network prediction model is:

2. Empirical Mode Decomposition (EMD) Model Design

Empirical Mode Decomposition (EMD) model is an adaptive signal screening method, which has the characteristics of simple calculation, intuitive, experience-based and adaptive. It can screen out the trends of different characteristics in the historical data information output by the time series DRNN neural network prediction model step by step, and obtain multiple high frequency fluctuation parts (IMF) and low frequency trend parts. The IMF components decomposed by the Empirical Mode Decomposition (EMD) model contain the components of different frequency segments from high to low information. The frequency resolution contained in each frequency segment varies with the information itself, and has the characteristics of adaptive multi-resolution analysis. The purpose of using the empirical mode decomposition (EMD) model is to extract the measurement history data information of the measurement sensor more accurately. The empirical mode decomposition method uses the following steps to "sieve" the historical data output from the time series DRNN neural network prediction model:

① Determine all the local extremum points of the output historical data information of the time series DRNN neural network prediction model, and then connect the left and right local maxima points with a cubic spline to form an upper envelope.

② The lower envelope is formed by connecting the local minimum points of the output historical data information of the cubic spline time series DRNN neural network prediction model, and the upper and lower envelopes should envelop all the data points.

③ The average value of the upper and lower envelopes is recorded as m ₁ (t), and the following is obtained:

x(t)-m ₁ (t)=h ₁ (t) (2)

Among them, x(t) is the original signal of historical data information output by the time series DRNN neural network prediction model. If h ₁ (t) is an IMF, then h ₁ (t) is the first IMF component of x(t). Denote c ₁ (t)=h _1k (t), then c ₁ (t) is the first component of the signal x(t) that satisfies the IMF condition.

④. Separate c ₁ (t) from x(t) to get:

r ₁ (t)=x(t)-c ₁ (t) (3)

Wherein, repeating steps ① to ③ with r ₁ (t) as the original data, to obtain the second component c ₂ of x(t) that satisfies the IMF condition. Repeat the cycle n times to obtain n components of the signal x(t) that satisfy the IMF condition. In this way, the empirical mode decomposition model decomposes the output historical data information of the time series DRNN neural network prediction model into low-frequency trend parts and multiple high-frequency fluctuation parts.

3. Elman neural network model design

The Elman neural network model can dynamically calibrate the measurement value of the predictive measurement sensor. The model is a forward neural network with local memory units and local feedback connections. The association layer receives feedback signals from the hidden layer. Each hidden layer node has a The corresponding association layer nodes are connected. The association layer uses the state of the hidden layer at the previous moment together with the network input at the current moment as the input of the hidden layer as the state feedback. The transfer function of the hidden layer is generally a sigmoid function, and the correlation layer and the output layer are linear functions. Let the number of input layer, output layer and hidden layer of Elman neural network model be m, n and r respectively; w ₁ , w ₂ , w ₃ and w ₄ represent the structure layer unit to the hidden layer, the input layer to the hidden layer, respectively , the connection weight matrix from the hidden layer to the output layer, and the structure layer to the output layer, the output value expressions of the hidden layer, the correlation layer and the output layer of the network are:

c _p (k) = x _p (k-1) (5)

The input data of the Elman neural network model is the actual value data of the low-frequency trend part and multiple high-frequency fluctuation parts output by the empirical mode decomposition (EMD) model at multiple different times over a period of time, and the output is the low-frequency trend part and multiple high-frequency fluctuation parts. The number of input layer, output layer and hidden layer of Elman neural network model is 10, 1 and 21 respectively. Elman neural network model realizes nonlinear prediction of low frequency trend part and multiple high frequency fluctuation parts.

