Energy storage battery working voltage monitoring method and system

Abstract

The invention provides an energy storage battery working voltage monitoring method and system, and the method comprises the steps: generating a coding list of cells, and constructing a data collection network matched with an energy storage battery in real time according to a plurality of cells based on the coding list; detecting a plurality of acquisition channels included in the data acquisition network in real time, and constructing a corresponding dimension reduction matrix in real time according to the plurality of acquisition channels and the plurality of battery cells; collecting a channel voltage corresponding to each collection channel in real time through a preset analog front-end chip, and converting the channel voltage into an initial voltage vector with a first vector dimension in real time; and performing real-time dimension reduction processing on the initial voltage vector through a dimension reduction matrix to correspondingly generate a target voltage vector with a second vector dimension, extracting a plurality of vector elements correspondingly contained in the target voltage vector in real time, and setting each vector element as a single voltage corresponding to each battery cell. According to the invention, the single voltage of the battery cell can be accurately detected, and the detection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage battery monitoring, and in particular to a method and system for monitoring the operating voltage of an energy storage battery. Background Art

[0002] In the field of energy storage battery voltage monitoring technology, accurately obtaining the operating voltage of individual cells is a key prerequisite for ensuring the safe operation of battery systems, optimizing charge and discharge strategies, and assessing battery health. With the rapid development of the new energy industry, energy storage batteries are being used in an increasingly diverse range of scenarios, from small household energy storage to large-scale grid-scale energy storage power stations. This places stringent demands on the accuracy, real-time nature, and reliability of voltage monitoring.

[0003] Among them, the analog front-end chip (AFE) is a key component responsible for voltage acquisition in the battery management system (BMS). Its performance directly determines the validity of the monitoring data and has now become the mainstream voltage sampling solution.

[0004] Furthermore, in actual applications, since existing AFE chips all use single-channel voltage sampling, the lower limit of the sampling voltage is usually higher than a certain value. However, some types of existing energy storage cells will enter a certain low-voltage section during normal operation, resulting in the existing AFE being unable to directly sample the single-cell voltage of such low-voltage cells, thereby having certain sampling limitations. At the same time, the corresponding voltage monitoring data generated is also incomplete, which correspondingly reduces the monitoring efficiency of the energy storage battery operating voltage. Summary of the Invention

[0005] Based on this, the purpose of the present invention is to provide a method and system for monitoring the working voltage of an energy storage battery to solve the problem that the existing technology cannot directly sample the single cell voltage of a low-voltage battery cell, resulting in certain sampling limitations and correspondingly reducing the efficiency of energy storage battery working voltage monitoring.

[0006] The first aspect of the embodiment of the present invention proposes: A method for monitoring the operating voltage of an energy storage battery, wherein the method comprises: Sequentially encode a plurality of battery cells connected in series within an energy storage battery to generate a corresponding code list, and construct a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; Detecting in real time a number of acquisition channels included in the data acquisition network, and constructing in real time a corresponding dimensionality reduction matrix according to the number of acquisition channels and the number of battery cells; Acquire channel voltages corresponding to each acquisition channel in real time through a preset analog front-end chip, and convert the channel voltages into an initial voltage vector having a first vector dimension in real time; The initial voltage vector is subjected to real-time dimensionality reduction processing by the dimensionality reduction matrix to generate a corresponding target voltage vector having a second vector dimension, and a number of corresponding vector elements contained in the target voltage vector are extracted in real time, and each of the vector elements is set as a single cell voltage corresponding to each of the battery cells.

[0007] The beneficial effect of the present invention is that by encoding the battery cells inside the energy storage battery in real time, a data acquisition network for reasonably collecting voltage data can be created in real time. Based on this, an acquisition channel and a dimensionality reduction matrix for subsequent dimensionality reduction processing can be created in real time. Based on this, the initial voltage vector collected in real time can be subjected to real-time dimensionality reduction processing, and a final target voltage vector can be generated. Based on this, the vector elements in the current target voltage vector can ultimately be set as the actual single cell voltage of each battery cell, thereby eliminating the limitations of monitoring and correspondingly improving the monitoring efficiency of the working voltage of the energy storage battery.

[0008] Furthermore, the step of constructing a corresponding dimensionality reduction matrix in real time according to the plurality of acquisition channels and the plurality of battery cells includes: Detecting in real time the cell serial number corresponding to each cell in the coding list, and synchronously recording the identification information corresponding to each acquisition channel, each acquisition channel corresponding to one or more cells; A corresponding matrix frame is created in real time with the total number of the battery cell serial numbers as the number of rows and the total number of the acquisition channels as the number of columns, wherein the row index of the matrix frame corresponds to the battery cell serial number and the column index corresponds to the identification information; A correspondence table between the battery cells and the acquisition channels is created in real time, and the matrix framework is filled in real time according to the correspondence table to generate the dimensionality reduction matrix accordingly.

[0009] Furthermore, the step of performing real-time filling processing on the matrix framework according to the correspondence table to generate the dimensionality reduction matrix includes: Setting a matrix element filling rule for the matrix frame, wherein the matrix element filling rule includes filling a preset non-zero value at an intersection of the battery cell number and the index of the corresponding acquisition channel in the matrix frame if there is a monitoring relationship between the acquisition channel and the battery cell, and filling a zero value otherwise; Traversing each mapping relationship in the corresponding relationship table, and filling corresponding positions of the matrix frame point by point according to the mapping relationship and the matrix element filling rule, so as to form an initial matrix containing zero values ​​and non-zero values; The initial matrix is ​​normalized to generate the dimension-reduced matrix accordingly.

