Dynamic measurement uncertainty quantification method based on physical information and evidence learning
By constructing four feedforward neural networks based on physical information and evidence learning, the problems of generalization ability and physical interpretability of deep neural networks in dynamic measurement uncertainty assessment are solved, and more efficient uncertainty quantification is achieved.
Patent Information
- Application Number
- CN202411157262.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-22
AI Technical Summary
Existing deep neural networks suffer from poor generalization ability and physical interpretability in dynamic measurement uncertainty assessment, and the uncertainty change trend is the same as that of the model input.
Based on physical information and evidence learning, this method establishes the partial differential equations of the target dynamic system, constructs four independent feedforward neural networks, trains them using training and test datasets, and optimizes network parameters and quantifies uncertainty by combining evidence regression theory and data-driven loss functions.
It improves the generalization ability and physical interpretability of neural networks, avoids dependence on the trend of uncertainty changes, and enhances the efficiency and accuracy of uncertainty modeling.
Smart Images

Figure CN119004033A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of dynamic measurement uncertainty estimation methods, and in particular to a dynamic measurement uncertainty quantification method based on physical information and evidence learning. Background Art
[0002] Dynamic measurement uncertainty is an important factor of measurement uncertainty, which itself depends on the dynamic characteristics of the measurement system and the frequency spectrum of the input signal. With the increasing importance of dynamic measurements, the evaluation of dynamic measurement uncertainty is considered to be an important scientific issue in metrology.
[0003] The GUM series of documents, "Guide to the Expression of Uncertainty in Measurement", established an uncertainty expression system, and a complete measurement framework for the evaluation of static measurement uncertainty has been established. However, they do not include reasonable estimates of measurement uncertainty in certain special cases, especially for dynamic measurements. Since the system state of a dynamic system changes over time, the measured quantity also changes over time, so the conditions for using GUM as a method for evaluating uncertainty in a dynamic system are no longer met.
[0004] At present, there are several different emerging mathematical methods for dynamic measurement uncertainty assessment. The more commonly used methods based on dynamic measurement error theory are the Monte Carlo method and the grey system theory. The Monte Carlo method is an improved statistical simulation method based on Bayesian statistical theory, which evaluates the uncertainty of dynamic measurement by simulating the dynamic measurement process through random sampling. This method requires a large number of samples to estimate the distribution of the measured value or the distribution of the measured value is known. The huge amount of calculation and the uncertainty introduced by repeated sampling increase the overall uncertainty of the final result. The applicability of the grey system theory depends on the correctness of the distribution of the sample data.
[0005] In recent years, deep neural networks have been widely used in the field of system modeling and identification due to their excellent nonlinear feature extraction and complex model recognition capabilities. Existing deep neural network uncertainty assessment methods mainly include network integration methods, Bayesian methods, test time data enhancement methods, single deterministic methods, etc. However, when these existing neural network uncertainty assessment methods are used in dynamic systems, there are problems with poor generalization ability and physical interpretability, and due to the dependence of uncertainty propagation on model input, it will also lead to the problem that the final uncertainty change trend is the same as the model input. Summary of the invention
[0006] The present invention provides a dynamic measurement uncertainty quantification method based on physical information and evidence learning to solve the problems of poor generalization ability and physical interpretability, and the same uncertainty change trend as model input in the existing deep neural network uncertainty assessment method in the prior art.
[0007] In order to achieve the above object, the technical solution adopted by the present invention is:
[0008] The dynamic measurement uncertainty quantification method based on physical information and evidence learning includes the following steps:
[0009] Step 1: Based on the structure and physical information of the target dynamic system, a partial differential equation of the output result of the target dynamic system is established, and an objective function for constraining the output result of the subsequent neural network is obtained according to the partial differential equation;
[0010] Furthermore, a simulation experiment is performed on the target dynamic system, and a training data set is constructed using an input signal input into the target dynamic system during the simulation experiment and a corresponding output signal of the target dynamic system;
[0011] Furthermore, a real experiment is conducted on the target dynamic system, and a test data set is constructed using input signals input into the target dynamic system during the real experiment and corresponding output signals of the target dynamic system;
[0012] Step 2: Based on the method of quantifying uncertainty based on evidence theory, four independent feedforward neural networks are constructed and generated, and then each feedforward neural network is trained using the training data set constructed in step 1. When the training data set is input into each feedforward neural network, each feedforward neural network outputs a tensor of predicted values and a tensor of evidence distribution parameters, and the dimensions of the tensors output by the four feedforward neural networks are the same;
[0013] In addition, a data-driven loss function is designed for each feedforward neural network training based on the theory of evidence regression, and a partial differential loss is constructed by minimizing the mean square error loss of the objective function obtained in step 1. The total loss function for each feedforward neural network training is constructed by the data-driven loss function and the partial differential loss. When each feedforward neural network is trained, the network parameters are updated and optimized based on the calculation results of the total loss function, thereby obtaining a dynamic system model consisting of four trained feedforward neural networks.