4. NARX neural network model design

The input of multiple NARX neural network models is the output of the corresponding multiple Elman neural network models, and multiple NARX neural network models realize the output of the corresponding multiple Elman neural network models for re-prediction, further improving the accuracy of the measurement sensor calibration value. Spend. NARX neural network model (Nonlinear Auto-Regression with External input neural network) is a dynamic feedforward neural network. Each NARX neural network model is a nonlinear autoregressive network with the output of the corresponding Elman neural network model as input. It has a multi-step delay dynamic characteristic, and connects several layers of the closed network through the NARX neural network model. The NARX neural network model is the most widely used dynamic neural network in nonlinear dynamic systems, and its performance is generally better than that of full Recurrent Neural Network. The NARX neural network model is mainly composed of input layer, hidden layer, output layer and input and output delay. Before application, the delay order of input and output and the number of hidden layer neurons should be determined in advance. The output at that time depends not only on the past output y(t-n), but also on the delay order of the output of the Elman neural network model of the input vector at that time. The NARX neural network model includes input layer, output layer, hidden layer and delay layer. The output of the Elman neural network model is transmitted to the hidden layer through the delay layer. The hidden layer processes the signal output by the Elman neural network model and then transmits it to the output layer. The output layer performs linear weighting on the output signal of the hidden layer to obtain the final NARX neural network model. The output signal, the delay layer delays the signal fed back by the network and the signal output by the input layer, and then sends it to the hidden layer. The NARX neural network model has the characteristics of nonlinear mapping ability, good robustness and adaptability. x(t) represents the external input of the NARX neural network, that is, the output value of the output of the Elman neural network model; m represents the delay order of the external input; y(t) is the output of the NARX neural network model, that is, the NARX of the next period The output control amount of the neural network model; n is the output delay order; s is the number of neurons in the hidden layer; thus, the output of the jth hidden unit can be obtained as:

In the above formula, w _ji is the connection weight between the ith input and the jth hidden neuron, b _j is the bias value of the jth hidden neuron, and the output of the NARX neural network model y(t +1) is:

y(t+1)=f[y(t),y(t-1),…,y(t-n),x(t),x(t-1),…,x(t-m+1) ;W] (8)

The input data of the NARX neural network model is the output of the Elman neural network model, and the output of the NARX neural network model is the future value of the low-frequency trend part and multiple high-frequency fluctuation parts, and the input layer, output layer and hidden layer of the NARX neural network model. The number is 2, 1 and 10 respectively. The NARX neural network model realizes the dynamic fusion of the output of the Elman neural network model and predicts the output of the Elman neural network model again, improving the dynamic performance, rapidity and accuracy of the measured values of the control and measurement sensors. and reliability.

5. Design of wavelet neural network fusion model

The input of the wavelet neural network fusion model is the output values of multiple NARX neural network models. The wavelet neural network fusion model realizes high-precision fusion of the output values of multiple NARX neural network models, and improves the fusion accuracy of the measurement values of the measurement sensors. The wavelet neural network The output value of the fusion model is used as the predicted value of the secondary calibration of the measurement sensor. The wavelet neural network fusion model is based on the theoretical basis of the wavelet neural network WNN (Wavelet Neural Networks) to construct a measurement sensor measurement value calibration and prediction fusion model. The wavelet neural network uses the wavelet function as the excitation function of neurons and combines with artificial neural network. type network. In the wavelet neural network fusion model, the wavelet scaling, translation factor and connection weight are adaptively adjusted in the process of optimizing the error energy function. Let the input of the wavelet neural network fusion model be multiple NARX neural network models. The output value can be expressed as a one-dimensional vector x _i (i=1,2,...,n), and the output signal is the measured sensor measurement value. The predicted fusion value is expressed as y _k (k=1,2,...,m), the calculation formula of the fusion value of the output layer of the wavelet neural network fusion model is:

