Power equipment multi-source heterogeneous perception data completion method and device based on Kalman filtering model, computer equipment, storage medium and computer program product

Abstract

The invention relates to a Kalman filtering model-based power equipment multi-source heterogeneous perception data completion method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of obtaining a sensing data set corresponding to power equipment; based on the perception data set, establishing a Kalman filtering model for the power equipment; performing parameter estimation on the Kalman filtering model by adopting an expectation maximization algorithm to obtain a parameter estimation result; performing Kalman filtering recursion based on the parameter estimation result to obtain a sensing data prediction result corresponding to the power equipment; on the basis of the perception data prediction result, performing data completion on the perception data set to obtain a completed perception data set; the complemented sensing data set is used for correcting the digital twinborn body of the power equipment. By adopting the method, data completion can be accurately and efficiently carried out.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and in particular to a method, device, computer equipment, storage medium and computer program product for completing multi-source heterogeneous perception data of power equipment based on a Kalman filter model. Background Art

[0002] As an emerging technical means, digital twin technology can achieve remote monitoring, fault diagnosis and performance optimization of power equipment through digital modeling of physical entities and real-time data fusion.

[0003] During the long-term operation of power equipment, its operating status and performance will change and evolve, which requires the digital twin model to have adaptive correction capabilities to maintain consistency and accuracy with the actual power equipment. The existing digital twin model adaptive correction method needs to be updated based on the actual operating data of the actual power equipment. The real-time operating data of the actual power equipment often contains certain abnormal data. Therefore, it is necessary to remove the actual operating data of the actual power equipment and complete the data after removing the data. However, the existing data completion methods such as mean completion method, polynomial completion method and nearest neighbor completion method, although they can supplement the missing data to a certain extent, have certain limitations. For example, the mean completion method ignores the fluctuation characteristics and physical meaning of the data, the polynomial completion method may produce certain errors when the data fluctuates greatly, and the nearest neighbor completion method may be affected by the data distribution when selecting neighboring data, resulting in inaccurate completion results.

[0004] Therefore, there is a problem in the traditional technology that data cannot be completed accurately. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for multi-source heterogeneous perception data completion of power equipment based on a Kalman filter model that can accurately complete data in response to the above-mentioned technical problems.

[0006] A method for completing multi-source heterogeneous sensing data of power equipment based on a Kalman filter model, the method comprising:

[0007] Acquire a perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing abnormal data in the original perception data set of the power equipment; the original perception data set includes original perception data of the power equipment in multiple dimensions collected by different sensors;

[0008] Based on the perception data set, a Kalman filter model for power equipment is established;

[0009] The parameter estimation of the Kalman filter model is carried out by using the expectation maximization algorithm to obtain the parameter estimation result;

[0010] Based on the parameter estimation result, Kalman filter recursion is carried out to obtain the predicted result of the sensed data corresponding to the power equipment;

[0011] Based on the predicted result of the sensed data, the sensed data set is completed to obtain the completed sensed data set; the completed sensed data set is used to correct the digital twin of the power equipment.

[0012] In one embodiment, a Kalman filter model for power equipment is established based on the sensed data set, including:

[0013] Based on the sensed data set, a state transition model and an observation model are constructed; the state transition model characterizes the law of change of the equipment state of the power equipment; the observation model characterizes the relationship between the equipment state of the power equipment and the sensed data;

[0014] Based on the state transition model and the observation model, a Kalman filter model is formed.

[0015] In one embodiment, the parameter estimation of the Kalman filter model is carried out by using the expectation maximization algorithm to obtain the parameter estimation result, including:

[0016] The Kriging model is used to construct the spatial field of the Kalman filter model;

[0017] The parameter estimation of the Kalman filter model is carried out by using the expectation maximization algorithm based on the spatial field to obtain the parameter estimation result.

[0018] In one embodiment, the sensed data set corresponding to the power equipment is obtained, including:

[0019] The original sensed data set is obtained;

[0020] The abnormal data in the original sensed data set is replaced to obtain the sensed data set.