[0014] Step 3: Input the test data set obtained in step 1 into the dynamic system model obtained in step 2, and calculate the total uncertainty of the prediction results of the dynamic system model based on the tensor of the evidence distribution output by the four trained feedforward neural networks in the dynamic system model.
[0015] In a further step 1, the input signals constituting the test data set are not contained in the input signals constituting the training data set.
[0016] In a further step 1, the signal lengths in the test dataset and the training dataset are kept consistent.
[0017] In the further step 1, for the long sequence input signals in the training data set and the test data set, the time dimension slices of the long sequence input signals are reconstructed, and the one-dimensional long sequence input signals with a length of N are reconstructed into a tensor with a shape of (1×H×W).
[0018] In the further step 2, the four feedforward neural networks output a total of four tensors of the same dimension of the evidence distribution, which are used to approximate the output signal mean μ and variance σ respectively. 2 The distribution parameters of the distribution are as follows: the mean μ follows a normal distribution and the variance σ 2 It follows a normal inverse gamma distribution.
[0019] In the further step 2, the data-driven loss function includes three losses, where:
[0020] Estimate the predictive distribution of the network output by minimizing the negative log marginal likelihood function NLL as the first part of the loss;
[0021] The second part of the loss uses an evidence regularizer to adjust the first part of the loss for training by increasing the prediction uncertainty where the prediction is wrong;
[0022] The third part of the loss is the MSE loss function, which is used to improve the prediction accuracy of the neural network and uses the gradient descent method for optimization approximation.
[0023] In the further step 3, the total uncertainty calculated includes epistemic uncertainty and arbitrary uncertainty.
[0024] The present invention applies the powerful modeling and uncertainty quantification capabilities of feedforward neural networks to dynamic systems, and integrates the physical information of dynamic systems, thereby improving the generalization and physical interpretability of the model and filling the gap in the use of neural network modeling of dynamic system uncertainty. Compared with the prior art, the present invention has the following advantages:
[0025] 1. The present invention uses the physical information of the dynamic system as a partial constraint to guide the neural network modeling, and integrates it with the loss function based on uncertainty evidence-aware learning, which supplements the missing knowledge and improves the generalization ability and physical interpretability of the neural network.
[0026] 2. The present invention uses the method of deep evidence regression to quantify the uncertainty of the estimated value of the neural network output, avoiding the situation in traditional methods where the final uncertainty change trend is the same as the model input due to the dependence of uncertainty propagation on the model input, and improves the efficiency of uncertainty modeling and quantification. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 It is a schematic diagram of the method of an embodiment of the present invention.
[0028] Figure 2 1 and 1. These are output result diagrams of the simulation model and the actual experiment obtained by five groups of voltage input signals at a frequency of 20 Hz according to an embodiment of the present invention, wherein (a) is the output result diagram of the simulation model, and (b) is the output result diagram of the actual experiment.
[0029] Figure 3 It is a performance diagram of the dynamic system model trained by the embodiment of the present invention on the test data set.
[0030] Figure 4 It is an error and uncertainty diagram of the prediction result of the dynamic system model trained in the embodiment of the present invention. DETAILED DESCRIPTION
[0031] The present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0032] like Figure 1 As shown, this embodiment discloses a dynamic measurement uncertainty quantification method based on physical information and evidence learning, comprising the following steps:
[0033] Step 1: Based on the structure and physical information of the target dynamic system, a partial differential equation of the output result of the target dynamic system is established, and the objective function used to constrain the output result of the subsequent neural network is obtained according to the partial differential equation.
[0034] The target dynamic system in this embodiment can be a nonlinear system or a linear system. Taking the piezoelectric ceramic drive system as an example, the piezoelectric ceramic drive system is approximated as a spring-mass-damper system according to the structure and physical information of the piezoelectric ceramic drive system. The partial differential equation of the output result of the piezoelectric ceramic drive system is shown in the following formula:
[0035]
[0036] Among them, y(x, t) is the output signal of the dynamic system, which is determined by the input signal x and time t; c and k are the damping coefficient and the elastic coefficient of the spring respectively.
