Mass Balancing

• Mass Balancing
• • JKMetAccount
  • Mass balancing overview
  • • Key concepts in mass balancing
    • Standard deviation in mass balancing
    • • SD estimation
  • Mass balance results

The mass balancing design is configured as part of project-specific site requirements. Refer to the Business Solution Design document supplied to your site for detailed information about stream and reconciliation configuration.

This section provides a generic overview of mass balancing theory and fundamentals.

JKMetAccount

Production Accounting uses the JKMetAccount statistical mass balancing engine to perform statistical mass balancing calculations. As a Production Accounting user, you do not interact directly with JKMetAccount. Statistical mass balancing data and results are viewed through the Production Accounting web application.

Note: Statistical mass balancing with JKMetAccount is only recommended when sampling practices, sample quality and lab assay results meet the requirements for the model to work effectively. Where the requirements for statistical mass balancing cannot be met, Production Accounting uses standard metallurgical formulae for mass balancing.

Mass balancing overview

Statistical mass balancing is a data reconciliation method used in metallurgical operations to ensure the accuracy and consistency of material flow and assay data. It involves adjusting raw data from various streams in a plant to make them consistent with the laws of conservation of mass. The statistical mass balance model aims to minimise the difference between measured and calculated (balanced) data.

Statistical mass balancing can help identify issues related to sampling, assaying, and other measurement inaccuracies. Insights from statistical mass balancing can be used to refine sampling practices, sample preparation and calibration of various field instruments.

Key concepts in mass balancing

• Data collection and analysis—Robust data collection is the foundation of effective statistical mass balancing. This includes assay data and flow rate data measured at steady-state conditions. Once collected, the data is statistically balanced to check for consistency and accuracy.
• Error identification—Statistical mass balancing helps identify variations or errors in data; for example, from sampling procedures or design, assaying techniques, sizing procedures, instrumentation issues, statistical effects and/or fluctuations in plant flow rates. By reconciling the data, statistical mass balancing highlights areas that need improvement.
• Model fitting—Statistical mass balancing uses mathematical modelling to 'fit' the data. The model includes simulated representations of process units like mixers, classifiers, and grinding mills. The aim is to produce best fit estimates of flow rates and a set of adjusted assay data consistent with those flow rates.

Standard deviation in mass balancing

Standard deviation (SD) is a measure of precision in statistical mass balancing. It quantifies the expected variation from the true value of a data point. High SD values indicate low accuracy, while low SD values indicate high accuracy. Accurate estimation of SD is crucial for reliable statistical mass balancing results.

SD estimation

Your site must perform an initial testing process to determine the SD values for the process units or streams in the statistical mass balancing model. These tests must be done when plant operations are at a steady state. Between five to ten complete observations (including sampling, sample preparation and analysis) are recommended to produce a usable SD estimate.

When a unit or stream's data accuracy is low, select a higher SD value. For example, for streams where representative sampling is difficult or samples are sometimes contaminated, you may need to set a higher SD value. When data accuracy is high, select a lower SD value. The statistical mass balancing model makes fewer or smaller adjustments to data points with low SD values compared to data points with high SD values.

Once SD values are determined, they must be defined as dated calculation constants in Production Accounting. See Update dated calculation constants. As conditions change; for example, sampling precision improves or instruments are re-calibrated, SD values can be adjusted to reflect the increased or decreased data accuracy.

Mass balance results

The JKMetAccount statistical mass balancing results can be configured to display in Production Accounting. Summary logsheets may be configured just for this balance data, or the data can be displayed as a section in other logsheets.

The following metrics are important for assessing the results:

• Iterations, also called Number of Steps—The number of iterations performed by the balance. High variability over time may indicate source data inconsistencies.

  Note: Comparing this value to the maximum iterations defined in the statistical mass balancing engine can help evaluate its performance.

• Balance Indicator, also called Weighted Sum of Squares (WSSQ) or Residual Error—The sum of the terms. Always a positive number. Lower values indicate a better balance result.
• Acceptance Indicator, also called Pooled Standard Error (PSE)—A measure of balance accuracy, with lower values indicating a better balance. The PSE can help you assess the suitability of the selected SD values.
• Sum of Differences, also called Error Sum—A measure of the total amount of adjustment required (in both directions) to achieve the balance. This value takes no account of SD values and is a raw measure of how close the original data was to being balanced. Lower values are better, but the PSE and WSSQ are more reliable indicators of balance performance.