[0021] In one embodiment, the abnormal data in the original sensed data set is replaced to obtain the sensed data set, including:

[0022] The abnormal data in the original sensed data set is removed by using the preset data rejection rule to obtain the processed sensed data set;

[0023] The null values in the processed sensed data set are replaced by using the preset data to obtain the sensed data set.

[0024] In one embodiment, the original perception data set includes current data, voltage data, temperature data, humidity data, device status data, and working mode data of the power equipment at each time point.

[0025] A multi-source heterogeneous perception data completion device for power equipment based on a Kalman filter model, the device includes:

[0026] An acquisition module for acquiring the perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors;

[0027] A building module for building a Kalman filter model for the power equipment based on the perception data set;

[0028] An estimation module for estimating the parameters of the Kalman filter model by using the expectation maximization algorithm to obtain a parameter estimation result;

[0029] A prediction module for performing Kalman filter recursion based on the parameter estimation result to obtain a perception data prediction result corresponding to the power equipment;

[0030] A completion module for completing the perception data set based on the perception data prediction result to obtain a completed perception data set; the completed perception data set is used to correct the digital twin of the power equipment.

[0031] A computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the above method are implemented.

[0032] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0033] A computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0034] The above-mentioned method, device, computer equipment, storage medium and computer program product for complementing multi-source heterogeneous perception data of power equipment based on the Kalman filter model obtain the perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors; based on the perception data set, a Kalman filter model for the power equipment is established; the expectation maximization algorithm is used to estimate the parameters of the Kalman filter model to obtain the parameter estimation result; Kalman filter recursion is performed based on the parameter estimation result to obtain the perception data prediction result corresponding to the power equipment; based on the perception data prediction result, the perception data set is complemented to obtain the complemented perception data set; the complemented perception data set is used to correct the digital twin of the power equipment; in this way, the null values of the perception data of the power equipment can be accurately filled through the Kalman filter model, so as to efficiently complete the complement of the perception data of the power equipment, which is beneficial to accurately correcting the digital twin of the power equipment. Brief Description of the Drawings

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0036] Figure 1 It is an application environment diagram of a method for complementing multi-source heterogeneous perception information of power equipment based on the Kalman filter model in an embodiment;

[0037] Figure 2 It is a schematic flowchart of a method for complementing multi-source heterogeneous perception information of power equipment based on the Kalman filter model in an embodiment;

[0038] Figure 3 It is a specific flowchart of a method for complementing multi-source heterogeneous perception information of power equipment in an embodiment;

[0039] Figure 4 It is a schematic flowchart of a method for complementing multi-source heterogeneous perception information of power equipment based on the Kalman filter model in another embodiment;

[0040] Figure 5 It is a structural block diagram of a device for complementing multi-source heterogeneous perception information of power equipment based on the Kalman filter model in an embodiment;

[0041] Figure 6 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0042] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0043] Existing methods for filling in missing data mainly include the mean filling method, the polynomial filling method, the nearest neighbor filling method, etc. The mean filling method uses the mean of the existing data sequence to fill in the missing data; the polynomial filling method first performs polynomial fitting on the existing data to obtain a fitting formula, and then predicts the data at the missing points based on the fitting formula; the nearest neighbor filling method searches for a value in the existing data sequence that is close to the missing value based on a distance function and uses this value as the value at the missing point. The above methods are all fillings based on the numerical characteristics of the existing data sequence, lacking consideration of the physical meaning of the data itself and the data fluctuation characteristics. The filling process is too rough and the filling accuracy is low. The present application can solve the above problems and effectively supplement the missing data in the state quantities of the multi-source heterogeneous perception data of power equipment.

[0044] The method for filling in multi-source heterogeneous perception information of power equipment based on the Kalman filter model provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed on the cloud or other network servers. The server 104 obtains the perception data set corresponding to the power equipment; the perception data set is obtained after preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors; the server 104 establishes a Kalman filter model for the power equipment based on the perception data set; the server 104 uses the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain the parameter estimation result; the server 104 performs Kalman filter recursion based on the parameter estimation result to obtain the perception data prediction result corresponding to the power equipment; the server 104 performs data completion on the perception data set based on the perception data prediction result to obtain the completed perception data set; the completed perception data set is used to correct the digital twin of the power equipment. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0045] In an exemplary embodiment, as Figure 2 shown, a method for completing multi-source heterogeneous perception data of power equipment based on a Kalman filter model is provided. Taking the server 104 in Figure 1 as an example, the following steps S202 to S210 are included. Among them:

[0046] Step S202, obtain the perception data set corresponding to the power equipment; the perception data set is obtained after preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors.