[0037] The partial differential equation shown in the above formula is expressed as the objective function used to constrain the output results of the subsequent neural network, and y(x, t) is used as the predicted output of the network, as shown in the following formula:
[0038]
[0039] Among them, g(x, t) is the objective function of establishing the network model.
[0040] Furthermore, in this embodiment, the Simulink toolbox of MATLAB is used to perform a simulation experiment on the target dynamic system, and a training data set is constructed by inputting an input signal of the target dynamic system during the simulation experiment and a corresponding output signal of the target dynamic system.
[0041] Specifically, the Simulink toolbox of MATLAB is used to establish a simulation model of the target dynamic system. K groups of dynamic time-related signals are input into the simulation model in batches. The duration of each group of dynamic time-related signals is t, and the frequencies of each group of dynamic time-related signals are different. Each group of dynamic time-related signals contains M input signals X1, X2, X3...X with the same frequency and different amplitudes. M , each input signal X of the dynamic measurement system i (i=1, 2, ...M) are all functions x of time t t (t=1, 2, ...N).
[0042] Each group of dynamic time-correlated signals contains the input signal X i Under this condition, the simulation model of the target dynamic system generates M corresponding output signals Y1, Y2, Y3...Y M , each output signal Y i are all functions y of time t t (t=1,2,..,.N). The function y at each time t t All obey the normal distribution, and the probability density function is η t (t=1,2,...,N), the joint probability density function of the entire time series is encoded as Then the probability distribution P(Y|X) of the output sequence of the simulation model of the dynamic system is as follows:
[0043]
[0044] Each group of dynamic time-correlated signals is matched with the output signal of the corresponding target dynamic system simulation model as a training data set.
[0045] Taking the piezoelectric ceramic drive system as the target dynamic system, 10 groups (i.e. K = 10) of sawtooth wave voltage signals with different frequencies (excluding 20 Hz) from 5 Hz to 40 Hz are used as dynamic time-related signals. Each group of sawtooth wave voltage signals has 40 (i.e. M = 40) input voltages with different amplitudes (i.e. input signal X i ), the amplitude increases from 30V to 120V, and 10 (ie N = 10) cycles of signals are taken at each amplitude (ie x t), the duty cycle is 0.9999, 10 groups of sawtooth wave voltage signals are input into the simulation model of the piezoelectric ceramic drive system, the output signal is obtained by the simulation model of the piezoelectric ceramic drive system, and the 10 groups of sawtooth wave voltage signals and the corresponding output signals are used as training data sets.
[0046] In addition, this embodiment also conducts a real experiment on the target dynamic system, and constructs a test data set with the input signal input into the target dynamic system during the real experiment and the corresponding output signal of the target dynamic system. The signal length of the test set data is consistent with the signal length of the training set data, but the input signal constituting the test data set is not included in the input signal constituting the training data set.
[0047] Taking the piezoelectric ceramic drive system as the target dynamic system, in the real experiment of the piezoelectric ceramic drive system, 10 sawtooth voltages with different voltage amplitudes at a frequency of 20 Hz that are not included in the input signal of the simulation model are used as input signals to obtain the corresponding output signals. The sawtooth voltages with 10 different voltage amplitudes at a frequency of 20 Hz and the corresponding output signals are used as the test set. In order to compare the performance of the network model in the test set, the output signals of the simulation model and the real experiment at 20 Hz are shown in the figure. Figure 2 shown.
[0048] After obtaining the training data set and the test data set, the signals in the training data set and the test data set are preprocessed to adjust each set of input and the corresponding output signal to the same dimension and size, and the long sequence input signal is sliced and reconstructed in the time dimension, and the one-dimensional long sequence input signal with a length of N is reconstructed into a tensor with a shape of (1×H×W).
[0049] Taking the piezoelectric ceramic drive system as the target dynamic system as an example, in the training data set and the test data set, the original input signal is an ultra-long time series signal x t (t=1,2,...,N,N=2000),the input dimension is changed by slicing and reconstructing in the time dimension, and the one-dimensional long sequence input signal is reconstructed into a tensor with a shape of (1×250×8). The subsequent neural network inputs the signal value of 8 time points each time until the signal ends, thus avoiding the huge amount of calculation caused by single time point processing and improving the operation speed.
[0050] Step 2: Based on the method of quantifying uncertainty using evidence theory, four independent feedforward neural networks are constructed. Each feedforward network consists of four hidden layers, each containing 256 neurons, and the output layer is activated with the softplus function to limit the range of the output results to positive numbers. Then, each feedforward neural network is trained using the training data set constructed in step 1 to obtain a dynamic system model consisting of four trained feedforward neural networks, thereby completing the modeling of the measurement process of the dynamic system.