[0010] Furthermore, the step of converting the channel voltage into an initial voltage vector having a first vector dimension in real time includes: The raw voltage signal collected by the acquisition channel is subjected to noise filtering and outlier correction, high-frequency interference is eliminated by a sliding average algorithm, and abnormal voltage values ​​outside the normal range are eliminated based on the 3σ criterion to obtain standardized channel voltage data; Detecting in real time the physical topological order corresponding to the plurality of acquisition channels, and creating a corresponding dimension index table according to the physical topological order; The channel voltage data is mapped to a preset dimensional space in real time according to the dimensional index table, so as to correspondingly form the initial voltage vector including the spatiotemporal feature markers.

[0011] Furthermore, the step of mapping the channel voltage data to a preset dimensional space in real time according to the dimensional index table to correspondingly form the initial voltage vector containing spatiotemporal feature markers includes: Assigning a unique spatial code to each of the acquisition channels, synchronously acquiring the timestamp of the channel voltage data, and fusing the spatial code and the timestamp through a hash algorithm to generate a feature tag with a fixed length; Automatically adjust the direction of the basis vectors of the preset dimensional space according to the number of series-connected cells and the channel distribution density in the data acquisition network; The channel voltage data is used as a vector element value and is sequentially filled into the interior of the preset dimensional space according to the direction of the basis vector to generate a corresponding metadata field, and the feature label is correspondingly embedded into the interior of the metadata field to generate the initial voltage vector.

[0012] Furthermore, the step of performing real-time dimensionality reduction processing on the initial voltage vector by using the dimensionality reduction matrix to correspondingly generate a target voltage vector having a second vector dimension includes: Analyzing the row and column distribution characteristics of the dimension reduction matrix in real time, and adjusting the operation dimension of the dimension reduction matrix according to the dimension parameter of the initial voltage vector; Obtaining health status parameters of the battery cell, and performing weighted correction on the dimensionality reduction matrix using the health status parameters to generate a corresponding target dimensionality reduction matrix; The initial voltage vector is subjected to multiple rounds of iterative operations using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space, and the target voltage vector is generated accordingly.

[0013] Furthermore, the step of performing multiple rounds of iterative operations on the initial voltage vector using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space and correspondingly generating the target voltage vector includes: Performing an iterative convergence operation on the initial voltage vector using the target dimensionality reduction matrix, and projecting the converged result into the interior of the low-dimensional space; Smoothing the projection results within the low-dimensional space using a Kalman filter algorithm to form corresponding target vector elements; The target vector elements are temperature compensated and calibrated according to the real-time operating temperature of the battery cell, and the temperature-calibrated target vector elements are integrated to generate the target voltage vector.

[0014] The second aspect of the embodiment of the present invention proposes: A system for monitoring the operating voltage of an energy storage battery, wherein the system comprises: A construction module is used to sequentially encode a plurality of battery cells connected in series within the energy storage battery to generate a corresponding code list, and to construct a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; A detection module, configured to detect in real time a plurality of acquisition channels included in the data acquisition network, and construct in real time a corresponding dimensionality reduction matrix according to the plurality of acquisition channels and the plurality of battery cells; a conversion module, configured to acquire a channel voltage corresponding to each acquisition channel in real time through a preset analog front-end chip, and convert the channel voltage into an initial voltage vector having a first vector dimension in real time; An extraction module is used to perform real-time dimensionality reduction processing on the initial voltage vector through the dimensionality reduction matrix to generate a target voltage vector having a second vector dimension, and to extract a number of corresponding vector elements contained in the target voltage vector in real time, and to set each of the vector elements as a single cell voltage corresponding to each of the battery cells.

[0015] Furthermore, the detection module is specifically used to: Detecting in real time the cell serial number corresponding to each cell in the coding list, and synchronously recording the identification information corresponding to each acquisition channel, each acquisition channel corresponding to one or more cells; A corresponding matrix frame is created in real time with the total number of the battery cell serial numbers as the number of rows and the total number of the acquisition channels as the number of columns, wherein the row index of the matrix frame corresponds to the battery cell serial number and the column index corresponds to the identification information; A correspondence table between the battery cells and the acquisition channels is created in real time, and the matrix framework is filled in real time according to the correspondence table to generate the dimensionality reduction matrix accordingly.

[0016] Furthermore, the detection module is specifically used to: Setting a matrix element filling rule for the matrix frame, wherein the matrix element filling rule includes filling a preset non-zero value at an intersection of the battery cell number and the index of the corresponding acquisition channel in the matrix frame if there is a monitoring relationship between the acquisition channel and the battery cell, and filling a zero value otherwise; Traversing each mapping relationship in the corresponding relationship table, and filling corresponding positions of the matrix frame point by point according to the mapping relationship and the matrix element filling rule, so as to form an initial matrix containing zero values ​​and non-zero values; The initial matrix is ​​normalized to generate the dimension-reduced matrix accordingly.

[0017] Furthermore, the conversion module is specifically used to: The raw voltage signal collected by the acquisition channel is subjected to noise filtering and outlier correction, high-frequency interference is eliminated by a sliding average algorithm, and abnormal voltage values ​​outside the normal range are eliminated based on the 3σ criterion to obtain standardized channel voltage data; Detecting in real time the physical topological order corresponding to the plurality of acquisition channels, and creating a corresponding dimension index table according to the physical topological order; The channel voltage data is mapped to a preset dimensional space in real time according to the dimensional index table, so as to correspondingly form the initial voltage vector including the spatiotemporal feature markers.

