Process Engg > Tools > PREDICT

An indispensable, user-friendly PROCESS EVALUATION AND DIAGNOSTICS TOOL for Minerals and Materials Processing Industries

PREDICT is a software tool developed by Tata Research Development and Design Centre (TRDDC), to assist plant engineers in monitoring and optimizing the plant performance. With the help of a powerful material balancing algorithm, it uses the data collected from the plant (in the form of assays and flow rates) to compute a complete set of data that is balanced, consistent and reliable. PREDICT calculates plant performance in terms of recoveries and distribution of the various components in the output streams in a complex circuit configuration. PREDICT is a user-friendly system that has a powerful graphics environment, menu-driven operations and on-line context sensitive help. PREDICT maintains a database of assays, flow rates and plant performance that enables studies of trends and correlations. It also comprises of a sieve module for analysis of sizewise data. The effect of grind size as well as the plant performance as a function of particle size can be studied with the help of PREDICT.

The collection of a data set (assays and flow rates of all the streams) from the plant is the primary step for analysis, evaluation, simulation, control and optimization of the plant performance. Invariably, due to unavoidable errors in measurements, inaccessibility of some sampling locations and difficulties involved in the flow rate measurements, the data collected is inconsistent, incomplete and inadequate for complete and meaningful evaluation of the plant performance. In addition, errors in the measurements, unless reconciled, introduce large uncertainty in the estimation of plant performance and any subsequent analysis. It is, therefore, imperative to systematically design a sampling campaign (decide sampling locations for collecting adequate data) and reconcile the measured data to remove inconsistencies through an efficient material balancing technique. PREDICT provides an easy-to-use sampling design and process evaluation tool that quantitatively assesses the plant performance node by node using a sophisticated material balancing algorithm.

PREDICT caters to a need that has long been felt by the minerals and materials processing industries.

Point and Click User Interface

As PREDICT is Windows based, its user interface is a joy to work with. Pop down menus, dialog boxes, a full fledged graphics editor and a moveable toolbar make it simple and intuitive to use. Aesthetically designed Dialog Boxes permits one to enter plant data with ease. No operation takes more than 4 mouse clicks. Comparison of measured and calculated data with component recovery is available at a click of the mouse button.


Design of Sampling Campaign

PREDICT can be used to systematically design sampling campaigns and thus ensure complete plant information through an appropriate choice of sampling locations. The plant engineer can check whether sufficient sampling locations have been identified in order to collect enough data for analysis.

As a Material Balancing Tool

An important application of PREDICT is as a Material Balancing Tool in Mineral Processing industries for data reconciliation. It takes measured data from the plant as inputs and computes a set of data that satisfies the Law of Conservation of Mass for all constituents and at all locations in the plant. Unlike the conventional mass balancing algorithms used in the industry, PREDICT uses a unique two-stage Mass Balancing technique. First, mass balancing is done for individual process units, and then for the overall circuit, maintaining the performance integrity of the individual process units. The advantage of this approach is that it provides more realistic and accurate values since the error in the measured value of a particular stream does not spread to other sampling steams.

Quantitative Assessment of Plant Performance

Apart from mass balanced data, PREDICT also provides quantitative estimates or indices of the overall plant performance as well as of the individual process units. Quantitative estimation of each process unit assists the plant engineer to identify the problem areas in the plant where more attention is required.


Size-wise Analysis

Predict has a powerful Sieve Analysis module that enables the user to carry out detailed size-wise analysis of the plant. This information is useful to plant engineers to identify the likely causes of losses in recovery and grade. For example, the size fraction responsible for losses of valuables or recovery of gangue in the concentrate and the effect of grind size on the plant performance can be studied.


PREDICT and Plant Efficiency

PREDICT can store and retrieve the plant performance profile from data collected over a period of time, as and when required, in the form of trend plots and histograms. PREDICT thus provides easy means to prepare reports regarding the efficiency, throughput, recovery and the grades achieved, the quantitative assessments of plant performance, as well as the efficiency of each unit, as and when desired by the management. This kind of data is useful to the plant engineers for assessing the stability and sensitivity of the various process units in the plant to the variations, if any, occurring in the plant feed.

Printing of Reports (Tabular and Graphical)

Reports can be printed on any printer supported by Windows. Plant flow sheets with or without data, overall and detailed performance of the plant in graphical as well as tabular form can be printed as and when required.

Features of PREDICT


  • Systematic plant sampling campaign design

  • 2 stage algorithm to localize errors and reconcile data with minimum deviation

  • Overall analysis for routine sampling campaigns and detailed analysis to explore problematic locations

  • Plant/Subplant analysis

  • Quantitative estimates of the performance of all process units

  • Powerful sieve module for size-wise analysis of plant performance

  • Storage and retrieval of plant data over a period of time for analyzing trends


  • Flowsheet editor to represent plant circuit

  • Menu driven operation puts all options a click away

  • Dialog boxes for data entry

  • Process unit Toolbar for drawing plant flowsheets

  • Zoom modes for magnifying portions of plant


  • Context sensitive field level help at all times

  • Concise help through status bar

  • On-line tutorial for first timers

  • Complete users manual and quick reference guide


  • Trends/correlation in the form of plots/histograms in order to assess stability and sensitivity of the plant

  • Concise and/or detailed reports in tabular form

  • System Requirements for PREDICT


Pentium 16 MB RAM

  • 5 MB minimum disk space required

  • MS-Windows compatible mouse

  • SVGA Color Monitor.

  • HP-Laserjet or Dot Matrix 80 or 132 column (optional)


  • MS-DOS Operating System (Version 4.0 or above)

  • MS-Windows 95