Geheyan dam is located at the downstream of the Qingjiang river, in Changyang county of Hubei Province, China. The Qingjiang river is the first major tributary of the Yangtze river after the Three Gorges, as shown in Figure 1. The main function of the project is power generation incorporated with comprehensive benefits such as flood control, navigation and tourism. The power plant has an installed capacity of 1,200MW. The reservoir has a total storage capacity of 3.4Bm3, available storage of 2.2Bm3 and flood control capacity of 0.87Bm3. Other important technical parameters of the reservoir are given in Table I.

The flood from Qingjiang river will largely affect the water level of the Yangtze river, especially that at the Jingjiang Section. Geheyan dam plays an important role in regulating the flood peak of the Yangtze river during the summer and helps to avoid or alleviate flood disasters downstream of the Qingjiang river and Changyang County. In 1998, Geheyan Project played an important role in decreasing the water level of the Yangtze river, and protected the lives and properties of people living downstream [Chen & Shun, 1998].

There are two bottom outlets, four deep outlets and seven surface spillways . In order to comply with the intensified flood control tasks, there are 11 flood gates (four for deep outlets and seven for surface spillways) which have to be controlled and operated with high reliability and availability. A flood gate integrated automation system (FGIAS) has been developed and put into service since 2002.

Conventional control systems

Most flood gates are controlled by conventional computer-based SCADA systems, with considerably improved performance and reliability. The schematic structure of the original flood gates control system is shown in Figure 1. The water gate of the surface spillway is driven by a dual cylinder, whereas the water gate of the deep outlet is driven by a single cylinder. Though the automatic control of water gates has been implemented, the automation level of diagnosis and maintenance decision-making is far below than that of control.

In order to ensure the flood gates act accurately on demand, the maintenance of flood gate mechanical and hydraulic systems is necessary. Unfortunately, the maintenance costs can be very high. Concerning the maintenance strategy, intervals are defined as fixed and independent of the condition of the mechanical and hydraulic systems and repairs are done when accidents occur. Because of this strategy, the hydraulic cylinder of the surface spillway No 2 was damaged in 1998, resulting in great economic losses. The mechanical and hydraulic systems are the most maintenance-intensive in the flood gate control system and better strategies are required to ensure the flood gates be operated with high reliability and availability.

Some novel maintenance strategies have been developed and put into use in industrial production processes, such as condition-based maintenance (Li et al., 2002; Guo et al., 2002) and optimal maintenance (Guo et al., vol 17). To date, it has been proven that condition-based maintenance can effectively improve the equipment reliability and availability, leading to better economic results. It is necessary to decide the extent and time of maintenance depending on the actual condition of the mechanical and hydraulic systems. Moreover, in order to maximise profits, it is necessary to combine the control domain with the maintenance domain by the technical management domain (Li & Ye, 1998). With the rapid development of control strategies, reliability techniques, information technology and computer power, all this is now feasible.

Along with the rapid advances in computer science and engineering, some novel maintenance strategies have been developed and put into use in industrial production process, such as condition-based maintenance [Li et al., 2002; Guo et al., 2002] and optimal maintenance [Guo et al., vol 17]. To date, it has been proven that condition-based maintenance can effectively improve the equipment reliability and availability so that better economic results can be achieved. It is necessary to decide the extent and the time of maintenance depending on the actual condition of the mechanical and hydraulic system of flood gates. Moreover, in order to maximise the enterprise profits, it is necessary to combine the control domain with the maintenance domain by the technical management domain [Li & Ye, 1998]. With the rapid development of control strategies, reliability techniques, information technology and computer power, all these can be feasible.

The conventional computer-based control system is only protected in real time by devices such as pressure relays, and as a result, reliability is not very high. And some important field parameters such as oil pressure can only be monitored in the field and cannot be stored in a database for further analysis. Without this information it is impossible to analyse the performance degradation of devices and equipment. Many units use electronic data processing for planning and controlling the operation and maintenance of their equipment, which is very convenient. Using data from existing computer systems reduces the expenditure on manual data acquisition.

FGIAS is designed to meet the increasing requirements in reliability and availability of flood gates and to reduce the maintenance costs. It has been developed in an innovative control, maintenance and technical management system (CMMS), which integrates the information from control, maintenance and technical management domains to maximise the enterprise profits.

