Control System rebuild
This document has been edited to remove any confidential information
To appraise the existing UV reporting system at an undisclosed wastewater treatment plant.
Recommend actions necessary to provide a simple reliable reporting system.
Shortcomings of the current system
We have received a great deal of anecdotal evidence as to the failings of this system and it is difficult to sort the wheat from the chaff.
We have had all sorts of failure causes suggested, the favourite being that the ‘black boxes’ get too hot, but we found the temperature of these units reasonable.
There is no doubt however that the system is unreliable (and may have always been so) and even if we had access to software source listings and circuit designs, correcting the faults may not be economically possible. There is also a risk that this approach could be a ‘dead end’ technically.
We also know that serial comms systems can be noise sensitive, influenced by cable length and type, and can get ‘out of step’ (synchronisation). This may well explain the most often reported fault i.e. If a module is “polled” after being taken offline for cleaning, the module fails to communicate and drops offline resulting in erroneous information to the SCADA.
We can see no effective way to correct or improve the current system; even the people who built it have never been able to make it reliable.
We could waste a great deal of time and effort on an abortive attempt to come to terms with this system.
Outline of Current System.
It is in our opinion unnecessarily complex.
The current system is a mixture of RS 422, RS 232 and MODBUS serial software at both ends of the comms chain
The’ black boxes’ mentioned above have two functions – the first is to convert the terminals from RS 422 to RS232 and the second, its main function, is as a data selector which selects each UV module in turn for polling. This is simple enough but the design is ‘1980’s, using many chips, some in sockets, which can be sensitive to corrosion.
These boxes are then linked to the ‘Modicon’ bridge multiplexer units installed in the Tubex enclosure. The Modicon units convert the serial data again, this time to MODBUS format, thence to the PLC. It is not clear why these units are used inaddition to the ‘black boxes’- it looks like these two parts of the system were designed in isolation.
The data is collected from the UV modules via two types of PCB cards:-
The first card, known as the “LCA card”
This is a distribution card, which splits the incoming 240-volt power supply into the eight connections to the ballasts, and provides the feedback on whether or not the lamps are lit. It is also fitted with a microcontroller and serial port to communicate the lamp status to the DCA card, and to control a relay for the flow-pacing programme.
The flow-pacing is never used.
There are five of these cards in each module.
The second card known as the “DCA card ”
This is also fitted with a microcontroller and A-D converter used to measure and record information such as: – lamp hours, lamp cycles and enclosure temperature. The measured UV intensity level is also routed through this card, which again uses serial comms under microprocessor control
This is a very complex card with unknown software, and reverse-engineering costs would be prohibitive, and the likelihood is that it would not prove successful.
A radical approach is required.
We propose abandoning the system in its entirety from the point where PLCs send the data to the SCADA, and build a new system from the UV modules up to that point.
(See attached drawings.)
We have a conceptual scheme although it needs a lot more work before we could start to design the cards we need.
It will, of necessity, be a major scheme.
There are 56 modules containing a total of 2240 lamps in four channels to monitor.
This means we will require 56 new design ‘DCA’ cards 280 new design ‘LCA’ cards as well as two new modules for the existing PLC (Telemechanique), which may be retained, but will need new software.
We would prefer to use Modicom PLCs as these are the Company’s standard.
Its output will still link into the SCADA system via the LAN.
We will also have to replace the wiring and connectors from the modules to the control cabinets, as there are insufficient ways in the current cables for our system to operate.
The new ‘DCA’ cards will eliminate the need for a lot of external equipment. The ‘black boxes’ (RS422/RS232 multiplexers), MODBUS conversion and associated wiring will not be required for the proposed system. This will permit a very simple I/O and 4-20mA signals to be used.
Our proposed system will monitor the following parameters:-
Module temperature, (necessary because of the wet cooling system)
UV lamp status, simple on or off.
Integrity tests for power supply, distribution etc.
To access to the UV plant data for maintenance, the local operator interface (LOI) would be replaced with a touch screen panel PC with access to the SCADA data via the works LAN.
The only data not directly available will be:- Lamp hours and lamp cycles, but these could be added to the SCADA program if required.
The only function not proposed is control of the relay for flow pacing.
This design has some drawbacks, these are:-
We cannot do the installation piecemeal so there will be downtime when reporting will not be available.
It will require new software for the PLCs (although similar software is in use at St Agnes, and may possibly be adapted)
We will have to install new cables and connectors between the modules and the control cabinets. (The cabinets will need to be sealed and air-conditioned, 4 number)
Cost. It will be a major project and will take many weeks to complete.
The advantages are:-
It is simple, understandable, and will work.
It will provide the essential data required.
It eliminates the unreliable elements in the current system.
It is much simpler electronically, giving more reliability and less maintenance.
Better noise immunity – the simple bus system has no critical timing requirements or practical cable length limitations.
The data will be more secure.
Missing “Offline modules” will not cause a problem.
Spare parts will be readily available within the UK at reasonable cost.
We have had only a very short period to investigate this project.
Whilst we are confident that we have correctly identified the only way forward, we do need to do more research to ensure there are no further services “buried “in the existing system which we have missed.
We also need to ensure that our changes will not prejudice any of the SCADA’s other functions.
We also need to speak to the people who will programme the PLCs for us, we must be certain that it can address the proposed data bus system, this is very different from the current serial communications and needs to be understood by the programmer.
Assuming this work is successful we can then prepare the necessary designs, this in turn will enable us to cost the project more accurately.