CN116756638B

CN116756638B - Method, device, equipment and storage medium for detecting electric load demand of electric vehicle

Info

Publication number: CN116756638B
Application number: CN202311034988.9A
Authority: CN
Inventors: 张英凯; 黄斌; 廖晓义; 汪泽峰
Original assignee: Shantou Power Supply Bureau of Guangdong Power Grid Co Ltd
Current assignee: Shantou Power Supply Bureau of Guangdong Power Grid Co Ltd
Priority date: 2023-08-17
Filing date: 2023-08-17
Publication date: 2023-11-14
Anticipated expiration: 2043-08-17
Also published as: CN116756638A

Abstract

The invention discloses a method, a device, equipment and a storage medium for detecting the electric load demand of an electric vehicle, wherein the method comprises the following steps: collecting original current data from a user belonging to resident life electricity in a power distribution network; clustering the users according to the electricity utilization behavior to obtain user categories; intercepting part of original current data according to the user category as target current data; loading a charging detection network according to the user category; encoding target current data in an encoder to obtain a first target current characteristic for distinguishing whether electric bicycle charging exists or not; decoding the first target current characteristics in a decoder to reconstruct in sequence a plurality of levels of second target current characteristics characterizing the user class when the electric bicycle is charged; in the classifier, under the enhancement effect of the second target current characteristics of a plurality of levels, whether the electric consumer has an electric load for charging the electric bicycle or not is detected according to the first target current characteristics. The accuracy of detecting electric bicycle charging is improved.

Description

Method, device, equipment and storage medium for detecting electric load demand of electric vehicle

Technical Field

The invention relates to the technical field of power grids, in particular to a method, a device and equipment for detecting electric load requirements of an electric vehicle and a storage medium.

Background

Electric bicycles (also called electric vehicles) are vehicles in which a control member such as a motor, a controller, a battery, a handle bar, and a display instrument system are mounted on the basis of a general bicycle by using a battery as an auxiliary energy source.

The electric bicycle generally comprises a motor, a battery, a controller, a charger, an instrument, a sensor, a frame, a decoration part and the like, and the end consumer downstream of the electric bicycle industry is mainly a personal resident or a household.

With the wide popularization of electric bicycles, the holding quantity of the electric bicycles is larger and larger, and is as large as tens of millions, part of users can push the electric bicycles back to the house for charging or pull the storage battery out, and the users can charge the house by independently carrying the storage battery, so that the fire risk of the house is increased, and hidden danger is brought to the life and property safety of the users.

At present, based on a neural network of deep learning, a non-invasive electrical signal load is analyzed to identify whether the electric bicycle charging feature exists in the residence of a user, so as to identify the electric bicycle charging behavior in the residence.

However, there are complicated electric signals in the resident electricity, and these electric signals may cause a great deal of interference to the electric bicycle charging, so that the accuracy of identifying the electric bicycle charging is low.

Disclosure of Invention

The invention provides a method, a device, equipment and a storage medium for detecting the electric load demand of an electric bicycle, which are used for solving the problem of how to improve the accuracy of detecting the charging of the electric bicycle.

According to an aspect of the present invention, there is provided a method for detecting an electric load demand for an electric vehicle, including:

collecting original current data from a user belonging to resident life electricity in a power distribution network;

clustering the power utilization users according to the power utilization behaviors to obtain user categories;

intercepting part of the original current data according to the user category as target current data;

loading a charging detection network according to the user category, wherein the charging detection network comprises an encoder, a decoder for training the user category and a classifier;

encoding the target current data in the encoder to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged or not;

decoding the first target current characteristics in the decoder to reconstruct, in turn, a plurality of levels of second target current characteristics characterizing the user class as there is electric bicycle charging;

And in the classifier, under the enhancement effect of the second target current characteristics of a plurality of levels, detecting whether the electric consumer has an electric load for charging the electric bicycle according to the first target current characteristics.

According to another aspect of the present invention, there is provided an electric load demand detection apparatus for an electric vehicle, including:

the primary current data acquisition module is used for acquiring primary current data from power consumers belonging to resident life power consumption in the power distribution network;

the power utilization user clustering module is used for clustering the power utilization users according to the power utilization behaviors to obtain user categories;

the target current data intercepting module is used for intercepting part of the original current data according to the user category and taking the original current data as target current data;

the charging detection network loading module is used for loading a charging detection network according to the user category, and the charging detection network comprises an encoder, a decoder for training the user category and a classifier;

the encoding module is used for encoding the target current data in the encoder to obtain a first target current characteristic used for distinguishing whether the electric bicycle is charged or not;

the decoding module is used for decoding the first target current characteristics in the decoder so as to reconstruct a plurality of levels of second target current characteristics representing the user class when the electric bicycle is charged;

And the classification module is used for detecting whether the electric consumer has an electric load for charging the electric bicycle according to the first target current characteristic under the enhancement effect of the second target current characteristics of a plurality of levels in the classifier.