[0047] Among them, the power equipment can refer to equipment with a voltage level of 1000V or above, mainly including high-voltage switchgear, high-voltage transformers, high-voltage circuit breakers, high-voltage disconnectors, high-voltage cables, etc., which are used for the transmission, distribution and control of high-voltage power systems.

[0048] Among them, the original perception data set can include physical perception data, equipment status data and working mode data collected by different sensors.

[0049] Among them, the perception data set can be a data set that can be directly used for data processing after preprocessing the original perception data set. After removing the abnormal data in the original perception data set, preset data needs to be filled in the positions corresponding to the abnormal data. The preset data can be 0 or other values.

[0050] Among them, the sensors can be voltage sensors, current sensors, temperature sensors, pressure sensors, humidity sensors, ultrasonic sensors, etc., and which sensors to use can be determined according to the characteristics of the power equipment.

[0051] Among them, the original perception data under multiple dimensions can refer to the original data under multiple perception dimensions.

[0052] Optionally, the server obtains the perception data set corresponding to the power equipment. For example, it obtains the perception data sets of the converter transformer and the GIS equipment.

[0053] Step S204: Based on the perception data set, establish a Kalman filter model for the power equipment.

[0054] Among them, the Kalman filter model can be used to describe the evolution of the state of the power equipment over time and the relationship between the observed values and the state of the power equipment. Other state-space models assume that both the state and the observed values of the power equipment are random variables and consider the effects of process noise and observation noise. In this application, by using the Kalman filter model, dynamic adjustment of data and generation of filling data can be achieved to complete the complementation of the perception data of the power equipment.

[0055] Optionally, the server establishes a Kalman filter model for the power equipment based on the perception data set of the power equipment.

[0056] Step S206: Use the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain the parameter estimation result.

[0057] Among them, the expectation maximization algorithm refers to the EM algorithm (Expectation-Maximization Algorithm), which is an iterative algorithm for parameter estimation of probability models containing latent variables and is widely used in fields such as machine learning and statistics. It can achieve good results when dealing with incomplete data or data sets with hidden variables.

[0058] Among them, the parameter estimation result can refer to the estimation result of the model parameters of the Kalman filter model.

[0059] Optionally, the server uses the EM algorithm to estimate the parameters of the Kalman filter model to obtain the estimated values of the model parameters of the Kalman filter model.

[0060] Step S208, perform Kalman filter recursion based on the parameter estimation result to obtain the prediction result of the sensed data corresponding to the power device.

[0061] Optionally, the server performs Kalman filter recursion processing based on the model parameter estimation value of the Kalman filter to obtain the prediction result of the sensed data for the power device.

[0062] Step S210, based on the prediction result of the sensed data, complete the data in the sensed data set to obtain the completed sensed data set; the completed sensed data set is used to correct the digital twin of the power device.

[0063] Among them, the completed sensed data set may refer to the result obtained after replacing the 0-value data in the sensed data set.

[0064] Optionally, the server completes the missing data in the sensed data set based on the prediction result of the sensed data to obtain the completed sensed data set, and subsequently, the completed sensed data set can be used to correct the digital twin of the power device.

[0065] In the above method for completing multi-source heterogeneous sensed data of power devices based on the Kalman filter model, by obtaining the sensed data set corresponding to the power device; the sensed data set is obtained after preprocessing the abnormal data in the original sensed data set of the power device; the original sensed data set includes the original sensed data of the power device in multiple dimensions collected by different sensors; based on the sensed data set, a Kalman filter model for the power device is established; the expectation-maximization algorithm is used to estimate the parameters of the Kalman filter model to obtain the parameter estimation result; perform Kalman filter recursion based on the parameter estimation result to obtain the prediction result of the sensed data corresponding to the power device; based on the prediction result of the sensed data, complete the data in the sensed data set to obtain the completed sensed data set; the completed sensed data set is used to correct the digital twin of the power device; thus, the null values of the sensed data of the power device can be accurately filled through the Kalman filter model, so as to efficiently complete the completion of the sensed data of the power device, which is beneficial to accurately correcting the digital twin of the power device.