为小波基函数，b_j为小波基函数的平移因子，a_j小波基函数的伸缩因子，ω_jk为隐含层j节点和输出层k节点间的连接权值。本专利中的小波神经网络融模型的权值和阈值的修正算法采用梯度修正法来更新网络权值和小波基函数参数，从而使小波神经网络融模型输出不断逼近测量传感器测量值的期望输出，小波神经网络融模型的输出为新陈代谢GM(1,1)灰色预测器的输入。In the formula, ω _ij is the connection weight between the input layer i node and the hidden layer j node,

is the wavelet basis function, b _j is the translation factor of the wavelet basis function, a _j is the scaling factor of the wavelet basis function, ω _jk is the connection weight between the hidden layer j node and the output layer k node. The correction algorithm for the weights and thresholds of the wavelet neural network fusion model in this patent uses the gradient correction method to update the network weights and wavelet basis function parameters, so that the output of the wavelet neural network fusion model is constantly approaching the expected output of the measurement sensor. The output of the wavelet neural network fusion model is the input of the metabolic GM(1,1) gray predictor.

6. Design of a gray predictor of metabolic GM(1,1)

The metabolic GM(1,1) gray predictor needs less modeling information, is easy to calculate and has high modeling accuracy, so it has a wide range of applications in various prediction fields. A good effect is obtained, and its output is used as the measurement calibration value of the measurement sensor to realize multiple calibration predictions of the measurement value of the measurement sensor. Metabolism GM(1,1) gray predictor takes the historical data output by the wavelet neural network fusion model as input, and its output is the next stage measurement sensor measurement calibration prediction value. Metabolism GM(1,1) grey predictor is a differential equation established after the historical data output from the wavelet neural network fusion model is generated. The series is generated and then modeled, so the metabolic GM(1,1) gray predictor is actually a generated series model, which is generally described by differential equations. Since the solution of the metabolic GM(1,1) gray predictor is an exponential curve, the solution of the differential equation is an exponential curve, so the generated sequence is required to be increasing and close to an exponential curve. The measurement values of the measurement sensors themselves are all positive values, which will become an incremental sequence after one accumulation and generation. Let the historical data output by the wavelet neural network fusion model be:

x ⁽⁰⁾ =(x ⁽⁰⁾ (1),x ⁽⁰⁾ (2)…x ⁽⁰⁾ (n))(10)

Make a build as:

x ⁽¹⁾ =(x ⁽¹⁾ (1),x ⁽¹⁾ (2)…x ⁽¹⁾ (n))(11)

For x ⁽¹⁾ , the linear differential equation of one variable of first order can be established as follows:

Solving this differential equation yields the measurement sensor calibration predictions:

x ⁽⁰⁾ (k+1)=x ⁽¹⁾ (k+1)-x ⁽¹⁾ (k)(13)

Metabolism GM(1,1) gray predictor must be equidistant, adjacent and must not have jumps. The latest wavelet neural network fusion model output data is used as a reference point to remove the oldest data prediction value. The next stage measures the sensor measurement value. In the measurement sensor measurement calibration prediction, the output value of the latest wavelet neural network fusion model can be used to model, thereby predicting the measurement sensor measurement value in the next stage. After using the above method to predict the measurement value of the measurement sensor in one stage, add the measurement value of the standard measurement sensor into the original sequence, remove a data modeling at the beginning of the sequence, and then predict the measurement value of the measurement sensor in the next stage in the future. Calibrate predictions. And so on, predict the future measurement value of the measurement sensor. Metabolism GM(1,1) gray predictor, which enables longer-term predictions and enables calibrated predictions of measurement sensor measurements.

In this embodiment, the first longitudinal sliding assembly includes a first longitudinal sliding guide rail 201 and a first longitudinal sliding platform 202. The first longitudinal sliding guide rail 201 is fixed on the base platform 1 through a bracket 203, and its lower surface is along the A first screw rod 204 is rotatably connected in the longitudinal direction, a first slider 205 is threadedly connected to the first screw rod 204, and a first longitudinal sliding platform 202 is sleeved on the first longitudinal sliding guide rail 201 and is connected with the first screw The rod 204 is fixedly connected with the first sliding block 205 at the corresponding position, and the structure of the first longitudinal sliding platform 202 is shown in FIG. 3 and FIG. 4 .