[0018] Furthermore, the conversion module is specifically used to: Assigning a unique spatial code to each of the acquisition channels, synchronously acquiring the timestamp of the channel voltage data, and fusing the spatial code and the timestamp through a hash algorithm to generate a feature tag with a fixed length; Automatically adjust the direction of the basis vectors of the preset dimensional space according to the number of series-connected cells and the channel distribution density in the data acquisition network; The channel voltage data is used as a vector element value and is sequentially filled into the interior of the preset dimensional space according to the direction of the basis vector to generate a corresponding metadata field, and the feature label is correspondingly embedded into the interior of the metadata field to generate the initial voltage vector.

[0019] Furthermore, the extraction module is specifically used to: Analyzing the row and column distribution characteristics of the dimension reduction matrix in real time, and adjusting the operation dimension of the dimension reduction matrix according to the dimension parameter of the initial voltage vector; Obtaining health status parameters of the battery cell, and performing weighted correction on the dimensionality reduction matrix using the health status parameters to generate a corresponding target dimensionality reduction matrix; The initial voltage vector is subjected to multiple rounds of iterative operations using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space, and the target voltage vector is generated accordingly.

[0020] Furthermore, the extraction module is specifically used to: Performing an iterative convergence operation on the initial voltage vector using the target dimensionality reduction matrix, and projecting the converged result into the interior of the low-dimensional space; Smoothing the projection results within the low-dimensional space using a Kalman filter algorithm to form corresponding target vector elements; The target vector elements are temperature compensated and calibrated according to the real-time operating temperature of the battery cell, and the temperature-calibrated target vector elements are integrated to generate the target voltage vector.

[0021] The third aspect of the embodiment of the present invention proposes: A computer comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the energy storage battery operating voltage monitoring method as described above when executing the computer program.

[0022] The fourth aspect of the embodiments of the present invention proposes: A readable storage medium stores a computer program thereon, wherein when the program is executed by a processor, the method for monitoring the operating voltage of an energy storage battery as described above is implemented.

[0023] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flowchart of a method for monitoring the operating voltage of an energy storage battery provided by the first embodiment of the present invention; Figure 2 This is a structural block diagram of the energy storage battery operating voltage monitoring system provided by the third embodiment of the present invention.

[0025] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0026] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0027] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be an intermediate element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or there may be an intermediate element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] See also Figure 1 , shown is a method for monitoring the operating voltage of an energy storage battery provided by a first embodiment of the present invention. The method for monitoring the operating voltage of an energy storage battery provided by this embodiment can accurately detect the single cell voltage of each battery cell inside the energy storage battery, thereby accurately determining the operating voltage of the energy storage battery, thereby correspondingly improving the detection efficiency of the energy storage battery.

[0030] Specifically, this embodiment provides: A method for monitoring the operating voltage of an energy storage battery comprises the following steps: Step S10, sequentially encoding a plurality of battery cells connected in series within the energy storage battery to generate a corresponding code list, and constructing a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; Among them, it should be noted that the interior of the existing energy storage battery is composed of several battery cells connected in series or in parallel, and each battery cell is a separate body and can work independently. Based on this, in order to be able to objectively and accurately monitor the working voltage of the energy storage battery, it is possible to objectively and accurately collect the single cell voltage of each battery cell, so that the working voltage of the energy storage battery can be accurately determined. Based on this, in order to accurately collect the single cell voltage of each battery cell, the present invention will first encode the several battery cells connected in series in sequence inside the current energy storage battery, and can generate a corresponding code list, such as 1#, 2#, 3#...n#, so as to ensure that each battery cell can be uniquely identified. Based on this, in order to facilitate subsequent detection, the present invention will be based on the code A mapping relationship between each cell code and the physical position of the cell is established, and the rated voltage, capacity and series quantity parameters of several of the cells are extracted in real time to generate a cell basic database containing cell location information and electrical parameters, and according to the series quantity parameters in the cell basic database, the cells are dynamically grouped according to preset grouping rules (such as 3-5 cells per group) to obtain several cell groups, and at least two independent acquisition channels are allocated to each cell group, where one channel is used to collect the total voltage of the cell group, and the other channel is used to collect the single cell voltage of the specified cell in the group, so that a data acquisition network adapted to the current energy storage battery can be constructed in real time. This structure can cover the voltage acquisition requirements of multiple cells at the same time (such as 1 acquisition channel corresponding to 2-3 adjacent cells), balance sampling accuracy and hardware cost, and facilitate subsequent processing.

[0031] Step S20, detecting in real time a number of acquisition channels included in the data acquisition network, and constructing in real time a corresponding dimensionality reduction matrix according to the number of acquisition channels and the number of battery cells; It should be noted that after the required data acquisition network is constructed in real time through the above steps, all acquisition channels actually contained in the data acquisition network (such as CH1, CH2, ..., CHm) are immediately identified, and the physical location and coverage range of each channel are clarified. In addition, based on the correspondence between the acquisition channels and battery cells (such as CH1 covers battery cells 1#-2#, and CH2 covers battery cells 2#-3#), a dimensionality reduction matrix is ​​constructed with battery cell numbers as rows and acquisition channels as columns. The matrix elements are non-zero only when there is an association between the channels and the battery cells, and are zero otherwise. The present invention establishes a mathematical association between the acquisition channels and the battery cells through the above dimensionality reduction matrix, providing an operational carrier for the subsequent dimensional conversion of the voltage data to facilitate subsequent processing.

