Smart solution for condition-based monitoring

Smart Limit Watcher provides real time machine monitoring and alerting to improve your plant preventive maintenance initiatives.

Implement artificial intelligence in manufacturing without a background in data science.
The 5-step wizard evaluates available data sources, applies the best fit model from our library, and trains and tests the model automatically.
Models use normal machine operating data, including cyclical process changes, and do not require fault or failure data.
Once configured, the Smart LImit Watcher continuously monitors the identified variables that affect the overall effectiveness of your machine and alerts you to process drifts that could compromise equipment health or product quality.

Proven statistical models

moneo’s Smart Limit Watcher uses 3 proven and widely accepted statistical algorithm models constructed specifically for industrial automation.

LassoCV: Linear Regression
Simple dynamic functions with clear distinctions of the target and support variables
Effective for many applications
Limitations in complex data relationships

LightGMB: Decision tree
More complex data relationship technique than linear regression
Highly optimized histogram-based decision tree learning algorithms such as Gradient-Based One-Side Sampling (GOSS) and Exclusive Feature Bundling (EFB)
Approximates continuous functions that have a direct correlation but are non-linear

MLPRegressor: Neural Net
Highest level of complexity
Multi-layer perceptron (MLP) learns using backpropagation to compute a gradient descent with respect to weights and approximates all functions with a correlation

But you don't need to be an expert on these algorithm models! Using its graphical interface designed specifically for industrial automation, the Smart Limit Watcher guides you through the entire process of assigning the necessary measurements to automatically implement the correct model to your application. The guided tour video below demonstrates this simplicity.

Guided tour: moneo Smart Limit Watcher

Clear and understandable alerting

Smart Limit Watcher automatically sets up confidence bands based on data from "good condition" training.

No need to simulate fault conditions.
Deviations from the allowable ranges triggers automated warnings and alarms.
Sensitivity ranges are user-adjustable for fine tuning to specific requirements.

Screenshot of the Smart LImit Watcher showing automatic setting of alarm limits

Benefits

Easy: No data science expertise required. Simple implementation for production and maintenance personnel.

Edge deployable: Compact edge algorithm models do not require powerful processors and memory allowing processing to be run close to the point of use.

Straightforward : Automated data preparation and quality check. No time-consuming data pre-processing required.

Intelligent: Selects the best-fit AI model based on inputs. Automatic model training and monitoring accuracy verification.

Reliable: Time and condition-based monitoring. Permanent background monitoring using dynamic predicted ranges for the measurement values.

Application methods

There are two main application methods for which the Smart Limit Watcher can be applied to most industrial machines.

Supportive (Vertical) measurements

Related measurements of different types on a single machine that affect its operation
Generalist approach suitable for most applications
Evaluation required to determine which parameters influence mechanical health
- Target variable is identified as the main parameter to monitor
- Support variables influence the health of the machine to varying degrees

Comparative (Horizontal) measurements

The same measurements from multiple devices of the same type
Devices are phyically connected and operate in unison on the machine
Potentially fewer sensors required than in the supportive method

Example applications

Compression system condition monitoring
Method: Supportive (Vertical) measurements

The Smart Limit Watcher tracks the mechanical health using the vibration signal V-RMS for potential wear. This measurement has the strongest influence on the compressor availability, so it is identified as the target variable.

Support variables include current amps and rotational speed of the motor, system pressure and system temperature.

An air compressor has callouts for variables used by the Supportive (or Verticle) Smart Limit Watcher method. The highest (or target) variable is vibration in units of V-RMS. Other support variables that influence machine health, in order of influence is motor current in amps, motor rotation in rpm, temperature in degrees and pressure in psi or bar. — Supportive (vertical) method representation

Potential component solutions
Item	Sensor	Variable	Measurement description
1	VVB vibration sensors	Target	Vibration is the primary measurement for mechanical condition. Excessive vibration indicates device construction faults Consider excessive vibration as energy lost by the compressor when performing its work.
2	Motor current sensors	Support	Motor current indicates total energy used to power the work produced by the compressor. Close connection to the level of vibration as lost energy
3	DI compact speed sensors	Support	Rotational speed is a significant indicator of motion and directly impacts changes in vibration. Tracking speed can account for changes in load
4	TS RTDs and TR tempeature sensors	Support	Temperature is linked as another form of energy loss. Indicates increased work performed by the machine
5	PN pressure sensors	Support	Pressure is the work output of the compressor. Indicates how well the compressor converts energy into work

Multiple conveyor drive motor condition monitoring
Method: Comparative (Horizontal) measurements

Applying the Smart Limit Watcher to a conveyor with multiple drive systems compares how each drive operates in relation to the others. The vibration measurement V-RMS compares the mechanical integrity of the rotating components and the quality of their connection to the complete conveyor system. Any drive system that performs differently than the others is identified.

In this implementation each of the variables have an equal weight to their relevance.

A single belt conveyoe with multiple drive motors has callouts for variables used by the Comparative (or horiztonal) Smart Limit Watcher method. The variables all have equal weight and are vibration measurements in units of V-RMS.otation in rpm, temperature in degrees and pressure in psi or bar. — Comparative (horizontal) method representation

Potential component solutions
Sensor	Variable	Measurement description
VVB vibration sensors	Target and suport	Vibration is used as the primary measurement for mechanical condition. Excessive vibration indicates device construction faults Excessive vibration indicates mechanical issues