The lateral sliding assembly includes a second lead screw 206 that is perpendicular to the first longitudinal sliding guide rail 201 and is rotatably connected to the upper surface of the first longitudinal sliding platform 202. The second lead screw 206 is threadedly connected with a second sliding block 207. A lateral sliding platform 208 is fixed on the two sliding blocks 207 , and the sensor mounting plate 2 is fixed on the lateral sliding platform 208 .

Further, in order to solve the situation that the first longitudinal sliding platform 202 slides on the first longitudinal sliding guide rail 201 and shakes, the two sides of the first longitudinal sliding guide rail 201 are set as inwardly recessed arc guide rails 209, The first longitudinal sliding platform 202 and the arc guide rail 209 are matched at the corresponding positions, so the corresponding positions of the first longitudinal sliding platform 202 and the arc guide rail 209 are provided with arc-shaped protrusions that protrude toward the arc guide rail 209, so that the circle The arc guide rail 209 is matched with the arc-shaped protrusion. When the first longitudinal sliding platform 202 slides on the first longitudinal sliding guide rail 201, its upper and lower positions are limited, and there will be no sliding and inclination, and it is difficult to A situation in which the first longitudinal sliding platform 202 shakes occurs.

Further, in order to reduce the frictional force between the first longitudinal sliding platform 202 and the first longitudinal sliding guide rail 201, a number of guide rail balls are also rolled on the first longitudinal sliding platform 202 at the matching position with the arc guide rail 209. 210. In this way, when the first longitudinal sliding platform 202 slides on the first longitudinal sliding guide rail 201, the guide rail balls 210 can reduce the friction force between the two.

Further, in order to more stabilize the first longitudinal sliding platform 202, a pair of longitudinal linear guide rails 211 are also provided on the upper surface of the first longitudinal sliding guide rail 201 along the sliding direction of the first longitudinal sliding platform 202. A pair of bar-shaped protrusions 212 are provided at the corresponding positions of the platform 202 and the longitudinal linear guide rails 211 , each bar-shaped protrusion 212 is matched with each longitudinal linear guide rail 211 , and the first longitudinal sliding platform 202 slides on the first longitudinal sliding guide rail 201 At the time, the bar-shaped protrusions 212 on the lower surface of the first longitudinal sliding platform 202 are constrained to slide in the longitudinal linear guide rails 211, and the first longitudinal sliding platform 202 is constrained to the left and right, and will not tilt.

Further, when adjusting the distance between the sensor and the reference object, the adjustment of the front and rear distance sometimes needs to be fine-tuned. The adjustment pitch of the first longitudinal sliding component is relatively large, and sometimes it is difficult to achieve the required adjustment accuracy. Therefore, we also set a second longitudinal sliding assembly on the lateral sliding assembly. The second longitudinal sliding assembly includes a third screw 213 that is rotatably connected to the lateral sliding platform 208 and is perpendicular to the second screw 206. A third sliding block 214 is threadedly connected to the third screw rod 213 , and the pitch of the third screw rod 213 is smaller than that of the first screw rod 204 . The sensor mounting plate 2 is fixed on the lateral sliding platform 208 .

Further, in order to realize that the first longitudinal sliding platform 202, the lateral sliding platform 208 and the sensor mounting plate 2 slide on the corresponding sliding rails to adjust the distance between the longitudinal and lateral directions of the sensor, the first longitudinal sliding assembly, the lateral sliding Both the sliding assembly and the second longitudinal sliding assembly are provided with driving mechanisms with the same structure. One end of the second screw 206 and the third screw 213 is provided with a main bevel gear 218 and a slave bevel gear 219 with the same structure, and the three output shafts of the stepping motor are fixedly connected to the center of the corresponding main bevel gear 218, The main bevel gear 218 meshes with its corresponding slave bevel gear 219, and the slave bevel gear 219 is sleeved and fixed on its corresponding screw rod.

