Autonomous/Semi-Autonomous Monitoring of Industrial Plants

 

Introduction

Heavy industrial plants operate in many sectors and industries worldwide. The manufacture of many materials, production of energy, gathering and processing of natural resources and chemical production involve potentially hazardous processes.

There have been efforts to improve the safety of workers. In many industries, there are regular checks to reduce the risk of exposure to hazardous materials. Workers have to wear personal protection equipment (PPE) to guard against:

  • hazardous materials
  • noise
  • brightness levels
  • extremes of temperature
  • radiation

Even with PPE, workers may be limited in the amount of time they can spend in the work environment. This can be as low as 20 minutes. Plant shutdown for inspections and high worker turnover can lead to significant financial losses in lost revenue.

It is rarely cost-effective or even possible to fit out existing plants with remote monitoring systems for all the checks that have to be made on a regular basis, such as checks for cracking in pressure vessels, wear on components in contact, misalignments in bearings and pipe leaks.

Task

Your task is to design a system to inspect an industrial plant of your choosing, in consultation with your industrial mentor or staff member. The system should be as automated as possible (at least semi-autonomous) to replace humans when carrying out risky procedures. This also has the advantage of reducing or eliminating the requirement for a major element of a plant to shut down for inspection reasons.

At the end of the week you will pitch your chosen solution to investors, and so you should aim to design a system which will be highly desirable for a large target market of potential customers. As such you might aim to produce a flexible system which can perform a range of inspections under different conditions. It is up to you to select your own industrial plant(s) and inspection(s), and then design a monitoring system to perform the inspections.

Considerations

First of all you should identify the problem you will attempt to solve. What is the inspection you wish to perform, and in what plant? What do you wish your system to measure, and how will it do so?

How will the data be collected? You could consider perhaps a system of simple, remote sensors which are placed and communicate wirelessly back to a central computer, or perhaps some more complex sensing equipment which can be moved around the plant by a robotic drone. How would such systems be deployed into a plant and operated?

Are there any constraints on the design due to the environment in which the system will operate? If the inspection can be performed without shutting down the process this will give a significant advantage to the operator, but may mean that your device will be exposed to a hazardous or inhospitable environment. This environment may also be in strong weather conditions, for example in the case of external inspections of oceanic oil rigs.

How will the collected data be transmitted back to a central computer? You will potentially be collecting a lot of plant information, and an autonomous system that can operate without requiring a plant shutdown will likely mean that inspections will be replaced semi-continuous monitoring. How will all the data coming in be compiled and analysed in real time?