In order to explore the possible integration of domains, several European projects were launched in the 1980s (Leger et al., 1997; 1999). And for generating units of hydro power plants, computerised maintenance in a CMMS framework has already been partially implemented (Li et al., 2002; Li & Guo, 1999).

All the activities can be divided into three categories: control, maintenance and technical management. It is also clear that it is necessary to combine the information from these three domains. The underlying idea of CMMS is to integrate the three domains from functional, engineering and technological points of view taking into account the communication, data, and information processing levels. The reference modeling of CMMS is given in Leger et al., 1999.

FGIAS system configuration

FGIAS consists of two workstations, nine programmable SIMATIC S7-300 logic controllers and peripheral devices; the structure is shown in Figure 2. Each flood gate of the surface spillways is controlled by one programmable logic controller (PLC), and two flood gates of the deep outlets are controlled by one PLC. Each of the PLCs have an operation panel to display information in the field. Simatic Profibus S7 is used as the communication network between all the PLCs and workstations. Each of the workstations have a communication card installed for data acquisition from field equipment. One of the workstations works as a control station, responsible for monitoring the operation state of the flood gates and giving control commands according to the flow requirement from the water dispatch centre or the plant control centre. The other workstation is developed as the maintenance and technical management workstation for the flood gates. The two can also exchange information through a local area network (LAN).

Several analogue signals of oil pressure and opening and some discrete signals are stored in real time and a history database in the workstations through Profibus. All the information in the database can be used to diagnose faults or analyse the performance degradation of equipment combined with the information from the knowledge bank and the expert system in the maintenance and technical management station. This data is also valuable in determining effective control strategies. Through LAN, the necessary information in the database of the maintenance and technical management station can be requested and displayed in the central control room of the power station and the water dispatch centre.

Hardware configuration

Besides the relevant technical factors that affect the selection of hardware, the economic aspects also have to be taken into consideration. To achieve high reliability, environmental influences also have to be considered when choosing the hardware components. Finally, restrictions on constructing or choosing the system?s components are imposed by other safety and reliability requirements. SIMATIC has always been the trendsetter for programmable logic controllers and new developments, and is famous for its high reliability and optimal price-performance ratio. So the SIMATIC S7-300 is chosen as the programmable logic controller in FGIAS to get high reliability.

Each PLC contains the power source, CPU and communication processor modules. The PLCs have 72 input and 16 output channels. Most of these channels are used to acquire data from field equipment to monitor the operation state and the equipment condition. Some of the channels are also reserved for possible future use. Each controller is also installed with an operation panel (OP, TD17) to display the operation state and release the fault alarm in the field.

In order to reduce the costs, the single-circle absolute encoders and the comparative encoders are used as transmitters to measure the opening of the flood gates instead of the multi-circle absolute encoders. The redundancy between the single-circle absolute encoders and the comparative encoders is used to improve the reliability of the flood gates and to avoid destroying the hydraulic cylinders. Consequently, better economic effects and high reliability are both achieved.

Software configuration

The control station, developed with Intouch of Wonderware Corporation, can monitor the operation state of the 11 flood gates working as a normal SCADA system. Its responsibilities include setting different kinds of operation type of the flood gates, setting

opening to reach, supervising the operation state, fault alarming and storing data in databases for further analysis.

The other workstation is responsible for maintenance and technical management. It also works as a redundancy control station to enhance the reliability if the control station fails. The maintenance and technical management workstation can communicate with the control station through LAN or Profibus network. If the control station failed, the maintenance and technical management station would work with the responsibility of the control station automatically. The control station can also work with the functions of the maintenance and technical

management station.

Functions of the system

FGIAS has three fundamental function domains: the control domain, the maintenance domain, and the technical management domain. These three are integrated and can communicate with each other and exchange necessary information. The process model of the FGIAS is given in Figure 4 in Integration Definition for Function Modeling (IDEF0) notation, the International Alliance for Interoperability (IAI) preferred convention for the creation of graphical process models for Industry Foundation Classes (IFC) specification projects.

3D and multimedia techniques are also employed to provide a powerful interactive environment between human and computers, which also provides a learning environment that enables operators to attain deep understanding and improve their own knowledge. It is especially useful to newcomers [Chen & Li, 2001].