[0051] When the training data set is input into each feedforward neural network, each feedforward neural network outputs a tensor of predicted value and evidence distribution, and the dimensions of the tensors output by the four feedforward neural networks are the same. The four feedforward neural networks output a total of four tensors of the same dimension of evidence distribution, which are used to approximate the output signal mean μ and variance σ respectively. 2 The distribution parameters of , where the mean μ follows a normal distribution Variance σ 2 Follows normal inverse gamma distribution (NIG) Γ -1 (α,β), is the conjugate prior of the normal distribution.
[0052] In this embodiment, the total loss function L during training of each feedforward neural network includes the data-driven loss function L data and the partial differential loss L pde , where the data-driven loss function L data It is designed for each feedforward neural network based on the theory of evidence regression. When training each feedforward neural network, the network parameters are updated and optimized based on the calculation results of the total loss function, thereby obtaining a trained feedforward neural network.
[0053] Among them, the data-driven loss function L in the total loss function L data The loss consists of the following three parts:
[0054] A. The parameter m = {γ,υ,α,β} of the output evidence distribution of each feedforward neural network, then the conditional probability distribution p(y|m) of the predicted value y output by each feedforward neural network is as follows:
[0055]
[0056] Among them: γ,υ,α,β represent mean μ and variance σ 2 The conditional probability distribution p(y|m) follows the Student t distribution, as shown below:
[0057]
[0058] By minimizing the negative log marginal likelihood function NLL, as the data-driven loss function L dataThe first part of the loss function L NLL , as shown below:
[0059]
[0060] Wherein, Ω=2β(1+υ); Γ() represents the gamma distribution.
[0061] B. Data-driven loss function L data The second part of the loss is the evidence regularizer L R , the NLL loss of the first part is adjusted for training by increasing the prediction uncertainty where the prediction error occurs, as shown in the following formula:
[0062] L R =|γ-y|(2υ+α)
[0063] C. Data-driven loss function L data The third part of the loss is the mean square error loss L MSE During training, the prediction accuracy of the feedforward neural network is improved based on the mean square error loss, and then the gradient descent method is used for optimization approximation. The calculation formula is:
[0064]
[0065] Where N pde is the total number of PDE sampling points.
[0066] Therefore, the data-driven loss function L consists of three parts of loss data As shown below:
[0067]
[0068] Among them, λ R is the regularization parameter; N data is the total number of training data points.
[0069] Partial differential loss L in the total loss function pde , which is constructed by minimizing the mean square error (MSE) loss of the objective function g(x, t) obtained in step 1. In the specific calculation, the partial differential of y(x, t) is calculated through the automatic differentiation technology of PyTorch and then inserted into g(x, t).
[0070] In this embodiment, the total loss function L during the training of each feedforward neural network is the data-driven loss function L data and the partial differential loss L pde The weighted sum of is as follows:
[0071] L=λ data L data +λ pde Lpde
[0072] Among them, λ data and λ pde are the weight coefficients for balancing the interaction between the two losses.
[0073] Step 3: Input the test data set obtained in step 1 into the dynamic system model consisting of the four trained feedforward neural networks obtained in step 2, and the dynamic system model outputs the prediction results and four sets of evidence distribution tensors, such as Figure 3 shown.
[0074] The variance of the evidence distribution is calculated based on the tensor of the evidence distribution output by the four trained feedforward neural networks, and then the epistemic uncertainty U of the prediction results of the dynamic system model is evaluated based on the variance. epi , arbitrary uncertainty U ale ,like Figure 4 shown.
[0075] Among them, epistemic uncertainty U epi As shown below:
[0076]
[0077] Where: Var[μ] represents the variance of the mean μ, that is, the epistemic uncertainty; p(μ) represents the distribution of the mean μ; E[μ] represents the expectation of the mean μ; μ represents the mean of the output value; α, β, υ are the parameters of the evidence distribution.
[0078] Arbitrary uncertainty U ale As shown below:
[0079]
[0080] Where: E[σ 2 ] represents the variance σ 2 The mean of p(σ 2 ) represents the variance σ 2 distribution; σ represents the standard deviation of the output value; α, β, υ are the parameters of the evidence distribution.
[0081] The total uncertainty U of the prediction results of the dynamic system model is the epistemic uncertainty U epi , arbitrary uncertainty U ale The sum is as follows:
[0082] U=U epi +U ale .
