System Identification

Upload your time-series response data to instantly classify linearity, estimate transfer functions, and quantify prediction certainty.

Cloud upload Logo

Drop your file here or Click to upload

Uploaded Files

No files uploaded yet.

How It Works

A simple, three-step process to get your files analyzed.

Upload

Drag and drop your .csv files securely into our upload zone. A minimum of 50 rows is required for better resuts.

AI Analysis

Our custom database and AI models validate your schema, classify linearity, and estimate the most suitable system model.

Results

Receive a detailed report in a matter of seconds.

Required CSV Column Format

Your file must include these exact columns:

time,input_signal,output_signal

Model Output Overview

This section explains exactly what is identified and what each reported parameter means.

Classification and Trust

  • linearity_classification: whether your system is linear, nonlinear, or uncertain.
  • prediction_certainty: confidence score from 0 to 1 for the classification.
  • identified_order: selected linear transfer-function order (1 to 4).
  • R^2, NRMSE, NMAE: fit-quality metrics used to validate how well the identified model matches measured data.

Linear Parameter Meanings

  • tau: time constant, how quickly a first-order system responds.
  • K: system gain, output change per unit input.
  • m: effective mass/inertia term.
  • b: damping term.
  • c: stiffness/restoring term.
  • a3, a2, a1, a0: higher-order denominator coefficients that shape poles, resonances, and transient behavior.
  • confidence_intervals: 95% uncertainty range for each linear parameter estimate.

Nonlinear Parameter Meanings

  • b(x,x'): nonlinear damping law as a function of state and state rate.
  • k(t): time-varying gain term when system gain changes over time.
  • nonlinear_restoring(x): nonlinear stiffness/restoring effect, often cubic.
  • saturation(u): estimated input clipping/saturation limits.
  • harmonic_distortion: amount of nonlinear harmonic content in the output spectrum.
  • fundamental_freq(Hz): dominant base frequency of the response.
  • superharmonics: detected energy at higher harmonics (2f, 3f, ...).

Frequently Asked Questions

What file types are supported?

Currently, we exclusively support .csv files. The file should contain time, input_signal, and output_signal columns.

Is my data secure?

Yes, all uploads are processed locally and temporarily stored strictly for analysis. Files are deleted immediately after the system report is generated.

How long does the analysis take?

Linear system analysis typically completes within seconds. Highly complex or nonlinear systems may take slightly longer due to advanced Monte Carlo Dropout computations.