According to another aspect of the present invention, there is provided an electronic apparatus including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor, so that the at least one processor can execute the electric load demand detection method for the electric vehicle according to any embodiment of the present invention.

According to another aspect of the present invention, there is provided a computer readable storage medium storing a computer program for causing a processor to execute the method for detecting electric load demand for an electric vehicle according to any one of the embodiments of the present invention.

In the embodiment, original current data are collected for power consumers belonging to resident life power consumption in a power distribution network; clustering the users according to the electricity utilization behavior to obtain user categories; intercepting part of original current data according to the user category as target current data; loading a charging detection network according to the user category, wherein the charging detection network comprises an encoder, a decoder for training the user category and a classifier; encoding target current data in an encoder to obtain a first target current characteristic for distinguishing whether electric bicycle charging exists or not; decoding the first target current characteristics in a decoder to reconstruct in sequence a plurality of levels of second target current characteristics characterizing the user class when the electric bicycle is charged; in the classifier, under the enhancement effect of the second target current characteristics of a plurality of levels, whether the electric consumer has an electric load for charging the electric bicycle or not is detected according to the first target current characteristics. The encoder provides a shared and basic first target current characteristic for the decoder and the classifier, the accuracy of the first target current characteristic for two classifications can be ensured, the influence on detecting electric bicycle charging caused by boundary blurring among user classes is reduced, the decoder reconstructs a plurality of levels of second target current characteristics when the electric bicycle is charged under the user classes, the second target current characteristics of different levels are sequentially enhanced for the first target current characteristic, the information quantity of the first target current characteristic is enriched, the accuracy of detecting electric bicycle charging is improved, the decoder and the classifier belong to a part of the whole charging detection network, the characteristics of the bottom layers are not concerned, the complexity of the structure is small, the electric bicycle charging behavior of whether electric bicycles exist under different user classes can be distinguished, the electric safety is improved, the operation quantity of detection can be reduced, and the occupation of resources is reduced.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flowchart of a method for detecting an electric load demand of an electric vehicle according to a first embodiment of the present invention;

fig. 2 is a schematic structural diagram of a charge detection network according to a first embodiment of the present invention;

fig. 3 is a flowchart of a method for detecting an electric load demand of an electric vehicle according to a second embodiment of the present invention;

fig. 4 is a schematic structural diagram of an electric load demand detection device for an electric vehicle according to a third embodiment of the present invention;

fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.

Detailed Description

In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Example 1

Fig. 1 is a flowchart of a method for detecting an electric load demand of an electric vehicle according to an embodiment of the present invention, where the method may be performed by an electric load demand detection device for an electric vehicle, and the electric load demand detection device for an electric vehicle may be implemented in hardware and/or software, and the electric load demand detection device for an electric vehicle may be configured in an electronic device. As shown in fig. 1, the method includes:

and 101, collecting original current data from power consumers belonging to resident life power consumption in the power distribution network.

In practical application, the electric consumers of the power distribution network can be classified into residential life electric consumption, general industrial and commercial electric consumption, large industrial electric consumption, agricultural production electric consumption and the like according to the electric consumption property of the industry.

The residential electricity consumption comprises urban and rural residential electricity consumption, urban and rural residential community public auxiliary facility electricity consumption, school teaching and student electricity consumption, social benefit place electricity consumption and the like.

In this embodiment, current data is collected for users divided into residential electricity in the distribution network, and recorded as raw current data.

For the original current data, preprocessing can be performed, such as removing redundant data, removing abnormal values, filling missing values and the like, so as to improve the quality of the original current data.

And 102, clustering the users according to the electricity consumption behavior to obtain user categories.

In this embodiment, data representing the electricity consumption behavior may be collected for the current user and recorded as the electricity consumption behavior data, for example, a time point of a switching event (i.e. an event that the electric device is turned on or off), electricity consumption power, electricity consumption duration, and the like, and the current user is clustered in the dimension of the electricity consumption behavior to obtain the user category to which the current user belongs.

In a specific implementation, clustering algorithms such as K-means (K-means clustering) and the like can be used in advance to cluster the users in the dimension of the electricity consumption behavior, so that a plurality of clusters representing user categories are obtained, and each cluster is provided with a center point.