[0032] Step S30, collecting the channel voltage corresponding to each of the collection channels in real time through a preset analog front-end chip, and converting the channel voltage into an initial voltage vector having a first vector dimension in real time; It should be noted that after the required dimensionality reduction matrix is ​​created in real time through the above steps, the channel voltage of each acquisition channel is immediately acquired in real time using the existing high-precision ADC chip. The channel voltages of all acquisition channels are arranged in channel order to form an initial voltage vector with a first vector dimension (number of dimensions = number of acquisition channels m). Each element in the vector corresponds to the voltage value of an acquisition channel, thereby converting the discrete channel voltage data into a standardized vector form for subsequent processing. It should be noted that in order to objectively and accurately create the required initial voltage vector, the detection order of the acquisition channels is used as the vector dimension order, and the digital voltage value of each channel is used as the "vector element" to construct a 1×m row vector (m is the number of acquisition channels, i.e., the first vector dimension). Specifically, for example, if there are three acquisition channels, channel 1 has a voltage of 3.2V, channel 2 has a voltage of 3.3V, and channel 3 has a voltage of 3.1V, then the initial voltage vector is [3.2, 3.3, 3.1]. Its core function is to convert the discrete channel voltage data into a mathematical form that conforms to matrix operation rules for subsequent processing.

[0033] In step S40, the initial voltage vector is subjected to real-time dimensionality reduction processing by the dimensionality reduction matrix to generate a target voltage vector having a second vector dimension, and a number of vector elements corresponding to the target voltage vector are extracted in real time, and each of the vector elements is set as a single cell voltage corresponding to each of the battery cells.

[0034] It should be noted that after the required dimensionality reduction matrix and initial voltage vector are obtained through the above steps, a matrix multiplication operation is immediately performed on the initial voltage vector using the dimensionality reduction matrix. Through the association relationship of non-zero elements, the high-dimensional (m-dimensional) initial voltage vector is reduced to a target voltage vector having a second vector dimension (number of dimensions = number of battery cells n), and each vector element in the target voltage vector is extracted. Each element corresponds to the single cell voltage of a battery cell (because the dimensionality reduction matrix has established a precise association between channels and battery cells, the vector elements after dimensionality reduction can be directly mapped to a single battery cell). Based on this, the single cell voltage of each battery cell can be inferred from the multi-channel mixed voltage data, realizing one-by-one monitoring of the series-connected batteries, so that the single cell voltage of each battery cell can be objectively and accurately calculated. Based on this, the current operating voltage of the energy storage battery can be accurately determined, which correspondingly improves the monitoring efficiency of the energy storage battery. Among them, it should be pointed out that after obtaining the dimensionality reduction matrix and the initial voltage vector respectively through the above steps, the initial voltage vector will be operated with the current dimensionality reduction matrix at this time, so as to obtain the above-mentioned target voltage vector, that is, a 1×n row vector. Based on this, each vector element of the target voltage vector is corresponded to each battery cell in turn according to the order of the above-mentioned battery cell coding list, and the element value is set to the single cell voltage of the corresponding battery cell, completing the conversion of the mathematical vector to the physical voltage value, so as to facilitate subsequent processing.

[0035] Second embodiment Furthermore, the step of constructing a corresponding dimensionality reduction matrix in real time according to the plurality of acquisition channels and the plurality of battery cells includes: Detecting in real time the cell serial number corresponding to each cell in the coding list, and synchronously recording the identification information corresponding to each acquisition channel, each acquisition channel corresponding to one or more cells; A corresponding matrix frame is created in real time with the total number of the battery cell serial numbers as the number of rows and the total number of the acquisition channels as the number of columns, wherein the row index of the matrix frame corresponds to the battery cell serial number and the column index corresponds to the identification information; A correspondence table between the battery cells and the acquisition channels is created in real time, and the matrix framework is filled in real time according to the correspondence table to generate the dimensionality reduction matrix accordingly.

[0036] Among them, it should be noted that in order to objectively and accurately construct a dimensionality reduction matrix adapted to the current number of battery cells, specifically, the present invention will extract the unique battery cell serial number corresponding to each battery cell from the code list (such as an integer from 1 to n), assign unique identification information to each acquisition channel (such as CH1, CH2 or digital codes 1 to m), and clarify the battery cell range covered by each channel (such as CH1 corresponds to 1# and 2# batteries). Based on this, the total number of battery cell serial numbers n is the number of matrix rows, and the total number of acquisition channels m is the number of matrix columns. A blank matrix frame is created, and the row index of the matrix frame is directly corresponded to the battery cell serial number (the 1st row corresponds to 1# battery cell, the 2nd row corresponds to 2# battery cell, ..., the nth row corresponds to n# battery cell), and the column index directly corresponds to the identification information of the acquisition channel (the 1st column corresponds to CH1, the 2nd column corresponds to CH2, ..., the mth column corresponds to CHm). Based on this, the basic structure of the dimensionality reduction matrix can be determined, the mapping relationship between rows and columns and batteries and channels can be clarified, and data dislocation can be avoided. On this basis, the present invention will sort out all "acquisition channels-battery cells" A structured correspondence table is formed based on the mapping relationship (for example, the table records that CH1 corresponds to cells 1# and 2#, and CH2 corresponds to cells 2# and 3#). According to the correspondence table, elements are filled in the matrix frame (non-zero values ​​are filled if there is a correlation, and zero values ​​are filled if there is no correlation). Finally, a dimensionality reduction matrix is ​​generated (most elements in the matrix are zero, and only a small number of associated positions are non-zero values) to facilitate subsequent processing.

