Input data and information sources
Operational data can reveal information about operators and, in fleets, drivers. Equipment and production data can be trade secrets.
Predictive maintenance
Also known as: Predictive maintenance and condition monitoring
The system detects impending machine or equipment failure from sensor and operational data so maintenance can be planned in time.
Understand the use case
What defines this use case?
The model learns normal operation, flags deviations or estimates when a component will fail. Deutsche Bahn monitors points this way; Utrecht uses it to plan vehicle fleet maintenance.
This page describes a typical system design. Actual data flows and permissions depend on your implementation.
- System design
- Analysis and processing
- Ability to act
- Action permissions depend on integration
- Fields of use
- railways, public administration
- Also known as
- condition monitoring, condition monitoring, industrial anomaly detection
How it works, step by step
System design and data flows
Input data
Sensors and controllers continually supply vibration, temperature, current draw or switching time measurements.
Some systems analyse vibration, temperature, sound and images together. Operational data can reveal information about operators and, in fleets, drivers.
Processing
A model compares measurements with learned normal operation and estimates whether and when failure is likely.
The model processes measurements and operating histories of equipment to detect deviations or estimate future maintenance needs. The estimate is combined with maintenance rules and available resources. An indication of wear and an actual maintenance work order are different process steps.
Results
Unusual assets generate alerts with suggested maintenance timing.
The result describes the defined model task. A match, category or forecast acquires its professional significance only in conjunction with the underlying data and intended use.
Use and impact
Warnings help maintenance teams plan staff and parts. Without warning, equipment runs until the next scheduled inspection.
Integration determines whether the result is simply displayed, adopted after approval or written directly into a business application. These variants differ in their data flows and permissions.
Explore the system design: Select a component to see its task, incoming and outgoing data, and role in the use case. The guided walkthrough explains each phase. Then use Yes, Uncertain and No to represent your specific system variant.
4 review first14 clarify7 baseline threats
- Present
- Depends on your system
- Absent
- Review first
- To clarify
How to use the model
The diagram shows the typical architecture. Select a component to see what happens there and adapt its features to your system. Enable “Analysis guidance” to see which threats need reviewing for this architecture. The page's checklist continues to describe the typical case.
Changes from the typical case
None yet. Each change will appear here.
Where most threats arise
- 4 clarify
- 2 review first, 1 clarify
- 3 clarify
- 3 clarify
In this system design: 4 to review first, 14 to clarify, 7 baseline threats.
Architecture as a list
External access: Depends on your system
To clarify
Input: Part of every AI system
To clarify
Baseline for every AI system
AI model: Part of every AI system
Review first
- Model Drift & Degradation Review priority
To clarify
Provider service: Depends on your system
To clarify
Training and feedback: Present
Review first
- Training Data Poisoning Review priority
- Backdoor ML Model
To clarify
Result: Part of every AI system
Baseline for every AI system
Business application or equipment: Depends on your system
To clarify
person: Part of every AI system
Review first
- Overreliance Review priority
Unapproved AI services: Depends on your system
To clarify
Supply chain and technical foundations: 5 baseline threats applying to every AI system
Baseline for every AI system
How it differs from similar use cases
Visual and acoustic quality inspection
Inspects manufactured components and products rather than equipment condition
AI-based cyberattack detection
Searches network, account and email data for signs of attacks
Data and information
Processing and operating environment
Some deployers develop models themselves; others use a provider's platform.
Storage and further use
Models for individual assets generally learn from those assets' own measurement series. Maintenance feedback on whether a warning was justified can teach the model.
Results and data sharing
Depending on integration, it creates maintenance orders in the maintenance system itself or only suggests them. Cloud platforms send equipment data outside the organisation's environment.
Using a provider does not automatically determine where processing takes place. Storage location, subcontractors, deletion periods and use for training must be checked for the specific deployment.
Assess risks and determine measures
From understanding the system to analysing risks
- Define the system boundary
Record data, participants, interfaces and permissions. Resolve uncertain characteristics in the model.
- Investigate threats
Relate review candidates to actual data flows: what could happen, and who would be affected?
- Assess risk
Assess impacts and likelihood in the deployment context. The number of review candidates is not a risk score.
- Provide evidence of measures
Select appropriate controls, assign responsibilities and define effectiveness checks. Document residual risks and approvals.