[0066] In an exemplary embodiment, establishing a Kalman filter model for a power device based on a sensed data set includes: based on the sensed data set, constructing a state transition model and an observation model; the state transition model characterizes the law of change of the device state of the power device; the observation model characterizes the relationship between the device state of the power device and the sensed data; based on the state transition model and the observation model, a Kalman filter model is formed.

[0067] In practical applications, a location point is given and a time point , where the state transition model can be expressed as , and the observation model can be expressed as , is the state transition matrix, is the system error at time t, is the state vector, representing the average error component at time t at each position point state, represents the average error component at time t - 1 at each position point state, represents the observation vector at time t, represents the observation matrix, is the observation error at each position point at time t.

[0068] Among them, the law of equipment state change can refer to the law of the operating condition change of power equipment over time, which is specifically reflected in the change law of perception data in multiple dimensions.

[0069] Among them, the equipment state can be the operating state of the power equipment.

[0070] Optionally, the server constructs a model that can characterize the law of equipment state change of the power equipment and an observation model that characterizes the numerical relationship between the equipment state of the power equipment and the perception data based on the perception data set of the power equipment, and forms a Kalman filter model based on the state transition model and the observation model.

[0071] In this embodiment, by constructing a state transition model and an observation model based on the perception data set; the state transition model characterizes the law of equipment state change of the power equipment; the observation model characterizes the relationship between the equipment state of the power equipment and the perception data; a Kalman filter model is formed based on the state transition model and the observation model; thus, by establishing a state transition model that can describe the law of equipment state change of the power equipment and an observation model that describes the relationship between the equipment state of the power equipment and the perception data, an accurate Kalman filter model can be obtained.

[0072] In an exemplary embodiment, the expectation - maximization algorithm is used to estimate the parameters of the Kalman filter model to obtain the parameter estimation result, including: using the Kriging model to construct the spatial field of the Kalman filter model; using the expectation - maximization algorithm to estimate the parameters of the Kalman filter model based on the spatial field to obtain the parameter estimation result.

[0073] Among them, the Kriging model can be the Kriging Model, which is an interpolation method based on the Gaussian process. It is an unbiased estimation model that predicts the values at unknown data points through the data of known data points. It can give the predicted value and the error of the predicted value, and is widely used in the fields of geostatistics, geological exploration, environmental science, etc. In this application, it is applied to the identification of abnormal data points of power equipment.

[0074] Among them, the spatial field can be the spatial field constructed by the Kriging model 。

[0075] In practical applications, at time point t, the universal Kriging model is expressed as: ,where, is the given spatial trend field, and its elements are functions of the s coordinate positions. Commonly used spatial trend fields include constant trend, linear trend, and quadratic trend. The subscript q is determined by the selected trend. When the constant trend is selected, the subscript q is 1. When the linear trend is selected, the subscript is 3. When the quadratic trend is selected, the subscript q is 6. The subscript q represents the number of parameters required. is the coefficient of, is the local spatial variation after removing the spatial trend.

[0076] Optionally, the server adopts the Kriging model to construct the spatial field of the Kalman filter model, and uses the expectation maximization algorithm to estimate the parameters of the Kalman filter model based on the spatial field to obtain the parameter estimation result.

[0077] In this embodiment, by adopting the Kriging model to construct the spatial field of the Kalman filter model; using the expectation maximization algorithm to estimate the parameters of the Kalman filter model based on the spatial field to obtain the parameter estimation result; in this way, the spatial correlation between normal data points can be fully utilized to estimate the values of abnormal data points, so as to achieve a more accurate estimation of the model parameters of the Kalman filter model.

