Vessel Stowage planning data Validation and Imputation

Introduction

The Vessel Manual Plan (Stowage Plan) is specially designed to stow the cargo onboard in the container vessel with better space utilisation, stability, and stress calculations. It means all essential factors which are required for effective stowage planning have to be validated. Hydrostatics and stability are one of the most important areas of focus in ship design and operation, not only to ensure the safety of the ship, cargo, crew and passengers but also to enable proper conditions for completion of all the processes on a ship. Hydrostatic tables of a ship are used to obtain the values of essential parameters required to calculate the ship’s stability for safe navigation. The business objectives and benefits expected after automation of the selected business process are:

• Reduce processing time of validation of the hydrostatic tables
• Accuracy of the process output by doing away with possibilities of human errors

Manual Process Overview

Hydrostatic input files for vessel stowage planning are received in PDF format. These PDF files are processed using in-house OCR. OCR output files are not in format for processing applications. Manual Editing and correction involve; Removing special characters, empty rows, repeated headers, wrongly interpreted. Once the data is cleaned up data will be manually corrected compared with the original hydrostatic tables available in PDFs. These output files are then used to plot graphs to check the spikes to check for any incorrect data. After the above manual corrections, the graph will be smooth for the correct data. Automated RPA Process

• Input Vessels in PDF format are first processed using an OCR.
• The output of OCR extracts are stored in folders in Excel format
• Each of the files in the input folders will go through Data correction, validation and Prediction.

Data correction, validation and prediction happens in several steps :

Data Formatting:

Converting the raw input data to General Format for the entire excel sheet.

Spikes Identification:

An incorrect data will result in spikes on the plotted graph. RPA process will detect those data cells which may be causing these spikes in the graph. All such cells will be marked and replaced with dummy placeholder data temporarily. These files are termed impute ready.

Interpolation and OCR error correction:

• The files in the Impute ready folder are picked for Interpolation.
• RPA process will invoke the Python Script for interpolation.

An output file is created imputed with values using polynomial interpolation techniques.

Example: Interpolated Data