Utility Data Anonymization Using Hashing and Endpoint Swapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Utility datasets containing personal identifiable information (PII) pose challenges for utility companies as they need to maintain privacy while preserving valuable grid topology and consumption data, and conventional methods often strip away useful information.
Innovation Solution
A two-phase anonymization technique involving a one-way hashing algorithm for encrypting identifiers and a mix-and-match methodology to swap data between endpoints, while maintaining topology and consumption information, ensuring PII removal is irreversible.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional approaches for removing PII are used, then privacy protection is improved, but useful information such as grid topology information and equipment specifications is stripped away
Solution Approach 1:
The patent segments the data anonymization process into distinct phases: (1) attaching pseudonyms to PII fields while preserving original values in a separate mapping structure, (2) swapping only the PII portions between records while maintaining non-PII fields intact, and (3) maintaining topology information separately. This segmentation allows selective anonymization of only the harmful PII portions without destroying useful grid topology and equipment specification data.
Solution Approach 2:
The patent introduces a pseudonym mapping structure as an intermediary between the original PII data and the anonymized output. This intermediary mapping table stores the correspondence between original PII values and their pseudonyms, allowing the system to replace PII with pseudonyms while preserving the ability to maintain relationships between records through the mapping structure, thus preventing information loss.
2Object-affected harmful factors
If PII is removed from utility datasets, then privacy protection is improved, but the ability to perform accurate analysis and prediction is worsened
Solution Approach 1:
The patent creates a copy of the original data structure where PII fields are replaced with pseudonyms rather than completely removed. The pseudonym mapping structure serves as a reference that preserves the original PII values and their relationships. This copying approach maintains the structural integrity and analytical value of the data while replacing only the sensitive PII portions, enabling accurate analysis without compromising privacy.
Solution Approach 2:
The patent applies local quality by differentiating between PII fields and non-PII fields in the utility dataset. Only the PII portions are anonymized through pseudonym replacement, while non-PII fields such as grid topology information, equipment specifications, and consumption data retain their original values and relationships. This selective local anonymization preserves the quality and accuracy needed for analysis while removing privacy risks.
3Object-affected harmful factors
If data swapping between endpoints is performed to anonymize PII, then privacy protection is improved, but the complexity of the anonymization process is worsened
Solution Approach 1:
The patent performs preliminary actions by first attaching pseudonyms to all PII fields and creating the pseudonym mapping structure before executing the data swapping operation. This preliminary preparation simplifies the subsequent swapping process, as the system only needs to exchange pseudonym values rather than reconstructing entire records. The preliminary pseudonym attachment phase separates the complex anonymization logic into manageable steps, reducing overall process complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The technique efficiently preserves utility dataset features by detaching real consumption data from its original location, maintaining topology for visualization, and enhancing security against reverse-engineering, thus providing accurate and secure anonymized datasets for analysis and commercial use.
Implementation Method 1
A first operation may involve assigning anonymous identifiers to various nodes in the dataset (e.g., assigning a new anonymous identifier to each house associated with a particular neighborhood's transformer).
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A data anonymization technique for datasets including personal identifiable information (PII) such as utility datasets may primarily include assigning anonymous identifiers to nodes in the dataset and swapping or otherwise moving portions of information between nodes of the dataset. The methodology of the swapping or moving operation may vary optionally based on a number of parameters, and may include swapping endpoints under a single parent, swapping endpoints between similar parents, and/or swapping similar endpoints between parents. The anonymization technique may output an anonymized dataset which reflects a topology of the original dataset and may optionally be updatable and modifiable to include additional data about existing or new nodes.