Key threats in the application context
- 3 particularly important
- 1 typical
- 14 depending on the system
- 7 additional threats for every AI system
A total of 25 threats from the threat catalogue may be relevant to this use case.
Priority review points
Modifications, replaced parts or frost alter measurements without implying damage. Until adapted, the model issues false alarms or misses real deviations.
Training measurements pass through plant controllers and networks. Undetected manipulation can teach the model that a damaged condition is normal, preventing later warnings.
Maintenance teams relying on predictions may extend fixed inspection intervals. Damage no sensor detects becomes apparent only at failure, disrupting operations and potentially endangering people at points or energy facilities.
1 further review candidates in the typical design
Other relevant threats
Training data or an adopted base model can introduce hidden triggers for misbehaviour.
Listed because of this characteristic
- Learns from your own data
Up to 14 further threats may apply, depending on your system design. Nine threats need to be reviewed for every AI system. Two of these are particularly important here and are therefore listed above.
System features and open questions
Each Yes adds further threats to your checklist, up to 14 in total. If you cannot yet answer a question, include its threats in your review.
- Added by this Yes
How your system is built
- Processes confidential data3 threats
- Processes personal data2 threats
- Processes text, images and speech together2 threats
- Changes data or triggers processes1 threat
- Learns from examples with correct answers1 threat
- Uses operational feedback1 threat
- Optimises a specified metric1 threat
How you operate the system
- Available to people outside the organisation5 threats
- Accepts queries directly through an interface4 threats
- Accessible over a network3 threats
- Incurs costs with each query1 threat
- Passes content to services outside your environment1 threat
- Shares computing resources with other customers or applications1 threat
- Used alongside unapproved AI services1 threat
How your system is built
Can the system change things itself, such as storing data, sending messages, initiating transactions or granting permissions?
Depending on integration, it creates maintenance orders in the maintenance system itself or only suggests them.
If yes, add
Does the system process information about identifiable people, including in inputs, outputs or logs?
Operational data can reveal information about operators and, in fleets, drivers.
If yes, add
Does the system process confidential information such as trade secrets, health or financial data?
Equipment and production data can be trade secrets.
If yes, add
Is the model trained on examples to which someone has assigned the correct answer?
Anomaly detection can work without labelled examples; failure prediction learns from documented damage.
If yes, add
Does operational feedback or output later feed into analysis or training?
Maintenance feedback on whether a warning was justified can teach the model.
If yes, add
Does the system process several input types together, such as text with images or speech?
Some systems analyse vibration, temperature, sound and images together.
If yes, add
Is the system intended to maximise a specified metric, such as revenue, utilisation or clicks?
A system planning maintenance windows itself pursues objectives such as high availability or low costs.
If yes, add
How you operate the system
These questions arise for almost every AI system.
Can people or programs query the model directly, for example through an interface?
If yes, add
Is the system accessible over a network or dependent on network services?
If yes, add
Do you pay per query or unit of computing time, so costs increase with use?
If yes, add
Can content leave your environment, for example to a cloud provider, web search or email?
Cloud platforms send equipment data outside the organisation's environment.
If yes, add
Does the system share servers or computing resources with other customers or applications?
If yes, add
Can people outside your organisation use the system, including without login?
If yes, add
Do staff also use AI services for this task that your organisation has not approved?
If yes, add
Controls and implementation
Measures for the priority review areas are expanded. Each linked measure explains how to implement it.
Review priorityModel Drift & Degradation3 measures
Review priorityTraining Data Poisoning2 measures
Review priorityOverreliance3 measures
Backdoor ML Model2 measures
Controls for the remaining threats are available on their pages in the threat catalogue.
AI Act provisions
-
Art. 4: AI literacy
For every AI system Applies since 2 February 2025.
-
Annex III no. 2: Critical infrastructure
Depending on use Review trigger from 2 December 2027.
-
Art. 6 para. 1, Art. 2 para. 2, Annex I Section B: Vehicles, aircraft, ships and railway assets
Depending on use Review trigger from 2 August 2028.
Applies to every AI system
Art. 4: AI literacy
Applies if staff or appointed representatives operate or use AI systems.
Applies since 2 February 2025.