[0078] For the convenience of understanding by those skilled in the art, the following provides a derivation method of the observation model. In practical applications, given the position point and the time point , the observed value can be decomposed into an average error component and an observation error component: , where, is the average error component, is the observation error component. The derivation process from to includes:

[0079] First, determine the expression of, specifically, define the spatial basis function matrix in the constructed spatial field (Each position s in the spatial field corresponds to p components), define as the time-varying weight coefficient vector. The average error component at the spatial field position s at time t can be expressed as , and can be expressed as a time-varying linear combination of the spatial field , that is , namely , represents the component value of the i-th component at the position s in the spatial field, represents the state of the i-th component at the spatial field position s;

[0080] Secondly, substitute the expression of , that is into , and we can get ;

[0081] Finally, it can be determined that:

[0082] , denoted as , , then .

[0083] In this way, the derivation of the observation model is completed. In the observation model , is the average error component at time t at each position point in the spatial field under each spatial component, represents the observation vector at time t, H represents the observation matrix, is the observation error at each position point in the spatial field at time t under each spatial component.

[0084] For the convenience of understanding by those skilled in the art, the following also provides a method for estimating the model parameters of the Kalman filter model using the Expectation-Maximization algorithm (EM algorithm). Maximum likelihood estimation (MLE) is an effective method for estimating the parameter . Among them, is the initial state vector, representing the initial hidden state of the system at time t = 0; is the initial error covariance matrix, describing the uncertainty of the initial state estimation; represents the state transition matrix, defining the dynamic law of the state vector evolving over time; is the process noise covariance matrix, characterizing the uncertainty of the system model; is the observation noise covariance matrix, which describes the statistical characteristics of the errors of the observation equipment. Since the distribution of the state vector is unknown, it is impossible to directly maximize the logarithm of the joint likelihood function log(l).

[0085] The EM algorithm provides an iterative method for finding the maximum value of log(l), which consists of the Expectation Step (E-step) and the Maximization Step (M-step). Assuming it is the (r + 1)-th iteration, the main steps of the EM algorithm are as follows:

[0086] 1) Use the Kalman filter to estimate the unknown state parameters relative to the (r)-th iteration value . Among them, (r) represents the number of iterations, and the iteration value represents the estimated parameter at the r-th iteration.

[0087] 2) E-step: Calculate the conditional expectation of log(l) under the distribution estimated in step 1, where is the expectation operator.

[0088] 3) M-step: Maximize , which will generate a new iteration value .

[0089] 4) Replace with , and repeat steps 1), 2), and 3) until the logarithm of the joint likelihood function log(l) stops increasing.

[0090] After estimating the matrix H and the parameter q, the Kalman filter recursion can be carried out. The interpolation of missing data can be calculated dynamically while filtering the noise. Suppose at time point t, the observation equation is:

[0091] ,

[0092] where , , are the observation data, the spatial field, and the observation data error respectively, , , are the missing data, the spatial field, and the missing data error respectively.

[0093] For the Kalman recursion, first set , as zero matrices, and the estimation method of is expressed as:

[0094] ,

[0095] Among them, represents the optimal estimate of the state vector based on all the observed data up to time t, represents the mathematical expectation of the state vector under the condition of the known observation sequences L1 to Lt, reflecting the optimal unbiased estimate of the state, represents the predicted value of the state at time t based only on the observed data up to time t−1, without including the observation information at time t, represents the weight that balances the observation residual and the prior estimate, and is used to update the posterior state estimate, represents the observation vector at time t, represents the observation matrix. The error matrix of this model is:

[0096] ,

[0097] Among them, represents the error uncertainty that describes the difference between the posterior state estimate and the true state. represents the prediction error uncertainty that describes the difference between the prior state estimate and the true state. The Kalman recursion formula can be written as:

[0098] ,

[0099] ,

[0100] Among them, is the state transition matrix, represents the noise or uncertainty in the system dynamic model (state transition equation).

[0101] The optimal estimated value at any position point and the time point of interest can be calculated by the following formula:

[0102] ,

[0103] Among them, represents the optimal estimated value (such as the predicted value or the fitted value) at time point t and position s. It is usually used in spatio-temporal models or regression analysis to describe the estimated result of the target variable at a specific spatio-temporal point. is a basis function vector (or eigenvector) related to position s, which is used to capture the features in space or other dimensions.

