using_ai_to_discover_explainable_industrial_analytics
Differences
This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
| using_ai_to_discover_explainable_industrial_analytics [2026/08/02 22:07] – [Technical summary] wikiadmin | using_ai_to_discover_explainable_industrial_analytics [2026/08/02 22:22] (current) – [The engineering process] wikiadmin | ||
|---|---|---|---|
| Line 19: | Line 19: | ||
| The AI discovers the analytic. | The AI discovers the analytic. | ||
| - | The analytic performs the monitoring. | + | The analytic performs the monitoring. |
| ===== Why this matters ===== | ===== Why this matters ===== | ||
| Line 27: | Line 27: | ||
| Traditionally there have been two common approaches: | Traditionally there have been two common approaches: | ||
| - | * Engineers manually explore the data until they discover something useful. | + | |
| - | * Machine learning attempts to discover patterns automatically from large labelled datasets. | + | * Machine learning attempts to discover patterns automatically from large labelled datasets. |
| Both approaches have drawbacks. | Both approaches have drawbacks. | ||
| Line 56: | Line 56: | ||
| The original paper investigates ML techniques because the data contains: | The original paper investigates ML techniques because the data contains: | ||
| - | * multiple machines | + | |
| - | * different cutting operations | + | * different cutting operations |
| - | * changing conditions over several years | + | * changing conditions over several years |
| - | * relatively few damaged examples | + | * relatively few damaged examples |
| Instead of training a model, we approached the data conversationally. | Instead of training a model, we approached the data conversationally. | ||
| Line 81: | Line 81: | ||
| Questions included: | Questions included: | ||
| - | * Does overall vibration amplitude separate good and bad tools? | + | |
| - | * Does energy move into particular frequency ranges? | + | * Does energy move into particular frequency ranges? |
| - | * Does the spindle speed matter? | + | * Does the spindle speed matter? |
| - | * Are the damaged tools exciting different mechanical behaviour? | + | * Are the damaged tools exciting different mechanical behaviour? |
| - | * Which measurements remain stable across different machines? | + | * Which measurements remain stable across different machines? |
| Most of these ideas were tested and rejected. | Most of these ideas were tested and rejected. | ||
| Line 107: | Line 107: | ||
| There is: | There is: | ||
| - | * no training phase | + | |
| - | * no statistical model | + | * no statistical model |
| - | * no neural network | + | * no neural network |
| - | * no weights | + | * no weights |
| - | * no inference engine | + | * no inference engine |
| Just conventional engineering mathematics based on the frequency content of the vibration signal. | Just conventional engineering mathematics based on the frequency content of the vibration signal. | ||
| Line 149: | Line 149: | ||
| That means: | That means: | ||
| - | * engineers can understand every calculation | + | |
| - | * results can be independently verified | + | * results can be independently verified |
| - | * thresholds can be tuned | + | * thresholds can be tuned |
| - | * failures can be investigated | + | * failures can be investigated |
| - | * behaviour | + | * behavior |
| {{ pasted: | {{ pasted: | ||
| Line 182: | Line 182: | ||
| Rather than expecting an AI to reason directly over millions of raw sensor samples, we can combine: | Rather than expecting an AI to reason directly over millions of raw sensor samples, we can combine: | ||
| - | * engineering context | + | |
| - | * deterministic analytics | + | * deterministic analytics |
| - | * conversational AI | + | * conversational AI |
| The LLM helps engineers discover better analytics. | The LLM helps engineers discover better analytics. | ||
| Line 198: | Line 198: | ||
| The workflow becomes: | The workflow becomes: | ||
| - | * Collect operational data. | + | |
| - | * Use conversational AI to explore hypotheses. | + | * Use conversational AI to explore hypotheses. |
| - | * Validate discoveries against real engineering data. | + | * Validate discoveries against real engineering data. |
| - | * Convert successful ideas into deterministic analytics. | + | * Convert successful ideas into deterministic analytics. |
| - | * Deploy those analytics in production. | + | * Deploy those analytics in production. |
| - | * Feed the resulting engineering features back into higher-level AI applications. | + | * Feed the resulting engineering features back into higher-level AI applications. |
| Rather than asking an LLM to analyse millions of sensor values every day, production systems execute simple, explainable calculations while AI continues helping engineers discover the next improvement. | Rather than asking an LLM to analyse millions of sensor values every day, production systems execute simple, explainable calculations while AI continues helping engineers discover the next improvement. | ||
| Line 211: | Line 211: | ||
| For readers interested in the implementation details: | For readers interested in the implementation details: | ||
| - | | + | * The analysis uses public Bosch CNC vibration recordings sampled from tri-axial accelerometers described in this paper [[https:// |
| - | | + | * Vibration is transformed into the frequency domain using standard Fast Fourier Transform (FFT) techniques. |
| - | | + | * Each machining run is scored by measuring the proportion of vibration energy around the spindle' |
| - | | + | * Runs are ranked automatically using MIStudio' |
| - | | + | * The highest-ranked run is compared with healthy reference runs using the Comparator. |
| - | | + | * Charts and reports are generated directly from the extracted features without custom processing. |
| ===== Conclusions ===== | ===== Conclusions ===== | ||
using_ai_to_discover_explainable_industrial_analytics.1785722827.txt.gz · Last modified: by wikiadmin