Control domain

The primary task of the control domain is to optimise the control

strategies of the flood gates according to the requirement from the water dispatch centre and/or the power station combining information from the maintenance and the technical management domains. During flood gate operation, the control domain monitors the operation state of the flood gates and sends real time data to the maintenance and technical management domains. The process model and information flow of the control domain is given in Figure 4 below.


Operation type setting – for remote control, all of the flood gates can work in three operation types according to the given water flow: single operation, group operation and all operation. Among these operation types, group operation is used most frequently. Thus No 1 and 7, 2 and 6, 3 and 5, 4 water gates of the surface spillway are grouped respectively in consideration of symmetrical distribution.


Flow distribution determination – in order to satisfy the sand-washing requirements, the flow distribution model must be used to determine the operation and opening of water gates. It is also necessary to consider the water forecast, power generation requirements and water level.


Control mode choice – each of the flood gates can work in test mode, manual control, automatic control and remote control. These modes are suitable for different operation requirements and different system states respectively.


Operation state of the flood gates monitoring – during flood gate operation, important hydraulic pressure parameters and alarm signals are monitored, recorded and analysed in the online mode. All kinds of parameters of the flood gates operation are displayed with the 3D graphic methods and are user-friendly. If failures or disturbances are detected, they are indicated and stored in the real time database

and then in the history database. The maintenance and technical

management domains can share these data for further analysis.


Real time simulation – the 3D and visual technique has been applied to the modelling of the operation state of the flood gates. The action of the simulation model is driven by the real time data acquired from equipments in the field.


Mechanical and hydraulic system monitoring – the mechanical and hydraulic systems are also modelled by means of the 3D technique. Their driven data are also acquired from real system at real time, and their important parameters are monitored and stored.

Maintenance domain

By analysing the data from the field and the database acquired by

the control domain, the maintenance domain can diagnose faults and help identify degradations in equipment performance. During this process, information from the technical management domain is also considered to facilitate these analyses. The detailed process model

and information flow of the maintenance domain is given in Figure 5 below.


Failure detection and fault diagnosis – according to the alarm signals and historical data stored by the control domain, and the knowledge in the expert system of the technical management domain, the workstation can accurately diagnose failures and indicate necessary action. The result of a diagnosis is the requirement for maintenance work to be done, either a repair or a further detailed inspection. The planned date and duration of maintenance are sent to the technical management domain, water dispatch centre and power plant.


Equipment degradation analysis – to monitor equipment behaviour and changes in its performance, combined with the dynamic and static data in the database, the system determines the maintenance actions to be taken. All these actions aim at moving the system maintenance strategies from breakdown maintenance to preventive maintenance, then ultimately optimal maintenance.


Graphical support for maintenance jobs – carrying out the maintenance job, the worker can retrieve the information from FGIAS about the job plans or the technical data about the special equipment displayed in the 3D graphic. With regard to the typical maintenance actions, FGIAS can give the 3D and visual indication and advise the maintenance actions to be taken.


Preventive maintenance task implementation – when setting up FGIAS, all targeted equipment needs to be identified for preventive maintenance. By using the manufacturer’s guidelines, specific tasks can be generated on either a time basis or on hours of operation. Weekly, monthly, quarterly, semiannual or annual tasks will be

generated so equipment can be properly maintained. Previously, machinery was very often forgotten until it failed. FGIAS will prompt that this equipment is ready for routine maintenance.

Technical management domain

The technical domain is used to coordinate the other two to get better economic results. The following data of the technical management domain are processed.


Water flow analysis – according to the water flow from the water dispatch centre, the offline flow analysis is used to determine the control strategies, operation type and opening to set in the control domain. The online water flow analysis is used to calculate and monitor the real time water flow.


Equipment information management – the technical documentations of the equipment and its components are digitalised and stored in the database and can be retrieved. This includes visual, manufacturing, installation, operation, maintenance, equipment condition and technical parameter information


Work order management – with regard to the repair and maintenance action, FGIAS offers functions for the generation and administration of work orders. Via FGIAS, the approved work orders are generated and then indicated to the workers.