[0037] Furthermore, the step of performing real-time filling processing on the matrix framework according to the correspondence table to generate the dimensionality reduction matrix includes: Setting a matrix element filling rule for the matrix frame, wherein the matrix element filling rule includes filling a preset non-zero value at an intersection of the battery cell number and the index of the corresponding acquisition channel in the matrix frame if there is a monitoring relationship between the acquisition channel and the battery cell, and filling a zero value otherwise; Traversing each mapping relationship in the corresponding relationship table, and filling corresponding positions of the matrix frame point by point according to the mapping relationship and the matrix element filling rule, so as to form an initial matrix containing zero values ​​and non-zero values; The initial matrix is ​​normalized to generate the dimension-reduced matrix accordingly.

[0038] It should be noted that after obtaining the required correspondence table and matrix framework through the above steps, the filling standard of the matrix elements will be clarified first. Specifically, if a certain acquisition channel has a monitoring association with a certain battery cell (that is, the channel covers the battery cell), the intersection of "battery cell number row-channel identification column" in the matrix frame is filled with a preset non-zero value (such as 1 or the voltage contribution coefficient of the channel to the battery cell); if there is no association, it is filled with zero value, and each "channel-battery cell" in the correspondence table is traversed row by row and column by column. Mapping relationship (such as first processing the association between CH1 and 1# and 2# cells, and then processing the association between CH2 and 2# and 3# cells), according to the filling rules, fill in zero values ​​or non-zero values ​​in the corresponding positions of the matrix frame to form an initial matrix containing zero values ​​and non-zero values. In addition, it should be pointed out that after the required initial matrix is ​​obtained in real time, the non-zero values ​​in the initial matrix will be standardized (such as dividing the non-zero values ​​by the number of cells covered by the channel to ensure that the single cell voltage can be accurately inferred during voltage calculation), while keeping the zero value unchanged. This can eliminate the error caused by the different number of cells covered by the channel, ensure the calculation accuracy of the dimensionality reduction matrix, and finally generate a dimensionality reduction matrix that can be directly used for dimensionality reduction for subsequent processing.

[0039] Furthermore, the step of converting the channel voltage into an initial voltage vector having a first vector dimension in real time includes: The raw voltage signal collected by the acquisition channel is subjected to noise filtering and outlier correction, high-frequency interference is eliminated by a sliding average algorithm, and abnormal voltage values ​​outside the normal range are eliminated based on the 3σ criterion to obtain standardized channel voltage data; Detecting in real time the physical topological order corresponding to the plurality of acquisition channels, and creating a corresponding dimension index table according to the physical topological order; The channel voltage data is mapped to a preset dimensional space in real time according to the dimensional index table, so as to correspondingly form the initial voltage vector including the spatiotemporal feature markers.

[0040] It should be noted that after the required channel voltage is collected in real time through the above steps, the current channel voltage will be immediately analyzed and processed. Specifically, the present invention will use a sliding average algorithm (such as a 5-point sliding average) to eliminate high-frequency interference (such as circuit noise, electromagnetic interference) on the original voltage signal collected by the acquisition channel to make the voltage data more stable. Based on this, based on the 3σ criterion (that is, eliminating voltage values ​​outside the range of "average value ± 3 times the standard deviation"), abnormal voltage data (such as extreme values ​​caused by sensor failure) is identified and eliminated, and normal voltage values ​​at adjacent moments are interpolated and replaced to obtain standardized channel voltage data, which can improve the reliability of channel voltage data and avoid the influence of noise and abnormal values ​​on subsequent vector conversion. In addition, the present invention will also determine the physical arrangement order of all acquisition channels in the energy storage battery (such as the channel order from left to right according to the direction of series connection of the battery cells), and create a dimension index table according to the physical topological order to clarify the position index of each acquisition channel in the subsequent vector (such as the first channel in the physical order corresponds to the first position of the vector, and the second channel corresponds to the second position of the vector). Based on this, a one-dimensional vector space matching the number of acquisition channels will be eventually constructed (the number of dimensions = the number of acquisition channels m), and the standardized channel voltage data will be filled into the preset dimensional space one by one according to the position index of the dimension index table to form an initial voltage vector. Each element in the vector contains spatiotemporal feature markers such as acquisition time and channel position, which can convert the preprocessed channel voltage data into a structured vector form, providing standardized input for subsequent dimensionality reduction operations to facilitate subsequent processing.

[0041] Furthermore, the step of mapping the channel voltage data to a preset dimensional space in real time according to the dimensional index table to correspondingly form the initial voltage vector containing spatiotemporal feature markers includes: Assigning a unique spatial code to each of the acquisition channels, synchronously acquiring the timestamp of the channel voltage data, and fusing the spatial code and the timestamp through a hash algorithm to generate a feature tag with a fixed length; Automatically adjust the direction of the basis vectors of the preset dimensional space according to the number of series-connected cells and the channel distribution density in the data acquisition network; The channel voltage data is used as a vector element value and is sequentially filled into the interior of the preset dimensional space according to the direction of the basis vector to generate a corresponding metadata field, and the feature label is correspondingly embedded into the interior of the metadata field to generate the initial voltage vector.