[0104] In an exemplary embodiment, a perception data set corresponding to the power equipment is obtained, including: obtaining the original perception data set; performing replacement processing on the abnormal data in the original perception data set to obtain the perception data set.

[0105] Among them, the abnormal data can be abnormal data such as over-range data, negative value data, consecutive identical data, etc.

[0106] Optionally, the server obtains the original perception data under multiple dimensions collected by different sensors, and replaces the abnormal data in the original perception data under multiple dimensions to obtain a perception data set after data replacement.

[0107] In this embodiment, by obtaining the original perception data set and then replacing the abnormal data in the original perception data set, a perception data set is obtained. In this way, the abnormal data in the original perception data set can be quickly replaced, thereby realizing the rapid preprocessing of the original perception data set.

[0108] In an exemplary embodiment, replacing the abnormal data in the original perception data set to obtain a perception data set includes: using a preset data elimination rule to eliminate the abnormal data in the original perception data set to obtain a processed perception data set; using preset data to replace the null values in the processed perception data set to obtain a perception data set.

[0109] Among them, the preset data elimination rule can be a data screening rule preset in the perception database.

[0110] Among them, the processed perception data set can be a data set after eliminating abnormal data.

[0111] Among them, the preset data can be the numerical value 0 or a 0 matrix.

[0112] Optionally, the server eliminates the abnormal data such as over-range data, negative value data, consecutive identical data, etc. in the original perception data set according to the preset data elimination rule to obtain a processed perception data set, and then uses the numerical value 0, 0 matrix, etc. to replace the null values in the processed perception data set to obtain a perception data set.

[0113] In this embodiment, by using the preset data elimination rule to eliminate the abnormal data in the original perception data set to obtain a processed perception data set; using preset data to replace the null values in the processed perception data set to obtain a perception data set; in this way, the interference of abnormal data on the actual estimation process can be reduced, which is beneficial to more efficiently complement the data of the perception data set.

[0114] In an exemplary embodiment, the original perception data set includes current data, voltage data, temperature data, humidity data, device status data, and working mode data of the power equipment at each time point.

[0115] Among them, the time point can refer to the sampling time point corresponding to each sensor when collecting data.

[0116] Among them, the current data may refer to the current value collected by a current sensor.

[0117] Among them, the voltage data may refer to the voltage value collected by a voltage sensor.

[0118] Among them, the temperature data may be the temperature value collected by a temperature sensor.

[0119] Among them, the humidity data may be the humidity value collected by a humidity sensor.

[0120] Among them, the equipment status data may refer to the equipment status of the power equipment at each time point.

[0121] Among them, the working mode data may refer to the working mode of the power equipment at each time point.

[0122] In this embodiment, since the original perception data of the power equipment in multiple dimensions is collected, it is beneficial to achieve more comprehensive data completion, and thus it is beneficial to more accurately correct the digital twin of the power equipment.

[0123] For the convenience of those skilled in the art to understand, Figure 3 a specific flowchart of a method for completing multi-source heterogeneous perception data of power equipment is provided. The method includes a spatial field construction step, a parameter estimation step, and a Kalman filter recursion step.

[0124] In the spatial field construction step, it is necessary to select an appropriate spatial trend to calculate the stationary residual, then calculate the average semivariance value, select an appropriate semivariance model for fitting, and construct an improved Kalman model.

[0125] The stationary residual is the random fluctuation part remaining after removing the spatial trend (such as systematic change or long-term trend). Assuming that the residual satisfies stationarity, that is, its statistical characteristics (mean, variance, covariance) do not change with position in space, which is convenient for subsequent modeling.

[0126] The average semivariance value is the average of all semivariance values within the same distance interval, which is used to draw a semivariogram and reflect the change of the strength of spatial dependence with distance, and can identify the spatial structure of the data.

[0127] An appropriate semivariance model is to fit the average semivariance value in the semivariogram through a mathematical function to quantitatively describe spatial correlation. Selecting the best fitting model according to the shape of the semivariogram directly affects the accuracy of Kriging interpolation or Kalman filtering.

