How to Dynamically Adjust Initial Thresholds to Adapt to Process Changes

Binding to process status and real-time baseline updates is the foundation of dynamically modifying thresholds. This is accomplished by using a three-layer logic to adjust to various process modifications. The particular implementation is straightforward and doable:

1. First Layer: Process Status-Based Automatic Threshold Switching Set up matching thresholds in advance for various process templates. When switching goods or processes, certain thresholds are automatically loaded, removing the need for human modification:

Use the basic threshold for conventional plastic parts, tighten it by 10% for precision plastic parts, and loosen it by 15% for high glass fiber/corrosive raw materials, according to the product processing difficulty category.

For low-pressure operations (<100 bar), tighten the threshold to 80% of the fundamental value; for high-pressure processes (>500 bar), loosen the barrier to 120% of the basic value;

In order to prevent false alarms brought on by adapting outdated thresholds to new processes, the system automatically reads the corresponding preset threshold when switching processes.

2. Second Layer: Statistical Data-Based Dynamically Updated Benchmark

Every week, the system automatically determines the two sensors' historical deviation distribution and modifies the threshold in response to real fluctuations:

It determines the most recent week's standard deviation (σ) of the difference between the two sensors. 99.7% of the variations will fall within ±3σ under typical steady-state settings.

The threshold has been changed to 3σ. The threshold is automatically expanded if actual fluctuations rise, and tightened if they fall.

To prevent threshold failure owing to excessive variations, the threshold contains upper and lower limits: the minimum is no less than 0.5 times the sensor's nominal accuracy, and the maximum is no more than 3 times the nominal accuracy.

For instance, the new threshold is changed to 0.24%FS if a sensor's nominal accuracy is 0.1%FS and the statistical standard deviation for the most recent week is 0.08%FS. This adjustment better reflects the operating conditions than a fixed threshold and adapts to the current real fluctuations.

3. Adaptive Threshold Relaxation for Unusual Operating Conditions in the Third Layer

The system automatically temporarily modifies the threshold in response to severe process fluctuations:

The current stage threshold is automatically loosened by 50% when the injection molding machine pressure fluctuates more than twice the typical steady-state value. Once the fluctuation has ended, the initial threshold is restored.

A deviation is identified as an operational condition disturbance if it is triggered three times in a row before automatically recovering. The threshold for that stage is then automatically decreased by 0.05%FS to minimize similar false alarms in the future.

In addition to maintaining a low false alarm rate over time, its dynamic adjustment logic can automatically respond to a variety of process changes, including product changeover, raw material changes, and pressure fluctuations, without the need for frequent manual adjustments.