[0042] It should be noted that after the required channel voltage data is obtained in real time through the above steps, a unique spatial code (such as a coordinate code based on the physical position of the channel) is assigned to each acquisition channel to identify the spatial position of the channel, and the acquisition timestamp of each channel voltage data is synchronously recorded (accurate to milliseconds) to reflect the time attribute of the data. The spatial code and the timestamp are fused through a hash algorithm (such as SHA-1) to generate a feature tag of a fixed length (such as 32 bits) to uniquely identify the spatiotemporal information of the channel voltage data. According to the number of series connected cells (n) and the channel distribution density (such as 1 channel for every 2 cells or 1 channel for every 3 cells) in the data acquisition network, the spatial code and the timestamp are fused to generate a feature tag of a fixed length (such as 32 bits) to uniquely identify the spatiotemporal information of the channel voltage data. Channels), automatically adjust the basis vector direction of the preset dimensional space (to ensure that the basis vector is consistent with the voltage change trend of the channel coverage range), and finally use the channel voltage data as the vector element value. According to the adjusted basis vector direction, the corresponding positions in the preset dimensional space are filled in sequence to generate metadata fields (including voltage value and position index). The generated spatiotemporal feature label is embedded in the head or tail of the corresponding metadata field to form a vector element containing spatiotemporal information. Finally, all elements are integrated to generate the initial voltage vector. This can ensure that the initial voltage vector contains not only voltage data but also carries complete spatiotemporal attributes, providing support for subsequent data tracing and anomaly analysis, so that the required initial voltage vector can be accurately obtained for subsequent processing.

[0043] Furthermore, the step of performing real-time dimensionality reduction processing on the initial voltage vector by using the dimensionality reduction matrix to correspondingly generate a target voltage vector having a second vector dimension includes: Analyzing the row and column distribution characteristics of the dimension reduction matrix in real time, and adjusting the operation dimension of the dimension reduction matrix according to the dimension parameter of the initial voltage vector; Obtaining health status parameters of the battery cell, and performing weighted correction on the dimensionality reduction matrix using the health status parameters to generate a corresponding target dimensionality reduction matrix; The initial voltage vector is subjected to multiple rounds of iterative operations using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space, and the target voltage vector is generated accordingly.

[0044] Among them, it should be noted that after the required dimensionality reduction matrix and the initial voltage vector are obtained respectively through the above steps, the corresponding dimensionality reduction processing will be performed immediately at this time. Specifically, the row and column distribution characteristics of the dimensionality reduction matrix (such as the number of rows n and the number of columns m) are analyzed to determine its operation dimension. If the dimension of the initial voltage vector (m-dimensional) is inconsistent with the number of columns (m) of the dimensionality reduction matrix, the vector dimension is adjusted by zero padding or truncation (according to the actual sampling situation) to ensure that the two can perform matrix multiplication operations (the vector dimension must match the number of matrix columns). Based on this, the health status parameters (SOH, such as capacity decay rate and internal resistance change rate) of each battery cell are obtained through the battery management system (BMS), and the non-zero elements of the dimensionality reduction matrix are weighted according to the SOH parameters (such as batteries with lower SOH, the non-zero values ​​of the corresponding associated channels are appropriately increased to compensate for their voltage acquisition errors). The target dimensionality reduction matrix is ​​generated, and finally the initial voltage vector and the target dimensionality reduction matrix are subjected to multiple rounds of matrix multiplication iterative operations (such as 3-5 Rounds of iteration are performed to gradually eliminate data redundancy and errors, and the results after iterative convergence are projected into a low-dimensional space with a dimension of n (the number of battery cells), forming a target voltage vector with a dimension consistent with the number of battery cells. It should be noted that through iterative optimization and dimensionality compression, the single cell voltage is accurately inferred from the mixed channel voltage, completing the dimensionality reduction process to facilitate subsequent processing.

[0045] Furthermore, the step of performing multiple rounds of iterative operations on the initial voltage vector using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space and correspondingly generating the target voltage vector includes: Performing an iterative convergence operation on the initial voltage vector using the target dimensionality reduction matrix, and projecting the converged result into the interior of the low-dimensional space; Smoothing the projection results within the low-dimensional space using a Kalman filter algorithm to form corresponding target vector elements; The target vector elements are temperature compensated and calibrated according to the real-time operating temperature of the battery cell, and the temperature-calibrated target vector elements are integrated to generate the target voltage vector.

[0046] It should be noted that, in order to quickly and effectively complete multiple rounds of iterative operations on the initial voltage vector, the present invention will use the target dimensionality reduction matrix to iteratively operate on the initial voltage vector, calculate the error value (such as the difference between the current result and the previous round result) after each round of operation, and stop the iteration when the error value is less than a preset threshold (such as 0.001V) to achieve result convergence, and project the converged operation result into a low-dimensional space (dimension number = n), obtain preliminary projection results (including the estimated voltage value of each battery cell). The projection results in the low-dimensional space are smoothed using the Kalman filter algorithm to eliminate random noise (such as voltage jumps caused by sampling fluctuations) and obtain more stable target vector elements. Finally, the real-time operating temperature of the battery cell is obtained (collected by the temperature sensor). According to the temperature-voltage characteristic curve (for example, for every 1°C increase in temperature, the voltage decreases by 0.001V), the target vector elements are temperature compensated to correct the voltage deviation caused by temperature. All target vector elements after temperature calibration are arranged in order of the battery cell serial number to generate the final target voltage vector. Each element in the vector is the precise single-cell voltage of the corresponding battery cell. Based on this, it can be ensured that the target voltage vector can truly reflect the operating voltage status of the battery cell, thereby improving the monitoring efficiency of the energy storage battery.

[0047] See also Figure 2 , the third embodiment of the present invention provides: A system for monitoring the operating voltage of an energy storage battery, wherein the system comprises: A construction module is used to sequentially encode a plurality of battery cells connected in series within the energy storage battery to generate a corresponding code list, and to construct a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; A detection module, configured to detect in real time a plurality of acquisition channels included in the data acquisition network, and construct in real time a corresponding dimensionality reduction matrix according to the plurality of acquisition channels and the plurality of battery cells; a conversion module, configured to acquire a channel voltage corresponding to each acquisition channel in real time through a preset analog front-end chip, and convert the channel voltage into an initial voltage vector having a first vector dimension in real time; An extraction module is used to perform real-time dimensionality reduction processing on the initial voltage vector through the dimensionality reduction matrix to generate a target voltage vector having a second vector dimension, and to extract a number of corresponding vector elements contained in the target voltage vector in real time, and to set each of the vector elements as a single cell voltage corresponding to each of the battery cells.