Obligations and details for Art. 4: AI literacy
- Who bears the obligations
- Provider, Deployer, Product manufacturer, Provider of a general-purpose AI model
Detailed obligations
May apply in particular uses
Annex III no. 2: Critical infrastructure
Applies if the system acts as a safety component in electricity, gas, heating or water supply installations, for example where warnings determine safe continued operation.
Applies from 2 December 2027.
Obligations and details for Annex III no. 2: Critical infrastructure
- The general condition is
- Applies if the system acts as a safety component in critical digital infrastructure, road traffic or water, gas, heating or electricity supply.
- Points to consider
- A fundamental rights impact assessment under Article 27 is not required here.
- Who bears the obligations
- Provider, Deployer, Product manufacturer
Detailed obligations
Art. 6 para. 1, Art. 2 para. 2, Annex I Section B: Vehicles, aircraft, ships and railway assets
Applies if it acts as a safety component of railway assets such as points. Annex I, Section B, and railway legislation then apply, rather than Annex III, point 2.
Applies from 2 August 2028.
Obligations and details for Art. 6 para. 1, Art. 2 para. 2, Annex I Section B: Vehicles, aircraft, ships and railway assets
- The general condition is
- Applies if the system is a safety component of a vehicle, aircraft, ship or railway asset covered by Annex I, Section B.
- Points to consider
- The AI Act applies through type approval or authorisation legislation here; its other chapters do not apply directly.
- Who bears the obligations
- Provider, Deployer, Product manufacturer
Detailed obligations
Classify your system with the quick checker All obligations in the obligations compass
Cross-cutting threats
Review these threats for every AI system, whatever its purpose.
Atypical system features
View 16 further questions
If any of these features does apply to your system, add the listed threats.
Does the system generate its own text, such as answers, drafts or summaries?
The system produces condition values, warnings and maintenance dates, rather than text.
If yes, add
Does the system generate or edit images, sounds, voices or videos?
If yes, add
Does the system create new content from open-ended requests that people read, see or hear directly? Translation, reading aloud and transcription do not count.
If yes, add
Can users give the system natural language tasks that it then performs?
Prediction uses measurements and does not accept natural language instructions.
If yes, add
Does the system search its own collection of documents or records for relevant content?
If yes, add
Does the system formulate answers from retrieved documents?
If yes, add
Does the system use a knowledge graph storing concepts and their relationships?
If yes, add
Does the system remember content from earlier conversations or tasks and reuse it later?
If yes, add
Does the system call other programs, interfaces or search services itself, even if only to read?
If yes, add
Does the system determine its next steps itself based on earlier results?
If yes, add
Does the system delegate tasks to other AI agents or incorporate their results?
If yes, add
Is the system connected to tools or data sources through the Model Context Protocol (MCP)?
If yes, add
Does the system process content outsiders may have written or altered, such as emails, uploads or websites?
Measurements come from the organisation's own sensors and controllers.
If yes, add
Do documents from several customers, departments or permission groups share one search store?
If yes, add
Does the system generate code, scripts or database queries for further use?
If yes, add
Does the system execute programs, scripts or commands itself?
If yes, add
Sources
- Report of a deployment
DB Engineering & Consulting, Predictive Maintenance: DIANA analysis and diagnosis platform at DB Netz AG According to the deployer
The rail subsidiary's project description cites over 40,000 connected point-machine sensors and failure prediction as the platform's purpose.
- Register or case collection
Dutch Algorithm Register, Onderhoud wagenpark, Gemeente Utrecht
The entry describes an operational algorithm using vehicle and maintenance data to suggest timing, with city staff making the decision.
- Manufacturer
Siemens, Press release on Senseye Predictive Maintenance, 5 February 2024 According to the manufacturer
The manufacturer announces generative AI enhancements intended for future use at BlueScope.
- Specialist study
OECD, Framework for the Classification of AI Systems, 2022, Annex A
The framework lists predictive maintenance as a main AI application.
Related use cases
Also in the field of Manufacturing, retail and logistics
- Visual and acoustic quality inspectionCameras or microphones capture components and products on the production line; a model checks defects, completeness and correct variants.
- Sales and demand forecastingRetailers use a model to predict how much of each item each store will sell, then turn those forecasts into suggested orders or completed orders.
Updated: 30 September 2026. Review checklist rules version 0.3.2, threat catalogue version 2026.10.24.2. How these pages are structured