When clustering current users, a plurality of clusters representing user categories may be loaded that are pre-trained.

The current power consumer is subjected to power consumption behavior data acquisition, the power consumption behavior data are subjected to power consumption behavior data in a mode of threshold value, type and the like, the first power consumption behavior data which are irrelevant to electric bicycle charging and the second power consumption behavior data which are relevant to electric bicycle charging are independently divided, the second power consumption behavior data which are relevant to electric bicycle charging can be increased, the duty ratio of the characteristics of the electric bicycle charging can be increased, and a group which can potentially charge the electric bicycle can be conveniently identified during clustering.

The first electrical behavior data is converted into a first vector and the second electrical behavior data is converted into a second vector by using a one-hot or other coding algorithm.

And splicing the first vector and the second vector into a third vector, and calculating the distance between the third vector and each center point by using an Euclidean distance equidistant algorithm.

And comparing the distances, dividing the users into clusters with the smallest distance, and setting the user category represented by the clusters as the user category to which the current user belongs.

And 103, intercepting part of original current data according to the user category to serve as target current data.

In this embodiment, the characteristics of the electricity behavior of the same user class may be analyzed for the user, and a portion of the original current data that may exist for charging the electric vehicle may be intercepted as the target current data.

The different user types of electric users have different corresponding electricity utilization behaviors, and accordingly, the electric bicycle is charged in different behaviors, for example, the occupation of some users is express men and takeaway, the electric bicycle is ridden for a long time, due to occupation limitation and higher battery loss, the users replace the battery in idle time and charge the replaced battery, the residual capacity of the battery is possibly higher, some users ride the electric bicycle in the morning and evening to work or pick up children and buying dishes, the battery loss is less, the battery is charged when the residual capacity is less, and the like.

The duration of charging the electric bicycle (especially the storage battery) is inconsistent under different conditions, so that corresponding windows can be preset for users in different user categories, and the length of the windows can cover a first inflection point and a second inflection point when the electric bicycle is charged by the users in the user categories.

At present, most electric bicycle's battery charge is three segmentation charges, is respectively:

1. constant current phase

And the constant current charging is adopted, so that the condition that the low electric quantity is too low to cause excessive current charging is avoided, and the heat, fire and irreparable damage of the storage battery are caused.

The charging mechanism is that the charging current of the charger is kept constant, the charging quantity is steadily increased, and the voltage of the storage battery is gradually increased.

2. Constant pressure stage

And constant voltage is adopted for charging, so that the storage battery finally reaches rated voltage.

The charging mechanism is that the charging voltage is kept constant, and the battery voltage is slowly increased until the rated voltage of the battery is reached.

3. Floating charge phase (trickle phase)

The charging current is small, and the maintenance of the storage battery is mainly performed.

The charging mechanism is that the storage battery is fully charged, the charging voltage is basically equal to the voltage of the storage battery, the charging current is very small, and the charging mechanism has a certain effect on the holding capacity of the storage battery.

Further, the first inflection point is an intersection point between the constant-current phase and the constant-voltage phase, the second inflection point is an intersection point between the constant-voltage phase and the trickle phase, and the window covers the first inflection point and the second inflection point, so that the waveform of charging can be recognized.

In a specific implementation, a window and a downsampling factor configured for a user category may be loaded, where the window and the downsampling factor belong to matched parameters.

And sequentially loading windows on the original current data, intercepting part of data positioned in the windows as candidate current data, and hopefully intercepting part of data comprising the first inflection point and the second inflection point, so that the integrity of a charging curve is ensured, and the detection accuracy is improved.

Because the lengths of the windows are not consistent, the lengths of the candidate current data are not consistent correspondingly, and the lengths of the input data are required to be consistent by the deep learning-based charge detection network, the downsampling operation can be performed on the candidate current data according to the downsampling factor, and the target current data with uniform length can be obtained.

Illustratively, downsampling may be expressed as y (n) =x (n×l), where y (n) is target current data, x is candidate current data, n is a time domain subscript, and L is a downsampling factor.

Step 104, loading the charge detection network according to the user category.

In this embodiment, a charging detection network may be constructed and trained in advance for different user categories, where the charging detection network is used to detect whether a user in the user category has a behavior of charging an electric bicycle.

The structure of the charge detection network is not limited to an artificially designed neural network, but a neural network optimized by a model quantization method, a neural network searched for characteristics of electric bicycle charge by a NAS (Neural Architecture Search, neural network structure search) method, and the like, which is not limited in this embodiment.

The charging detection network comprises an Encoder Encoder, a Decoder trained on user classes, and a classifier.