[0048] Furthermore, the detection module is specifically used to: Detecting in real time the cell serial number corresponding to each cell in the coding list, and synchronously recording the identification information corresponding to each acquisition channel, each acquisition channel corresponding to one or more cells; A corresponding matrix frame is created in real time with the total number of the battery cell serial numbers as the number of rows and the total number of the acquisition channels as the number of columns, wherein the row index of the matrix frame corresponds to the battery cell serial number and the column index corresponds to the identification information; A correspondence table between the battery cells and the acquisition channels is created in real time, and the matrix framework is filled in real time according to the correspondence table to generate the dimensionality reduction matrix accordingly.

[0049] Furthermore, the detection module is specifically used to: Setting a matrix element filling rule for the matrix frame, wherein the matrix element filling rule includes filling a preset non-zero value at an intersection of the battery cell number and the index of the corresponding acquisition channel in the matrix frame if there is a monitoring relationship between the acquisition channel and the battery cell, and filling a zero value otherwise; Traversing each mapping relationship in the corresponding relationship table, and filling corresponding positions of the matrix frame point by point according to the mapping relationship and the matrix element filling rule, so as to form an initial matrix containing zero values ​​and non-zero values; The initial matrix is ​​normalized to generate the dimension-reduced matrix accordingly.

[0050] Furthermore, the conversion module is specifically used to: The raw voltage signal collected by the acquisition channel is subjected to noise filtering and outlier correction, high-frequency interference is eliminated by a sliding average algorithm, and abnormal voltage values ​​outside the normal range are eliminated based on the 3σ criterion to obtain standardized channel voltage data; Detecting in real time the physical topological order corresponding to the plurality of acquisition channels, and creating a corresponding dimension index table according to the physical topological order; The channel voltage data is mapped to a preset dimensional space in real time according to the dimensional index table, so as to correspondingly form the initial voltage vector including the spatiotemporal feature markers.

[0051] Furthermore, the conversion module is specifically used to: Assigning a unique spatial code to each of the acquisition channels, synchronously acquiring the timestamp of the channel voltage data, and fusing the spatial code and the timestamp through a hash algorithm to generate a feature tag with a fixed length; Automatically adjust the direction of the basis vectors of the preset dimensional space according to the number of series-connected cells and the channel distribution density in the data acquisition network; The channel voltage data is used as a vector element value and is sequentially filled into the interior of the preset dimensional space according to the direction of the basis vector to generate a corresponding metadata field, and the feature label is correspondingly embedded into the interior of the metadata field to generate the initial voltage vector.

[0052] Furthermore, the extraction module is specifically used to: Analyzing the row and column distribution characteristics of the dimension reduction matrix in real time, and adjusting the operation dimension of the dimension reduction matrix according to the dimension parameter of the initial voltage vector; Obtaining health status parameters of the battery cell, and performing weighted correction on the dimensionality reduction matrix using the health status parameters to generate a corresponding target dimensionality reduction matrix; The initial voltage vector is subjected to multiple rounds of iterative operations using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space, and the target voltage vector is generated accordingly.

[0053] Furthermore, the extraction module is specifically used to: Performing an iterative convergence operation on the initial voltage vector using the target dimensionality reduction matrix, and projecting the converged result into the interior of the low-dimensional space; Smoothing the projection results within the low-dimensional space using a Kalman filter algorithm to form corresponding target vector elements; The target vector elements are temperature compensated and calibrated according to the real-time operating temperature of the battery cell, and the temperature-calibrated target vector elements are integrated to generate the target voltage vector.

[0054] A fourth embodiment of the present invention provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the energy storage battery operating voltage monitoring method described above when executing the computer program.

[0055] A fifth embodiment of the present invention provides a readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method for monitoring the operating voltage of an energy storage battery as described above is implemented.

[0056] In summary, the energy storage battery operating voltage monitoring method and system provided by the above embodiments of the present invention can accurately detect the single cell voltage of each battery cell inside the energy storage battery, thereby accurately determining the operating voltage of the energy storage battery, which correspondingly improves the monitoring efficiency of the energy storage battery.

[0057] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0058] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0059] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0060] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0061] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0062] The above-described embodiments merely illustrate several embodiments of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for monitoring the operating voltage of an energy storage battery, characterized in that: The method comprises: Sequentially encode a plurality of battery cells connected in series within an energy storage battery to generate a corresponding code list, and construct a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; Detecting in real time a number of acquisition channels included in the data acquisition network, and constructing in real time a corresponding dimensionality reduction matrix according to the number of acquisition channels and the number of battery cells; Acquire channel voltages corresponding to each acquisition channel in real time through a preset analog front-end chip, and convert the channel voltages into an initial voltage vector having a first vector dimension in real time; The initial voltage vector is subjected to real-time dimensionality reduction processing by the dimensionality reduction matrix to generate a corresponding target voltage vector having a second vector dimension, and a number of corresponding vector elements contained in the target voltage vector are extracted in real time, and each of the vector elements is set as a single cell voltage corresponding to each of the battery cells.