[0128] In the parameter estimation step, it is necessary to use the EM algorithm to estimate the model parameters.

[0129] In the Kalman filter recursion step, it is necessary to use a 0 matrix to replace the missing data and spatial field information, and use Kalman recursion to filter the data and complete the missing data or unexamined data. Replacing the missing data and spatial field information with a 0 matrix is essentially to maintain the recursive operation of the filter through mathematical simplification while ignoring the influence of the missing part.

[0130] In another embodiment, as Figure 4 shown, a method for completing multi-source heterogeneous perception data of power equipment based on a Kalman filter model is provided. Taking the application of this method to Figure 1 the server 104 in as an example, the method includes the following steps:

[0131] Step S402, obtain the perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors.

[0132] Step S404, based on the perception data set, construct a state transition model and an observation model; the state transition model characterizes the law of change of the device state of the power equipment; the observation model characterizes the relationship between the device state of the power equipment and the perception data.

[0133] Step S406, based on the state transition model and the observation model, form a Kalman filter model.

[0134] Step S408, use the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain the parameter estimation result.

[0135] Step S410, perform Kalman filter recursion based on the parameter estimation result to obtain the perception data prediction result corresponding to the power equipment.

[0136] Step S412, based on the perception data prediction result, complete the data of the perception data set to obtain the completed perception data set; the completed perception data set is used to correct the digital twin of the power equipment.

[0137] It should be noted that the specific limitations of the above steps can refer to the specific limitations of a method for completing multi-source heterogeneous perception data of power equipment based on a Kalman filter model above.

[0138] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are sequentially shown according to the indications of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless clearly stated herein, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0139] Based on the same inventive concept, an embodiment of the present application further provides a Kalman filter model-based power equipment multi-source heterogeneous perception data completion device for implementing the above-mentioned Kalman filter model-based power equipment multi-source heterogeneous perception data completion method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the Kalman filter model-based power equipment multi-source heterogeneous perception data completion device provided below can refer to the limitations on the Kalman filter model-based power equipment multi-source heterogeneous perception data completion method in the above text, and will not be elaborated here.

[0140] In an exemplary embodiment, as Figure 5 shown, a Kalman filter model-based power equipment multi-source heterogeneous perception data completion device is provided, including: an acquisition module 502, a building module 504, an estimation module 506, a prediction module 508, and a completion module 510, where:

[0141] The acquisition module 502 is configured to acquire a perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing the abnormal data in the original perception data set of the power equipment; the original perception data set includes the original perception data of the power equipment in multiple dimensions collected by different sensors;

[0142] The building module 504 is configured to build a Kalman filter model for the power equipment based on the perception data set;

[0143] The estimation module 506 is configured to perform parameter estimation on the Kalman filter model by using the expectation maximization algorithm to obtain a parameter estimation result;

[0144] The prediction module 508 is configured to perform Kalman filter recursion based on the parameter estimation result to obtain a perception data prediction result corresponding to the power equipment;

[0145] A completion module 510 is configured to complete the perception data set based on the prediction result of the perception data to obtain a completed perception data set; the completed perception data set is used to correct the digital twin of the power equipment.

[0146] In an exemplary embodiment, a building module 504 is specifically configured to build a state transition model and an observation model based on the perception data set; the state transition model characterizes the change law of the device state of the power equipment; the observation model characterizes the relationship between the device state of the power equipment and the perception data; a Kalman filter model is formed based on the state transition model and the observation model.

[0147] In an exemplary embodiment, an estimation module 506 is specifically configured to use a Kriging model to build a spatial field of the Kalman filter model; use the expectation maximization algorithm to estimate the parameters of the Kalman filter model based on the spatial field to obtain a parameter estimation result.

[0148] In an exemplary embodiment, an acquisition module 502 is specifically configured to acquire an original perception data set; perform replacement processing on the abnormal data in the original perception data set to obtain a perception data set.

[0149] In an exemplary embodiment, an acquisition module 502 is specifically configured to use a preset data elimination rule to eliminate the abnormal data in the original perception data set to obtain a processed perception data set; use preset data to replace the null values in the processed perception data set to obtain a perception data set.