The stepping motor drives the main bevel gear 218 to rotate, and the main bevel gear 218 drives the secondary bevel gear 219 to rotate to realize the rotation of the lead screw. The first longitudinal sliding platform 202 , the lateral sliding platform 208 and the sensor mounting plate 2 can slide on the corresponding first screw rods 204 , second screw rods 206 and third screw rods 213 .

Further, if the first longitudinal sliding platform 202, the lateral sliding platform 208 and the sensor mounting plate 2 are not electrically controlled to slide on the corresponding sliding rails to adjust the distance between the front, rear, left and right of the sensor, the first screw rod 204, One end of the second screw rod 206 and the third screw rod 213 are respectively fixed with a rotating handle 220. The rotation of the corresponding screw rod is realized by rotating the handle 220, thereby realizing the first longitudinal sliding platform 202, the lateral sliding platform 208 and the sensor mounting plate 2. Slide on the corresponding lead screw.

working principle:

First, the measuring sensor is fixed on the sensor mounting plate 2 through a tripod, the reference object is set on the reference object mounting platform 3, and the measuring sensor and the For the distance between the reference objects, move the measurement sensor to the set position according to the preset needs, first start the first step motor 215 to drive the first screw rod 204 to rotate to initially adjust the longitudinal distance between the reference object and the measurement sensor; Start the second stepper motor 216 to drive the second screw rod 204 to rotate to adjust the lateral distance between the measurement sensor and the reference object; finally start the third stepper motor 217 to drive the third screw rod 213 to rotate to fine-tune the longitudinal direction between the reference object and the measurement sensor Distance, to achieve precise positioning of the measurement sensor position.

When the measurement sensor measures a certain point parameter of the reference object, the measurement value of the measurement sensor is calibrated and predicted through the calibration algorithm in the monitoring unit of the MSP430 single-chip microcomputer, so that the measurement value of the measurement sensor is closer to the accurate parameter value of the reference object.

The above-mentioned embodiments are only intended to illustrate the technical concept and characteristics of the present invention, and the purpose is to enable those who are familiar with the art to understand the content of the present invention and implement it accordingly, and cannot limit the protection scope of the present invention. All equivalent transformations or modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.

Claims (10)

