Transcoding and storage method and system for high-dimensional time-series energy storage data

By dividing high-dimensional time-series data into column groups according to the semantic type of physical quantities and using differential encoding, the problem of high-dimensional time-series data being difficult to compress is solved, and the optimization of storage space and transmission bandwidth is achieved.

CN122133606APending Publication Date: 2026-06-02新源智储能源发展(北京)有限公司 +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
新源智储能源发展(北京)有限公司
Filing Date
2026-01-22
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

The row-based storage format of high-dimensional time-series data is difficult to compress and does not match the direct write mode of cloud databases, resulting in low efficiency in storage space and network transmission bandwidth.

Method used

The single-row records of high-dimensional time series data are divided into column groups according to the semantic type of physical quantities, and different columns are encoded differently to generate multi-column format files which are sent to cloud storage. These include differential encoding, dictionary encoding, and GORILLA encoding for the serial number column, measurement point column, and numerical column.

Benefits of technology

Transforming high-dimensional row-format storage into smaller column-format files reduces storage overhead and transmission bandwidth requirements, and improves data compression efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122133606A_ABST
    Figure CN122133606A_ABST
Patent Text Reader

Abstract

This invention provides a transcoding and storage method and system for high-dimensional time-series energy storage data. The method involves acquiring a single-line record from a high-dimensional time-series data file, where each line includes a timestamp and multiple heterogeneous raw numerical points. Based on the physical quantity semantic type of each raw numerical point, the raw numerical points are divided into at least one column group, where raw numerical points within the same column group have the same physical quantity semantics and data type. Differential encoding is applied to each column based on its data characteristics. A multi-column format file is generated based on the timestamp and the encoded values ​​in each corresponding column, and this multi-column format file is sent to cloud storage. This invention can transform high-dimensional data stored in row format into a smaller column format file, solving the problems of overhead caused by the difficulty in compressing row data and insufficient transmission bandwidth. Combined with the encoding and compression advantages of column-based storage, it can reduce storage overhead and save transmission bandwidth for high-dimensional data.
Need to check novelty before this filing date? Find Prior Art