[0150] In an exemplary embodiment, the original perception data set includes current data, voltage data, temperature data, humidity data, device state data, and working mode data of the power equipment at each time point.

[0151] Each module in the above multi-source heterogeneous perception data completion device for power equipment based on the Kalman filter model can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0152] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 6As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the multi-source heterogeneous perception data completion data of power equipment based on the Kalman filter model. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it realizes a method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model.

[0153] Those skilled in the art can understand that Figure 6 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0154] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the steps of the above-mentioned method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model. Here, the steps of the method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model may be the steps in the method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model in each of the above embodiments.

[0155] In one embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by the processor, the processor executes the steps of the above-mentioned method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model. Here, the steps of the method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model may be the steps in the method for completing multi-source heterogeneous perception data of power equipment based on the Kalman filter model in each of the above embodiments.

[0156] In one embodiment, a computer program product is provided, including a computer program which, when executed by a processor, causes the processor to execute the steps of the above-mentioned method for complementing multi-source heterogeneous sensing data of power equipment based on a Kalman filter model. The steps of the method for complementing multi-source heterogeneous sensing data of power equipment based on a Kalman filter model here may be the steps in the method for complementing multi-source heterogeneous sensing data of power equipment based on a Kalman filter model in each of the above embodiments.

[0157] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0158] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0159] The embodiments described above merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for completing multi-source heterogeneous sensing data of power equipment based on a Kalman filter model, characterized in that, The method includes: Obtaining a perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing abnormal data in the original perception data set of the power equipment; the original perception data set includes original perception data of the power equipment in multiple dimensions collected by different sensors; Based on the perception data set, establishing a Kalman filter model for the power equipment; Using the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain a parameter estimation result; Based on the parameter estimation result, performing Kalman filter recursion to obtain a perception data prediction result corresponding to the power equipment; Based on the perception data prediction result, performing data completion on the perception data set to obtain a completed perception data set; the completed perception data set is used to correct the digital twin of the power equipment.

2. The method according to claim 1, wherein The establishing a Kalman filter model for the power equipment based on the perception data set includes: Based on the perception data set, constructing a state transition model and an observation model; the state transition model characterizes the change law of the equipment state of the power equipment; the observation model characterizes the relationship between the equipment state of the power equipment and the perception data; Based on the state transition model and the observation model, forming the Kalman filter model.

3. The method according to claim 1, wherein The using the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain a parameter estimation result includes: Using a Kriging model to construct a spatial field of the Kalman filter model; Using the expectation maximization algorithm to estimate the parameters of the Kalman filter model based on the spatial field to obtain the parameter estimation result.

4. The method according to claim 1, wherein The obtaining a perception data set corresponding to the power equipment includes: Obtaining the original perception data set; Performing replacement processing on the abnormal data in the original perception data set to obtain the perception data set.

5. The method according to claim 4, characterized in that, The performing replacement processing on the abnormal data in the original perception data set to obtain the perception data set includes: Using a preset data rejection rule to reject the abnormal data in the original perception data set to obtain a processed perception data set; Using preset data to replace the null values in the processed perception data set to obtain the perception data set.

6. The method according to claim 1, characterized in that, The original perception data set includes current data, voltage data, temperature data, humidity data, equipment state data, and working mode data of the power equipment at each time point.

7. A multi-source heterogeneous perception data completion device for power equipment based on a Kalman filter model, characterized in that, The device includes: An obtaining module, configured to obtain a perception data set corresponding to the power equipment; the perception data set is obtained by preprocessing abnormal data in the original perception data set of the power equipment; the original perception data set includes original perception data of the power equipment in multiple dimensions collected by different sensors; A establishing module, configured to establish a Kalman filter model for the power equipment based on the perception data set; An estimating module, configured to use the expectation maximization algorithm to estimate the parameters of the Kalman filter model to obtain a parameter estimation result; A prediction module, configured to perform Kalman filtering recursion based on the parameter estimation result to obtain a prediction result of the sensed data corresponding to the power device; A completion module, configured to perform data completion on the sensed data set based on the prediction result of the sensed data to obtain a completed sensed data set; the completed sensed data set is used to correct the digital twin of the power device.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images