Further, the Encoder is common to different user classes, the Decoder is adapted to the assigned user class, and the classifier is common to different user classes.

Then, during training, the target current data of the user who does not distinguish the user category and the current data of the electric bicycle charged separated from the target current data may be used as a sample to mark whether the electric bicycle charged tag exists, the Encoder, the Decoder and the classifier may be trained, and the loss function may use relative entropy (relative entropy).

When training is completed, the current data of the electric bicycle charged separated from the target current data under the condition of maintaining the encoder and the classifier without updating parameters is used as a sample, whether a label of the electric bicycle charged exists or not is marked, fine-tuning is performed on the Decoder, and MSE Loss (mean square error Loss) can be used as a Loss function.

Step 105, encoding the target current data in the encoder to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged or not.

In this embodiment, the target current data is input to the Encoder, which encodes the target current data, and extracts high-level features from the target current data, and marks the high-level features as first target current features.

The boundary between each user category is fuzzy, the old electric appliance is aged and even obsolete, the new electric appliance is purchased, and the weather is converted, the electricity utilization behavior of a user can change to a certain extent, if the charge detection network is trained by using current data marked with different user categories as a sample, the charge detection network can be poor in performance and lack of generalization, the electric bicycle is charged in a long time sequence operation, the current data in a longer window can obviously show the characteristic of charging the electric bicycle, and whether the electric bicycle is charged or not can be distinguished to a certain extent, therefore, when the Encoder is trained, the Encoder can be trained to the first target current characteristic of encoding the Encoder, and can be used for distinguishing whether the electric bicycle is charged or not.

In one embodiment of the present invention, as shown in FIG. 2, the Encoder Encoder includes a long and short term memory network (Long Short Term Memory, LSTM), a first convolutional layer Conv_1, a second convolutional layer Conv_2, a third convolutional layer Conv_3, a first Pooling layer Pooling_1, and a second Pooling layer Pooling_2.

For LSTM, the following key variables can be divided:

input: h is a _t-1 (hidden layer at time t-1) and x _t (feature vector at time t)

And (3) outputting: h is a _t (adding an activation function such as softmax can be used as a real output, otherwise, the activation function can be used as a hidden layer)

Mainline/memory: c _t-1 And c _t

The inside of the long-term and short-term memory network mainly comprises three stages:

1. forgetting stage

The forgetting stage is mainly used for selectively forgetting input transmitted by the last node, in particular z obtained by calculation _f (f represents forgetting about) c for controlling the last state as forgetting gating _t-1 Which leaves which ones forgotten.

2. Selection memory stage

The selective memory stage selectively "memorizes" the input, mainly the input x _t The selection memory is performed, which important matters are recorded, which are not important, and which are recorded. The current input content is represented by z calculated previously. And the selected gating signal is z _i (i represents information) to perform control.

Adding the results obtained in the two stages to obtain c which is transmitted to the next state _t 。

3. Output stage

The output phase will determine which will be the output of the current state. Mainly by z _o To control. And also for c obtained in the previous stage _o Scaling (by a tanh activation function).

The first convolution layer Conv_1, the second convolution layer Conv_2 and the third convolution layer Conv_3 belong to convolution layers (Convolutional Layer), the layer convolution layers are usually composed of a plurality of convolution units, parameters of each convolution unit are obtained through back propagation algorithm optimization during training, the lower-level convolution layers can extract some low-level features, such as edges, lines, angles and other levels, and the higher-level convolution layers can iteratively extract higher-level and more complex features from the lower-level features.

The first Pooling Layer pooling_1 and the second Pooling Layer pooling_2 belong to Pooling layers (Pooling layers), the Pooling layers can provide sampling operation, and the Pooling can compress input data in a certain mode so as to accelerate the operation speed of the charge detection network. Some way of doing this is in fact an algorithm for pooling operations, such as maximum pooling operations, average pooling operations, or minimum pooling operations.

In this embodiment, the pooling layer provides a minimum pooling operation, i.e. a region is represented by a certain local region minimum.

In general, the original current data is preprocessed, so that the abnormal condition with the value of 0 does not exist, the influence of the minimum pooling operation on the gradient is limited, and for electric bicycle charging, other electric signals belong to noise, and noise suppression is facilitated by taking the minimum value.

In this embodiment, the target current data is input into the long-short-period memory network LSTM, and time sequence processing is performed on the target current data in the long-short-period memory network LSTM, so as to obtain the first reference current characteristic, where the target current data belongs to a time sequence, and the LSTM can extract the characteristic in time sequence.

