Optimisation of costs and performance of a sludge destruction plant
M D Everest [1] MSc CEng MInstE. Meniscus Systems Ltd. Tel: 01480 433714. Fax: 01480 433715.
E-mail: meverest@meniscus.co.uk
S Harris [2] Severn Trent Water Ltd. Tel: 01162 340340. Fax: 01162 671887.
E-mail: Sam.harris@severntrent.co.uk
1. Abstract
To help drive down operating costs at the Coleshill Sludge Destruction Plant in Birmingham, Severn Trent Water Ltd use existing operational data to understand process interactions within the plant and to obtain a “helicopter” view of site performance.
A data management service is used that collects the routine operational data and routine analysis undertaken by the operators. This data is used to automatically derive cost and performance information for each separate process unit. The system also calculates a heat and mass balance throughout the plant to help measure and monitor key performance parameters. Using these techniques it is possible, for example, to determine increased operating costs caused by non availability of plant. Data mining techniques are used to identify “hidden” relationships in the data and in conjunction with the operators, used to derive a set of operating rules for the plant. These rules are input into an expert system that automatically applies them to the incoming data on a day to day basis.
2. Keywords
Sludge destruction, incineration, optimisation, key performance parameters, data mining, expert system, data collection, best operating practice.
3. Introduction
The sludge destruction plant (SDP) operated by Severn Trent Water Ltd at Coleshill in Birmingham, processes over 28,000 tonnes of digested sludge a year from Minworth STW and the Black Country Trunk Sewer – having a combined population equivalent for sludge treatment of 2.5 million.
The plant comprises 5 centrifuges producing thickened sludge with target dry solids content of 20%. Thickened sludge passes through rotary driers to increase the dry solids to a target of 30% before passing to two fluidised bed incinerators, rated at 2.5 tonnes of dry solids per hour each. The incinerators are supplemented with natural gas to maintain a temperature of 870 deg C in the incinerator head space. An economiser provides hot water used in the driers and an air pre-heater raises combustion air from ambient temperature to around 550 deg C. Comprehensive gas cleaning ensures that the flue gases are compliant with the operating conditions as set out by the Environment Agency.
Ash and any thickened sludge not burnt in the incinerator are sent to landfill.
The overall plant, and in particular the sludge thickening centrifuges, incinerator and flue gas handling systems are fully automated, monitored and controlled with a sophisticated Yokagawa control system. However, in order to continually drive down operating costs at the site, Severn Trent Water Ltd are investigating how day to day changes in operating conditions affect the overall cost and performance of the plant. The monitoring programme to evaluate these opportunities is two thirds complete. This paper highlights some of the interim findings of this work.
4. Macro vs micro view of data
The Yokagawa control system on the plant provides comprehensive alarm management for the plant and undertakes monitoring and control of the incinerator operating conditions on a real time basis.
This form of “micro” control of the plant is absolutely necessary for the control of complicated process plant. However, for process monitoring and to obtain overall information on costs and performance criteria a more “macro” or “helicopter” view of the data is helpful. This presents data in a more concise format and allows calculations that use data not easily collected by standard monitoring instrumentation, e.g. dry solids readings. Presentation of data in this way also facilitates the use of more complex data analysis techniques such as data mining. Traditionally, process or plant managers have undertaken such analysis using spreadsheets or simple databases.
By adopting this “macro” view of cost and performance, the plant operator has a much better understanding of the individual costs of each process within the SDP and of the total overall operating cost of the site. This in turn helps to justify modifications and improvements to the plant so as to reduce operating costs.
5. Data collection and monitoring
A total of 19 parameters are collected from site on a daily basis with a further 23 on a weekly basis. Daily readings comprise routine operational data including sludge, polymer, water and centrate flow readings as well as inlet and outlet dry solids readings for the centrifuge and the dry solids of sludge leaving the driers.
Weekly readings comprise key hours run data and consumption figures for gas, electricity, water and caustic.
Data is entered into a Psion Workabout handheld computer by the night shift operations team and the data is downloaded to a central computer during the night. The central computer derives a broad range of process and cost based information and uploads a new data file and changes to the reading forms on the Psion automatically.
Key parameters are assigned high/low limits which generate daily exception reports for the operators. These can be uploaded to the Psion if required.
All the data sent to the Psion and all the derived parameters can be viewed over the internet and customised reports summarise key performance information and the daily operating costs for the plant.
The monitoring regime entailed no additional monitoring or data collection for the operations team since all the data was collected prior to this monitoring exercise and recorded on traditional paper log sheets.
6. Heat and mass balance
A mass balance derived around the whole of the plant, enables the calculation of key performance information, such as mass of dry solids burnt in the incinerator. Such information is not available under the standard monitoring systems since it would require measurement of sludge flows from the centrifuge and driers at high dry solids conditions. Such flow measurement is difficult to achieve in a reliable and consistent way.
The mass balance uses inlet sludge flow and dry solids readings to derive the mass of water and solids processed at each process in the plant. Readings derived in this way are accurate to an estimated +/- 15%. The accuracy of the mass balance is dependent upon the accuracy of the centrate flow meter (only one such meter) and on the amount of thickened sludge “stock” held on site ready for disposal at the landfill. Monthly results are summarised in a one page report.
The heat balance is based around the incinerator and uses solids and water loading data derived from the mass balance as inputs to calculate heat losses in the flue gases and in the associated water vapour. These heat losses are then compared to the heat inputs into the incinerator in the way of natural gas (gross calorific value used), sludge solids load and pre-heated inlet air. The difference in the heat balance between heat into and out of the incinerator is 5%.