1. A parameter calibration system for measurement, comprising a reference object mounting table (3), a sensor mounting plate (2) and a base platform (1), the reference object mounting table (3), the sensor mounting plate (2) ) is arranged on the base-level platform (1), characterized in that, the base-level platform (1) is further provided with a first longitudinal sliding component and a lateral sliding component; the sensor mounting plate (2) is arranged on the On the lateral sliding assembly, the sensor mounting plate (2) slides laterally with the lateral sliding assembly, and the lateral sliding assembly slides longitudinally with the first longitudinal sliding assembly; the sensor mounting plate The board (2) is also provided with a measurement sensor and an MSP430 single-chip monitoring unit. The MSP430 single-chip monitoring unit includes a time-series DRNN neural network prediction model, an empirical mode decomposition (EMD) model, multiple Elman neural network models, and multiple NARX neural network models. network model, wavelet neural network fusion model and metabolic GM(1,1) gray predictor; the output of the measurement sensor is used as the input of the time series DRNN neural network prediction model, and the output of the time series DRNN neural network prediction model is used as the empirical mode The input of the decomposition (EMD) model, the output of the empirical mode decomposition (EMD) model measures the low-frequency trend part and multiple high-frequency fluctuation parts of the sensor as the input of the corresponding multiple Elman neural network models, respectively. The output is used as the input of the corresponding multiple NARX neural network models, the output of multiple NARX neural network models is used as the input of the wavelet neural network fusion model, and the output of the wavelet neural network fusion model is used as the output of the metabolic GM (1,1) gray predictor. Input, the output of the metabolic GM(1,1) gray predictor serves as the measurement calibration value of the measurement sensor. 2. A parameter calibration system for measurement according to claim 1, wherein the first longitudinal sliding assembly comprises a first longitudinal sliding guide rail (201) and a first longitudinal sliding platform (202) , the first longitudinal sliding guide rail (201) is fixed on the base platform (1) through a bracket (203), and a first screw rod (204) is rotatably connected to its lower surface along its length direction. A first sliding block (205) is threadedly connected to the screw rod (204), and the first longitudinal sliding platform (202) is sleeved on the first longitudinal sliding guide rail (201) and is connected to the first longitudinal sliding rail (201). The corresponding position of the lead screw (204) is fixedly connected with the first sliding block (205). 3 . The parameter calibration system for measurement according to claim 2 , wherein the lateral sliding component comprises a vertical sliding guide rail ( 201 ) that is perpendicular to the first longitudinal sliding and is rotatably connected to the first A second lead screw (206) on the upper surface of the longitudinal sliding platform (202), the second lead screw (206) is threadedly connected with a second slider (207), and the second slider (207) is fixed on There is a lateral sliding platform (208), and the sensor mounting plate (2) is fixed on the lateral sliding platform (208). 4. A parameter calibration system for measurement according to claim 2, characterized in that, both sides of the first longitudinal sliding guide rail (201) are set as inwardly recessed arc guide rails (209), The first longitudinal sliding platform (202) and the arc guide rail (209) are arranged in a corresponding position matching. 5 . The parameter calibration system for measurement according to claim 4 , wherein the first longitudinal sliding platform ( 202 ) and the arc guide rail ( 209 ) are also provided with a number of rolling rail balls (210). 6. A parameter calibration system for measurement according to claim 2, characterized in that the upper surface of the first longitudinal sliding guide rail (201) slides along the first longitudinal sliding platform (202) A pair of longitudinal linear guide rails (211) are also provided in the direction, and a pair of bar-shaped protrusions (212) are provided at the corresponding positions of the first longitudinal sliding platform (202) and the longitudinal linear guide rails (211). The protrusions (212) are matched with each of the longitudinal linear guide rails (211). 7 . The parameter calibration system for measurement according to claim 3 , wherein the lateral sliding assembly is further provided with a second longitudinal sliding assembly, and the second longitudinal sliding assembly comprises a rotational connection. 8 . A third lead screw (213) of the lateral sliding platform (208) that is perpendicular to the second lead screw (206), and a third slider (214) is threadedly connected to the third lead screw (213). ), the pitch of the third screw (213) is smaller than the pitch of the first screw (204), and the sensor mounting plate (2) is fixed on the lateral sliding platform (208). 8 . The parameter calibration system for measurement according to claim 7 , wherein the first longitudinal sliding assembly, the lateral sliding assembly and the second longitudinal sliding assembly are all provided with drives with the same structure. 9 . The driving mechanism is a first stepper motor (215), a second stepper motor (216) and a third stepper motor (217), the first screw rod (204), the second screw rod ( 206) and one end of the third screw (213) are provided with a main bevel gear (218) and a follower bevel gear (219) with the same structure, and the three stepper motor output shafts are all corresponding to the main bevel gear (218) ) are fixedly connected at the center, the main bevel gear (218) is meshed with the corresponding secondary bevel gear (219), and the secondary bevel gear (219) is sleeved and fixed on its corresponding lead screw. 9 . The parameter calibration system for measurement according to claim 7 , wherein one end of the first screw rod ( 204 ), the second screw rod ( 206 ) and the third screw rod ( 213 ) are respectively fixed at one end. 10 . There is a swivel handle (220). 10. A parameter calibration system for measurement according to any one of claims 1 to 9, wherein the reference object mounting platform (3) further comprises a pad height fixed on the base platform (1) A table (101), a plurality of vertically arranged electric push rods (102) fixed on the top of the booster fixing table (101), the reference object mounting table (3) is provided at the top of the electric push rods (102) ).

Images