And inputting the first reference current characteristic into a first convolution layer Conv_1, and performing convolution operation on the first reference current characteristic in the first convolution layer Conv_1 to obtain a second reference current characteristic.

And inputting the second reference current characteristic into the first Pooling layer pooling_1, and executing minimum Pooling operation on the second reference current characteristic in the first Pooling layer pooling_1 to obtain a third reference current characteristic.

And inputting the third reference current characteristic into a second convolution layer Conv_2, and performing convolution operation on the third reference current characteristic in the second convolution layer Conv_2 to obtain a fourth reference current characteristic.

And inputting the fourth reference current characteristic into the third convolution layer Conv_3, and performing convolution operation on the fourth reference current characteristic in the third convolution layer Conv_3 to obtain a fifth reference current characteristic.

And inputting the fifth reference current characteristic into the second Pooling layer pooling_2, and executing minimum Pooling operation on the fifth reference current characteristic in the second Pooling layer pooling_2 to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged or not.

Step 106, decoding the first target current characteristics in a decoder to reconstruct the second target current characteristics of a plurality of layers, which represent the user category when the electric bicycle is charged.

The first target current feature output by the Encoder Encoder is processed in two ways, wherein one way is input into the Decoder, the Decoder decodes the first target current feature, and in the decoding process, the second target current features which characterize the user category and are in multiple levels when the electric bicycle is charged are sequentially reconstructed.

The electric users of different user categories have certain differences in selecting the electric bicycle and in using the electric bicycle, and also have certain differences in charging the electric bicycle, so that the electric users of different user categories have different characteristics when charging the electric bicycle, and the Decoder can be trained in different user categories, so that the Decoder can reconstruct the second target current characteristic aiming at the first target current characteristic of the corresponding user category, and the reconstruction process further strengthens the charging characteristic and noise weakening of the electric bicycle in the first target current characteristic.

In general, noise is mainly used in target current data, an Encoder Encoder mainly performs downsampling operation, noise can be weakened, characteristics of an electric bicycle during charging are highlighted in time sequence, a Decoder Decode mainly performs upsampling operation, a residual connection (skip connection) is not established between the Encoder Encoder and the Decoder Decode, and influence of noise on upsampling is reduced.

In one embodiment of the present invention, as shown in fig. 2, the encoder Decoder includes a first deconvolution layer deconv_1, a second deconvolution layer deconv_2, and a third deconvolution layer deconv_3.

The deconvolution layers deconv_1, deconvolution_2 and deconvolution_3 are deconvolution layers (deconvolution), also called transpose convolution (transposed convolution), which can provide deconvolution operation, upsample feature map of the previous layer, and implement interpolation of the neighborhood points, wherein the interpolation process is parameterized by a kernel representing the influence range and contribution of the feature points to the neighborhood. For efficient interpolation, the kernel should be large enough to cover the output.

Wherein the interpolation process multiplies the value of the kernel by each input and adds an overlapping response in the output.

In this embodiment, the first target current characteristic is input into the first deconvolution layer deconv_1, and deconvolution operation is performed on the first target current characteristic in the first deconvolution layer deconv_1, so as to obtain a second target current characteristic of the first level, which characterizes that the user class has the electric bicycle charging.

And inputting the second target current characteristic of the first level into a second deconvolution layer deconv_2, and performing deconvolution operation on the second target current characteristic of the first level in the second deconvolution layer deconv_2 to obtain the second target current characteristic of the second level, which characterizes the existence of electric bicycle charging in the user category.

And inputting the second target current characteristic of the second level into a third deconvolution layer deconv_3, and performing deconvolution operation on the second target current characteristic of the second level in the third deconvolution layer deconv_3 to obtain the second target current characteristic of the third level, which characterizes the existence of electric bicycle charging in the user category.

And step 107, detecting whether the electric consumer has an electric load charged by the electric bicycle according to the first target current characteristic under the enhancement effect of the second target current characteristics of a plurality of levels in the classifier.

The first target current feature output by the Encoder Encoder is processed in two ways, and the other way is input into the classifier, and the second target current features of multiple levels output by the Decoder Encoder are also input into the classifier together.

When the Decoder is used for reconstruction, the characteristics of electric bicycle charging in the first target current characteristics are enhanced, noise is weakened, and certain distortion exists relative to target current data, so that the classifier is used for classifying the first target current characteristics mainly and the second current characteristics of a plurality of layers secondarily, at the moment, the second current characteristics of the plurality of layers can enhance the first target current characteristics, the classifier classifies the first target current characteristics, and a result of whether the electric load of electric bicycle charging exists for a user or not is output.