Operation and fault reports – FGIAS can generate operation and fault reports in any specified period according to requirements. The

statistics data about the equipment can also be queried from the database, which is very useful in determining the maintenance actions and helpful to analyse the equipment performance. These automated reports can be analysed to increase efficiency and reliability.


Expert system and knowledge bank – expert operators execute their operations unconsciously, but they choose the most appropriate knowledge according to the situation. Such maintenance expertise of operators is summed up and stored in the knowledge bank. Furthermore, the expert system provides effective support for the maintenance worker in finding equipment faults.

The implemented knowledge concerning specific problems of the mechanical and hydraulic system, e.g. the fault of system over

pressure, is available to all workers. The knowledge in the

expert system can be added or updated according to the

requirements.

Field test results

Field tests of FGIAS have been performed three times in the Geheyan hydro power plant, and the system was successfully put into service in 2002. The regular function and performance and fault detection of the FGIAS were comprehensively tested with satisfactory results. In order to prevent water flowing away unexpectedly, the other three control modes except no-load test have to be tested after the maintenance gate has been in position or the water level upstream is less than 175m. There are so many items to be tested for each of the control modes that cannot be listed below. Therefore only the primary test details and their purposes are presented in the following.

No-load tests

In order to protect the equipment, the control system must first be tested with no load. Another purpose is to test if the control sequence of the programme is correct. Before the no-load tests, efforts must be made to check the circuitry of transmitters and

actuator channels to avoid destroying components, especially the modules of the PLC. Before further field tests, it must be ensured that the programme can enable the mechanical and hydraulic components to act accurately by the no-load tests.

Manual control tests

When the flood gates are operating in this mode, the faults resulting from encoders are left alone and the opening of the flood gates has to be checked by eye. This kind of control mode is primarily used for emergency and is seldom taken normally. One of the most important test items of the manual control is to see if the down-limit relays have been installed correctly and if they can act accurately.

This test item can be ignored in the automatic control and the remote control for it has been tested here. Another test is that oil motors stop running and the flood gate comes down by its weight when the opening is less than 20cm, or else the oil cylinders may be damaged.

The flood gate must retain its original opening to pass the given flow when the motor stops running. Therefore, the system must have the ability to raise the flood gate automatically if the flood gate has come down by its weight and the opening has been decreased more than 10cm. In the field this was tested until a satisfactory performance was achieved. This has to be tested further in automatic control and remote control modes because these three control modes use different control functions.

Automatic control tests

This control mode can be used when the communication network cannot work and the PLCs cannot receive commands from the control workstation. There are many items to be tested in this kind of control mode, the most important of which is to test if the system has the ability to correct the opening difference between the right side and the left side of the flood gate and to assess its performance.

Another item is to test the up-limit relays and if the motors stop running when the opening is more than 1700cm. Before this test item has been carried out, the full-journey of the flood gate cannot be performed.

Because the down limit relays have been tested in the manual control, it can be ignored here. But it has to be tested again that the motors stop running when the opening is less than 20cm.

Remote control tests

This is the most common control mode. Table II and Figure 7 give one of the remote test results of the flood gate on the 3# surface spillways. The data in Table II and Figure 4 demonstrate that the opening difference between the right and the left is very little and the flood gates are operated perfectly. The test results of other flood gates are also satisfactory. In all of the tests, the difference between the right and the left is much less than the limits. This results from the better control strategies and the safety and reliability of the system is improved.

Fault detection and diagnosis

In all control modes, the system’s ability to detect different faults on-line has also been tested. When man-made failures have occurred, it has been demonstrated that the control station can give an alarm correctly and that the maintenance and technical management station perform fault detection and give the proper maintenance advice and action to be taken.

Some other functions of the FGIAS have also been tested off-line, such as report generation, work order generation and system simulation.

Conclusions and prospects

FGIAS has been developed in the CMMS scheme for the flood gates of the dam of the Geheyan hydro power plant, integrating control and maintenance with technical management to improve the system’s economic performance. The new system makes important contributions to further enhance the availability and reliability of the flood gates. The test results and successful applications demonstrate the significance of FGIAS. The application of CMMS to the entire hydro power plant is also feasible with today’s developments in computer science and engineering.


Tables

Table 1
Table 2