For a sludge with a calorific value (gross) of 16,050 kJ/kg, the incinerators require 4.7MJ of energy to evaporate each kg of water. 63% of the energy input into the incinerator is required to evaporate the water with the balance comprising heat losses in flue gases 35% and 1% each for heat losses in ash and for general heat losses (estimated) around the plant.
Stoichiometric air requirements for the sludge (analysis per 100kg of dry solids: C 34%, H 6.6%, O 13.8%, N 3.8%, S 1.3% and Cl 0.3%) are 5.7 kg of air per kg of dry solids burnt. For the combined gas and sludge figures the stoichiometric air requirements are 6.3 kg of air per kg of dry solids burnt.
7. Energy losses due to high excess air values
Control of the combustion air required in the incinerator is managed by the central Yokagawa control system on the plant. Control is based upon a key assumption that dry solids into the centrifuge is 31% at which point the sludge is assumed to be autothermic – this is confirmed with the heat balance. In practice, the solids content of the sludge into the incinerators is always less than this due to operational problems with the rotary driers. The plant presently generates sludge with a dry solids content of 22-26% against a target of 30%.
As a consequence, the control system continually overstates the combustion air requirement. Based on a carbon balance which assumes that no residual carbon is left in the ash, the total mass of air used is 12.7 kg per kg dry solids burnt which equates to an excess air figure in the incinerator of over 100%, considerably more than is required for optimal performance of the incinerator. This high excess air figure is confirmed by the high oxygen readings in the flue gases of between 12 and 15%.
Too much excess air wastes energy in two ways.
a. The combustion air blowers operate for longer than required
b. The energy required to heat the excess air to the operating temperature of the head space is wasted. Note that the economiser and the air pre-heater downstream of the incinerator will extract heat from the flue gases and so recover a proportion of this heat. However, if the demands of these two units are met with a reduced excess air requirement then any surplus excess air can be considered as wasted.
For every 10% increase in the amount of excess air used the annual savings on blower power are £7,200. If the target excess air requirement in the incinerator is 75% (8 to 10%) of oxygen in the flue gases) then the additional blower power costs are around £18,750 per annum.
The additional energy required to heat all the excess combustion air from 550 to 870 deg C is 8% of the total energy input. Based on the target excess air requirement of 75% then the wasted energy represents 21% of the natural gas burnt in the incinerator or around £25,000 per annum.
8. Cost of non operation of the sludge driers
As a consequence of operational problems associated with the sludge conveyors to the driers, the reliability of the driers is poorer than required, resulting in sludge with too low a dry solids content being fed to the incinerator.
By increasing the dry solids of sludge to the incinerator, the site reduces the volume of sludge sent to landfill. With the use of the mass balance it is calculated that a 1% increase in dry solids from the drier reduces the sludge disposal to landfill by 63 tonnes per month.
In addition, by increasing the dry solids content of the sludge the process moves towards an autothermic condition, whereby the sludge will burn without the need for supplemental fuel, the gas use per kg of dry solids required to maintain the head space temperature will reduce. At present it is difficult to quantify this saving. There is a relationship between gas use per kg of dry solids and the dry solids content of the sludge to the driers but its impact is masked by a more dominant relationship between the gas use and the blower hours.
It is feasible that cost savings identified by use of the heat and mass balances may be sufficient to justify additional capital expenditure to improve the reliability of the driers.
9. Polymer make up water
Polymer make up water contributes 11% of the total volume of sludge processed by the centrifuges. The annual additional power cost to centrifuge this volume of water increases the centrifuge power by some £10,000 per annum.
Whilst this volume is accepted as normal by experts in the industry, it effectively means that a sludge with a 3% dry solids content will be reduced down to 2.7% prior to thickening in the centrifuges. If additional polymer is used to thicken digested sludge to 3% prior to processing then this represents an additional cost.
10. Future work
Future work involves incorporating operating knowledge gained from analysis of the data and from the experience of the local operators and managers into a site specific expert system. Such a system has two objectives.
a. To apply the operating rules to the incoming data on a day by day basis so as to help identify potential operating conditions on a proactive basis. In this way the system can help establish and maintain best operating practice for the plant.
b. To develop a cost model for the entire plant that takes into consideration the impact that one process has on the subsequent process rather than viewing each process as an individual stage.
11. Conclusion
The monitoring programme has demonstrated that daily operational data, collected as part of routine operational duties, can provide a valuable insight into the overall cost and performance of a plant.
With minimal costs of acquisition of the data, this approach maximises the use of existing data and helps to capture operating knowledge about the process.
The ability to derive operating information about individual processes facilitates inter plant cost and performance benchmarking, helps to establish and maintain best operating practice and provides managers and operators with readily accessible cost and performance information.
12. References
Wastewater Engineering 3ed. Metcalf & Eddy
Kemps Engineers Yearbook. Edited by J P Quayle
Heat Transfer from Flames. W A Gray, I K Kilham and R Muller
Sewage Sludge II. Conditioning, Dewatering and Thermal Drying. Institute of Water Pollution Control
Process Plant Design. J R Backhurst, J H Harker
Thermodynamic and Transport Properties of Fluids. G F C Rogers, Y R Mayhew
13. Acknowledgements
The authors would like to thank the operators and management at Severn Trent Water, in particular those at the Coleshill Sludge Destruction Plant, for their assistance and support throughout this project.
Views presented in this paper are those of the authors and do not necessarily represent views help by either Meniscus Systems Ltd or Severn Trent Water Ltd.
[1] Meniscus Systems Ltd. Unit 3 Blotts Barn, Brooks Road, Raunds Northants. NN9 6NS. Http://www.meniscus.co.uk
[2] Severn Trent Services Ltd. Senior Technician, Service Delivery East, Wanlip STW, Leicester.