In one embodiment of the present invention, as shown in fig. 2, the classifier includes a first full connection layer fc_1, a second full connection layer fc_2, a third full connection layer fc_3, a fourth full connection layer fc_4, a fifth full connection layer fc_5, a first separable convolution layer scl_1, a second separable convolution layer scl_2, and a third separable convolution layer scl_3.

The first full-connection layer fc_1, the second full-connection layer fc_2, the third full-connection layer fc_3, the fourth full-connection layer fc_4 and the fifth full-connection layer fc_5 all belong to full-connection layers (Fully Connected Layers, FC), neurons in the full-connection layers are fully connected with neurons in the previous layer, and the full-connection layers can map the learned "distributed feature representation" into a specified multidimensional space, in particular, a sample marking space.

The first separable convolution layer scl_1, the second separable convolution layer scl_2 and the third separable convolution layer scl_3 are all separable convolution layers (Separable Convolutions Layers), in particular spatially separable convolution layers (Spatial Separable Convolutions Layers), which are convolution operations performed by splitting a standard convolution operation into a plurality of small convolution kernels in a spatial dimension.

In this embodiment, the second target current feature of the first level is input into the first fully connected layer fc_1, and the second target current feature of the first level is mapped into the first candidate current feature in the first fully connected layer fc_1, wherein the dimension of the first candidate current feature is smaller than or equal to the dimension of the first target current feature.

And fusing the first target current characteristic and the first candidate current characteristic into a first fused current characteristic through a Concate function and the like.

Inputting the first fusion current characteristic into the first separable convolution layer SCL_1, and performing separable convolution operation on the first fusion current characteristic in the first separable convolution layer SCL_1 to obtain the first separation current characteristic.

And inputting a second target current characteristic of the second level into the second fully connected layer FC_2, and mapping the second target current characteristic of the second level into a second candidate current characteristic in the second fully connected layer FC_2, wherein the dimension of the second candidate current characteristic is smaller than or equal to the dimension of the first separation current characteristic.

And fusing the first separation current characteristic and the second candidate current characteristic into a second fusion current characteristic through a Concate function and the like.

Inputting the second fusion current characteristic into a second separable convolution layer SCL_2, and performing separable convolution operation on the second fusion current characteristic in the second separable convolution layer SCL_2 to obtain a second separation current characteristic.

And inputting the second target current characteristic of the third level into a third fully connected layer FC_3, and mapping the second target current characteristic of the third level into a third candidate current characteristic in the third fully connected layer FC_3, wherein the dimension of the third candidate current characteristic is smaller than or equal to the dimension of the second separation current characteristic.

And fusing the second separation current characteristic and the third candidate current characteristic into a third fusion current characteristic through a Concate function and the like.

Inputting the third fusion current characteristic into the third separable convolution layer SCL_3, and performing separable convolution operation on the third fusion current characteristic in the third separable convolution layer SCL_3 to obtain a third separation current characteristic.

The third split current characteristic is input into the fourth fully connected layer fc_4, which is mapped into a fourth split current characteristic in the fourth fully connected layer fc_4.

Inputting the fourth separation current characteristic into a fifth full-connection layer FC_5, and mapping the fourth separation current characteristic into the probability that the electric bicycle charges by a user in the fifth full-connection layer FC_5;

if the probability is larger than a preset first threshold value, determining that the electric load charged by the electric bicycle exists in the user.

Example two

Fig. 3 is a flowchart of a method for detecting an electric load demand of an electric vehicle according to a second embodiment of the present invention, where an alarm operation is added on the basis of the embodiment. As shown in fig. 3, the method includes:

step 301, collecting original current data for power consumers belonging to resident life power consumption in a power distribution network.

And step 302, clustering the users according to the electricity consumption behavior to obtain user categories.

Step 303, intercepting part of the original current data according to the user category as target current data.

Step 304, loading the charging detection network according to the user category.

The charging detection network comprises an encoder, a decoder for training the user category and a classifier.

Step 305, encoding the target current data in the encoder to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged.

Step 306, decoding the first target current characteristics in the decoder to reconstruct in sequence a plurality of levels of second target current characteristics characterizing the user class when there is electric bicycle charging.

Step 307, in the classifier, under the enhancement effect of the second target current characteristics of the multiple levels, detecting whether the electric consumer has an electric load charged by the electric bicycle according to the first target current characteristics.

Step 308, counting the proportion of the electric load of the electric bicycle in the preset time period for the user.

In this embodiment, it may be detected for multiple times, for example, when the electric bicycle charging is detected for the first time, the timing of the starting period (for example, 60 min) is started, and during the timing, the target current data is intercepted from the original current data at certain intervals (for example, 1 min) to detect whether the electric bicycle charging is present.