2. The method for monitoring the operating voltage of an energy storage battery according to claim 1, wherein: The step of constructing a corresponding dimensionality reduction matrix in real time according to the plurality of acquisition channels and the plurality of battery cells comprises: Detecting in real time the cell serial number corresponding to each cell in the coding list, and synchronously recording the identification information corresponding to each acquisition channel, each acquisition channel corresponding to one or more cells; A corresponding matrix frame is created in real time with the total number of the battery cell serial numbers as the number of rows and the total number of the acquisition channels as the number of columns, wherein the row index of the matrix frame corresponds to the battery cell serial number and the column index corresponds to the identification information; A correspondence table between the battery cells and the acquisition channels is created in real time, and the matrix framework is filled in real time according to the correspondence table to generate the dimensionality reduction matrix accordingly.

3. The method for monitoring the operating voltage of an energy storage battery according to claim 2, wherein: The step of performing real-time filling processing on the matrix framework according to the corresponding relationship table to generate the dimensionality reduction matrix accordingly includes: Setting a matrix element filling rule for the matrix frame, wherein the matrix element filling rule includes filling a preset non-zero value at an intersection of the battery cell number and the index of the corresponding acquisition channel in the matrix frame if there is a monitoring relationship between the acquisition channel and the battery cell, and filling a zero value otherwise; Traversing each mapping relationship in the corresponding relationship table, and filling corresponding positions of the matrix frame point by point according to the mapping relationship and the matrix element filling rule, so as to form an initial matrix containing zero values ​​and non-zero values; The initial matrix is ​​normalized to generate the dimension-reduced matrix accordingly.

4. The method for monitoring the operating voltage of an energy storage battery according to claim 1, wherein: The step of converting the channel voltage into an initial voltage vector having a first vector dimension in real time comprises: The raw voltage signal collected by the acquisition channel is subjected to noise filtering and outlier correction, high-frequency interference is eliminated by a sliding average algorithm, and abnormal voltage values ​​outside the normal range are eliminated based on the 3σ criterion to obtain standardized channel voltage data; Detecting in real time the physical topological order corresponding to the plurality of acquisition channels, and creating a corresponding dimension index table according to the physical topological order; The channel voltage data is mapped to a preset dimensional space in real time according to the dimensional index table, so as to correspondingly form the initial voltage vector including the spatiotemporal feature markers.

5. The method for monitoring the operating voltage of an energy storage battery according to claim 4, wherein: The step of mapping the channel voltage data to a preset dimensional space in real time according to the dimensional index table to correspondingly form the initial voltage vector containing spatiotemporal feature markers includes: Assigning a unique spatial code to each of the acquisition channels, synchronously acquiring the timestamp of the channel voltage data, and fusing the spatial code and the timestamp through a hash algorithm to generate a feature tag with a fixed length; Automatically adjust the direction of the basis vectors of the preset dimensional space according to the number of series-connected cells and the channel distribution density in the data acquisition network; The channel voltage data is used as a vector element value and is sequentially filled into the interior of the preset dimensional space according to the direction of the basis vector to generate a corresponding metadata field, and the feature label is correspondingly embedded into the interior of the metadata field to generate the initial voltage vector.

6. The method for monitoring the operating voltage of an energy storage battery according to claim 1, wherein: The step of performing real-time dimensionality reduction processing on the initial voltage vector by using the dimensionality reduction matrix to correspondingly generate a target voltage vector having a second vector dimension includes: Analyzing the row and column distribution characteristics of the dimension reduction matrix in real time, and adjusting the operation dimension of the dimension reduction matrix according to the dimension parameter of the initial voltage vector; Obtaining health status parameters of the battery cell, and performing weighted correction on the dimensionality reduction matrix using the health status parameters to generate a corresponding target dimensionality reduction matrix; The initial voltage vector is subjected to multiple rounds of iterative operations using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space, and the target voltage vector is generated accordingly.

7. The method for monitoring the operating voltage of an energy storage battery according to claim 6, wherein: The step of performing multiple rounds of iterative operations on the initial voltage vector using the target dimensionality reduction matrix to project the initial voltage vector into a low-dimensional space and correspondingly generating the target voltage vector includes: Performing an iterative convergence operation on the initial voltage vector using the target dimensionality reduction matrix, and projecting the converged result into the interior of the low-dimensional space; Smoothing the projection results within the low-dimensional space using a Kalman filter algorithm to form corresponding target vector elements; The target vector elements are temperature compensated and calibrated according to the real-time operating temperature of the battery cell, and the temperature-calibrated target vector elements are integrated to generate the target voltage vector.

8. A storage battery operating voltage monitoring system, characterized in that: The system comprises: A construction module is used to sequentially encode a plurality of battery cells connected in series within the energy storage battery to generate a corresponding code list, and to construct a data acquisition network adapted to the energy storage battery in real time based on the code list and the plurality of battery cells; A detection module, configured to detect in real time a plurality of acquisition channels included in the data acquisition network, and construct in real time a corresponding dimensionality reduction matrix according to the plurality of acquisition channels and the plurality of battery cells; a conversion module, configured to acquire a channel voltage corresponding to each acquisition channel in real time through a preset analog front-end chip, and convert the channel voltage into an initial voltage vector having a first vector dimension in real time; An extraction module is used to perform real-time dimensionality reduction processing on the initial voltage vector through the dimensionality reduction matrix to generate a target voltage vector having a second vector dimension, and to extract a number of corresponding vector elements contained in the target voltage vector in real time, and to set each of the vector elements as a single cell voltage corresponding to each of the battery cells.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for monitoring the operating voltage of the energy storage battery according to any one of claims 1 to 7 is implemented.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for monitoring the operating voltage of an energy storage battery as described in any one of claims 1 to 7 is implemented.

Images