[0083] Although there are significant differences between the uncertainty quantification based on deep neural networks and the uncertainty assessment method of GUM, and there is no clear uncertainty propagation process, compared with traditional nonlinear dynamic system modeling, deep neural networks have good recognition and modeling capabilities for complex input-output mapping relationships, and the integration of physical information improves the interpretability and generalization ability of neural networks.
[0084] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. The embodiments described in the present invention are merely descriptions of the preferred embodiments of the present invention, and do not limit the concept and scope of the present invention. The various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction, and such combinations should also be regarded as the contents disclosed in the present disclosure as long as they do not violate the concept of the present invention. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.
[0085] The present invention is not limited to the specific details of the above-mentioned embodiments. Within the technical concept of the present invention and without departing from the design concept of the present invention, various modifications and improvements made to the technical solution of the present invention by technical personnel in this field should fall within the protection scope of the present invention. The technical contents for which protection is sought in the present invention have been fully recorded in the claims.
Claims
1. A dynamic measurement uncertainty quantification method based on physical information and evidence learning, characterized by: The following steps are involved: Step 1: Based on the structure and physical information of the target dynamic system, a partial differential equation of the output result of the target dynamic system is established, and an objective function for constraining the output result of the subsequent neural network is obtained according to the partial differential equation; Furthermore, a simulation experiment is performed on the target dynamic system, and a training data set is constructed using an input signal input into the target dynamic system during the simulation experiment and a corresponding output signal of the target dynamic system; Furthermore, a real experiment is conducted on the target dynamic system, and a test data set is constructed using input signals input into the target dynamic system during the real experiment and corresponding output signals of the target dynamic system; Step 2: Based on the method of quantifying uncertainty based on evidence theory, four independent feedforward neural networks are constructed and generated, and then each feedforward neural network is trained using the training data set constructed in step 1. When the training data set is input into each feedforward neural network, each feedforward neural network outputs a tensor of predicted values and a tensor of evidence distribution parameters, and the dimensions of the tensors output by the four feedforward neural networks are the same; In addition, a data-driven loss function is designed for each feedforward neural network training based on the theory of evidence regression, and a partial differential loss is constructed by minimizing the mean square error loss of the objective function obtained in step 1. The total loss function for each feedforward neural network training is constructed by the data-driven loss function and the partial differential loss. When each feedforward neural network is trained, the network parameters are updated and optimized based on the calculation results of the total loss function, thereby obtaining a dynamic system model consisting of four trained feedforward neural networks. Step 3: Input the test data set obtained in step 1 into the dynamic system model obtained in step 2, and calculate the total uncertainty of the prediction results of the dynamic system model based on the tensor of the evidence distribution output by the four trained feedforward neural networks in the dynamic system model.
2. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1 is characterized in that: In step 1, the input signal constituting the test data set is not included in the input signal constituting the training data set.
3. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1 is characterized in that: In step 1, the signal lengths in the test dataset and the training dataset are kept consistent.
4. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1 is characterized in that: In step 1, for the long sequence input signals in the training data set and the test data set, the time dimension slices are reconstructed for the long sequence input signals, and the one-dimensional long sequence input signals with a length of N are reconstructed into a tensor with a shape of (1×H×W).
5. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1, characterized in that: In step 2, the four feedforward neural networks output a total of four tensors of the same dimension for the four evidence distributions, which are used to approximate the output signal mean μ and variance σ respectively. 2 The distribution parameters of the distribution are as follows: the mean μ follows a normal distribution and the variance σ 2 It follows a normal inverse gamma distribution.
6. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1, characterized in that: In step 2, the data-driven loss function includes three parts of loss, where: Estimate the predictive distribution of the network output by minimizing the negative log marginal likelihood function NLL as the first part of the loss; The second part of the loss uses an evidence regularizer to adjust the first part of the loss for training by increasing the prediction uncertainty where the prediction is wrong; The third part of the loss is the MSE loss function, which is used to improve the prediction accuracy of the neural network and uses the gradient descent method for optimization approximation.
7. The method for quantifying dynamic measurement uncertainty based on physical information and evidence learning according to claim 1, characterized in that: In step 3, the calculated total uncertainty includes epistemic uncertainty and arbitrary uncertainty.
Citation Information
Patent Citations
Text entity extraction method and system based on self-training semi-supervised learning
CN115270797A
Apparatus and method for model quality estimation and model adaptation in multivariable process control
US20110130850A1
Cited By
Inertia and damping online evaluation method, device and system of power system
CN121456741A