During this period, the statistical result is the proportion of the electric load of the electric bicycle charging in all the results.

Step 309, if the ratio is greater than the preset second threshold, generating alarm information for the user.

Comparing the proportion of the electric load charged by the electric bicycle in all results with a preset second threshold value, if the proportion of the electric load charged by the electric bicycle in all results is larger than the preset second threshold value, the probability of the electric bicycle charging behavior of the user is larger, the electric bicycle charging behavior of the user can be confirmed, the false detection rate is reduced, at the moment, alarm information can be generated for the user, and the user can be checked by a designated manager.

Example III

Fig. 4 is a schematic structural diagram of an electric load demand detection device for an electric vehicle according to a third embodiment of the present invention. As shown in fig. 4, the apparatus includes:

the original current data acquisition module 401 is used for acquiring original current data for power consumers belonging to resident life power consumption in the power distribution network;

the electricity utilization user clustering module 402 is used for clustering the electricity utilization users according to electricity utilization behaviors to obtain user categories;

a target current data intercepting module 403, configured to intercept a portion of the original current data according to the user category, as target current data;

a charging detection network loading module 404, configured to load a charging detection network according to the user category, where the charging detection network includes an encoder, a decoder for training the user category, and a classifier;

the encoding module 405 is configured to encode the target current data in the encoder, so as to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged;

a decoding module 406, configured to decode the first target current characteristic in the decoder, so as to reconstruct, in sequence, a plurality of levels of second target current characteristics that characterize the user class when there is electric bicycle charging;

The classification module 407 is configured to detect, in the classifier, whether the user has an electric load for charging the electric bicycle according to the first target current feature under the enhancement effect of the second target current features of multiple levels.

In one embodiment of the present invention, the user clustering module 402 is further configured to:

loading a plurality of pre-trained clusters representing user categories, each cluster having a center point therein;

collecting first electricity behavior data which are irrelevant to the charging of the electric bicycle and second electricity behavior data which are relevant to the charging of the electric bicycle for the user;

converting the first electrical behavior data into first vectors and the second electrical behavior data into second vectors, respectively;

splicing the first vector and the second vector into a third vector;

calculating the distance between the third vector and each center point;

dividing the user into the clusters with the minimum distance, and setting the user category represented by the clusters as the user category to which the user belongs.

In one embodiment of the present invention, the target current data interception module 403 is further configured to:

Loading a window and a downsampling factor configured for the user category, wherein the length of the window covers a first inflection point and a second inflection point when a user in the user category charges an electric bicycle, the first inflection point is an intersection point between a constant-current stage and a constant-voltage stage, and the second inflection point is an intersection point between the constant-voltage stage and a trickle stage;

loading the window on the original current data, and intercepting part of data in the window as candidate current data;

and performing downsampling operation on the candidate current data according to the downsampling factor to obtain target current data with uniform length.

In one embodiment of the invention, the encoder comprises a long and short term memory network, a first convolution layer, a second convolution layer, a third convolution layer, a first pooling layer, and a second pooling layer;

the encoding module 405 is further configured to:

performing time sequence processing on the target current data in the long-short-period memory network to obtain a first reference current characteristic;

performing convolution operation on the first reference current characteristic in the first convolution layer to obtain a second reference current characteristic;

performing a minimum pooling operation on the second reference current characteristic in the first pooling layer to obtain a third reference current characteristic;

Performing convolution operation on the third reference current characteristic in the second convolution layer to obtain a fourth reference current characteristic;

performing convolution operation on the fourth reference current characteristic in the third convolution layer to obtain a fifth reference current characteristic;

and performing a minimum pooling operation on the fifth reference current characteristic in the second pooling layer to obtain a first target current characteristic for distinguishing whether electric bicycle charging exists or not.

In one embodiment of the invention, the encoder includes a first deconvolution layer, a second deconvolution layer, and a third deconvolution layer;

the decoding module 406 is further configured to:

performing deconvolution operation on the first target current characteristic in the first deconvolution layer to obtain a second target current characteristic of a first level;

performing deconvolution operation on the second target current characteristic of the first level in the second deconvolution layer to obtain a second target current characteristic of the second level;

and performing deconvolution operation on the second target current characteristic of the second level in the third deconvolution layer to obtain the second target current characteristic of the third level.

In one embodiment of the present invention, the classifier includes a first fully connected layer, a second fully connected layer, a third fully connected layer, a fourth fully connected layer, a fifth fully connected layer, a first separable convolutional layer, a second separable convolutional layer, and a third separable convolutional layer;

The classification module 407 is further configured to:

mapping the second target current feature of a first level to a first candidate current feature in the first fully connected layer;

fusing the first target current feature and the first candidate current feature into a first fused current feature;

performing separable convolution operation on the first fusion current characteristic in the first separable convolution layer to obtain a first separation current characteristic;

mapping the second target current feature of a second level to a second candidate current feature in the second fully connected layer;

fusing the first separation current feature and the second candidate current feature into a second fused current feature;

performing separable convolution operation on the second fusion current characteristic in the second separable convolution layer to obtain a second separation current characteristic;

mapping the second target current feature of a third level to a third candidate current feature in the third fully connected layer;

fusing the second separation current feature and the third candidate current feature to a third fused current feature;

performing separable convolution operation on the third fusion current feature in the third separable convolution layer to obtain a third separable current feature;

Mapping the third split current feature to a fourth split current feature in the fourth fully connected layer;

mapping the fourth separation current characteristic to the probability that the electric bicycle charges exist for the user in the fifth full connection layer;

In one embodiment of the present invention, further comprising:

the proportion statistics module is used for counting the proportion of the electric loads of the electric bicycle in a preset time period;

and the alarm information generation module is used for generating alarm information for the user if the proportion is larger than a preset second threshold value.

The electric load demand detection device for the electric vehicle provided by the embodiment of the invention can execute the electric load demand detection method for the electric vehicle provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the electric load demand detection method for the electric vehicle.

Example III

Fig. 5 shows a schematic diagram of the structure of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.

As shown in fig. 5, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.

Various components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, etc.; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the electric vehicle electrical load demand detection method.

In some embodiments, the electric vehicle electrical load demand detection method may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the electric load demand detection method for an electric vehicle described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the electric vehicle electrical load demand detection method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.

The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.

Example N

The embodiment of the invention also provides a computer program product, which comprises a computer program, and the computer program realizes the electric load demand detection method for the electric vehicle according to any embodiment of the invention when being executed by a processor.

Computer program product in the implementation, the computer program code for carrying out operations of the present invention may be written in one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.

The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.

Claims

1. The method for detecting the electric load demand of the electric vehicle is characterized by comprising the following steps of:

in the classifier, under the enhancement effect of the second target current characteristics of a plurality of levels, detecting whether the electric consumer has an electric load for charging the electric bicycle according to the first target current characteristics;

the encoder comprises a long-term memory network, a first convolution layer, a second convolution layer, a third convolution layer, a first pooling layer and a second pooling layer;

the encoding of the target current data in the encoder to obtain a first target current characteristic for distinguishing whether the electric bicycle is charged or not includes:

performing a minimum pooling operation on the fifth reference current feature in the second pooling layer to obtain a first target current feature for distinguishing whether electric bicycle charging exists;

the decoder comprises a first deconvolution layer, a second deconvolution layer and a third deconvolution layer;

decoding the first target current characteristic in the decoder to reconstruct, in sequence, a plurality of levels of second target current characteristics characterizing the user class as there is electric bicycle charging, comprising:

2. The method of claim 1, wherein clustering the users according to the electricity usage behavior to obtain a user category comprises:

splicing the first vector and the second vector into a third vector;

calculating the distance between the third vector and each center point;

3. The method of claim 1, wherein said intercepting a portion of said raw current data as target current data in accordance with said user category comprises:

4. The method of claim 1, wherein the classifier comprises a first fully connected layer, a second fully connected layer, a third fully connected layer, a fourth fully connected layer, a fifth fully connected layer, a first separable convolutional layer, a second separable convolutional layer, and a third separable convolutional layer;

in the classifier, under the enhancement effect of the second target current characteristics of a plurality of levels, detecting whether the electric bicycle is charged by the user according to the first target current characteristics, including:

5. A method according to any one of claims 1-3, further comprising:

counting the proportion of the electric loads of the electric bicycle in a preset time period;

and if the proportion is larger than a preset second threshold value, generating alarm information for the user.

6. An electric vehicle electrical load demand detection device, comprising:

the classification module is used for detecting whether the electric consumer has an electric load for charging the electric bicycle according to the first target current characteristic under the enhancement effect of the second target current characteristics of a plurality of levels in the classifier;

the encoding module is further configured to:

the decoding module is further configured to:

7. An electronic device, the electronic device comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the electric vehicle electrical load demand detection method of any one of claims 1-5.

8. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program for causing a processor to execute the electric load demand detection method for an electric vehicle according to any one of claims 1 to 5.