There are many considerations when looking at improvement to governor systems at plants, explains David L Kornegay, senior project engineer with North American Hydro


Original governor cabinet

If a governing system upgrade or replacement is necessary, the first decision to make is how much of the existing system should be replaced. The simplest approach is to replace any parts of the system that are historically troublesome, either in performance or in maintenance requirements.

Especially important to consider are any system components that have caused unscheduled outages of the unit. Unscheduled outages have a double impact upon revenue from the unit: lost generation revenue during the outage; unscheduled outage can result in penalties imposed by the power system authority, such as fines or a lower revenue rate due to the lower reliability of the unit.

The performance of many older governing systems does not meet current system requirements of contribution to grid frequency stability, speed of generation dispatch, and time required from a stopped condition to generating power into the interconnected power system. Even if the older governing systems can be restored to like-new operating condition (which can be difficult or impossible due to the diminishing availability of spare parts), their performance may fall short of current operating requirements.

Current technology offers many advantages in using programmable digital controllers. Often, special functions needed to meet site-specific performance requirements can be implemented in software with no hardware changes. Older mechanical and analog electronic governing systems required additional hardware, along with the required integration engineering and installation, for each additional function needed. Programming changes to digital governor controllers can generally be made in much less time.

An important issue for many owners is the long-term maintainability of a governing system. Historically, many governor manufacturers have provided systems that use proprietary hardware and software platforms. This means that the owner needs to go to a single source (who controls the price and delivery) for any replacement parts or software support. It is particularly appealing to many owners if a governing system uses only commercially available components (wherever possible) and a digital controller programming platform that has easily available programming expertise.

Governor strategy – debate and development

Historically, governing theory debated the merits of conventional PID control strategy versus the US Bureau of Reclamation’s (USBR) Double Derivative governing strategy, which was developed around 1970. A recent (circa 2004) improvement to governing technology is the TEPID, or “Transient-Enhanced”, governing strategy that builds on the familiar “Proportional, Integral, Derivative” (PID) format, and was developed by North american-hydro. As a modification of USBR’s Double Derivative strategy and while functionally similar, TEPID differs in that governor gains are structured in the PID format and so making it easy for someone familiar with tuning a conventional PID governor to tune a TEPID governor.

The TEPID strategy offers greater speed stability, along with the ability to provide rapid response to generation dispatching commands without sacrificing the governor’s contribution to the frequency stability of the interconnected power system.

The concept of TEPID was developed over some 20 years during which time there was the ongoing debates between governor designers in commercial firms and USBR. During this time, I constructed an analog governor simulator, which was useful in finding proper tuning parameters for field service personnel preparing to commission a new governor installation.

During time spent running simulations of governor operation under varying operating conditions, I noticed that the governors were much more stable when they used gate position as the feedback (as opposed to generation feedback) for developing the governor’s speed droop characteristic. Later on, during my IEEE committee activities and from conversations with USBR personnel, I began to see some advantages in the federal agency’s approach but the stabilising gains were structured differently from the conventional PID governor strategy.

Around 2004, after joining North American Hydro, I investigated ways of improving the performance of hydro governors. Propelled by the information compiled in the then soon-to-be-published IEEE Standard 1207, the committee for which I chaired, I looked into re-structuring USBR’s Double Derivative governor strategy into a format that would be more intuitive to tune for someone familiar with the conventional PID governor. The result was the TEPID governor strategy, which was proven to be successful in many challenging applications, with performance superior to many conventional PID governors.

Implementing TEPID

The first commercial installation of the TEPID governor strategy was at the Bhote Koshi plant near Jhirpu, Nepal. The plant is jointly owned by montgomery-watson-harza (MWH) and Panda Energy, and operated by the Bhote Koshi Power Co, based in Kathmandu. The station was built around 2000 and by 2004 the original governors for the two 22.5MW units needed to be replaced due to matters of reliability and insufficient or adequate documentation and manufacturer support.

Aside from the remote location of Bhote Koshi, the main challenge with the work was to integrate the new digital control equipment into the existing Supervisory Control And Data Acquisition (SCADA) system at the station. After the governor replacement, the time to synchronise a unit to the relatively unstable grid (sometimes varying in frequency by as much as ±4 Hz) was reduced from as long as 30 minutes to as little as 30 seconds. Also, the units more accurately maintained their online dispatched generation setpoint, helping to stabilise the country’s very small power grid (approximately 500 MW total load).

Some recent projects where the TEPID governor routine was used include Newfoundland Power’s Petty Harbour and Rattling Brook plants, which were upgraded as part of a general power generation controls upgrade programme.

At Petty Harbour, two units were converted to digital governors from a previous upgrade from mechanical-hydraulic governors to analog electronic governors. The main challenge was to interface the new digital electronic controller to an existing electrohydraulic valve, which required a custom-designed driver amplifier assembly.

Two units were also converted at Rattling Brook to digital governors this time from the original mechanical-hydraulic Gateshaft type governors. The principle challenge was to interface the existing distributing valve to the digital electronic controller via an electrohydraulic pilot valve manifold.

Author Info:

Mr Kornegay also addressed various aspects of these issues at Waterpower XV in July 2007, in his paper: ‘Upgrade or replace? Considerations for Improvements to Hydroelectric Governors and Unit Controls’.


Governor controls – maintaining, improving

The governor system controls the offline speed of the hydroelectric generating unit, and the online output
of the hydroelectric generating unit. The system comprises a governor head (accepts setpoint commands from the unit control
system) and an actuator (mechanically executes
the governing action).
The governor head is a controller that is typically a digital control system, either a commercially available PLC-based system, or a proprietary digital control system. The actuator may be either a hydraulic actuator (typically using an hydraulic pressure supply system) or an electromechanical actuator. An hydraulic pressure supply system provides a supply of pressurised oil to operate a hydraulic governor actuator. A cleanliness control system keeps oil contamination from causing premature wear of hydraulic control components, and improves the reliability of the hydraulic control components.
A Unit Control System controls the starting, synchronising, online operation, and stopping of a hydroelectric generating unit. A Station Control System controls the starting, online operation, and stopping of multiple units within a hydroelectric generating station. A SCADA system is often used as a Station Control System for either single-unit or multiple-unit hydroelectric generating stations. It allows for communication with remotely-located computer-based control systems and Energy Management Systems (EMS).
Factors that can indicate a possible need to improve control equipment at a plant include changes in ability to: meet current control needs; meet water level and flow control needs; meet generation dispatch needs; operate with islanded load; meet efficiency criteria; meet fish survivability criteria; meet offline stability criteria; synchronise to line in acceptable time; meet online stability criteria for pool connection; be remotely controlled and monitored; coordinate with centralised control systems.
The decision process regarding the governors is in two stages. First, is an improvement required for a plant. Second, what level of improvement is needed (i.e. conversion or replacement). The processes of rehabilitation (or overhauling), upgrading (or converting), governors would require a number of factors to be addressed, such as: changes to the functional requirements for the plant; replacement versus retention of the existing hydraulic pressure supply systems; low pressure versus high pressure hydraulic systems; and, converting existing control valves versus replacing them.
Further factors to address in the assessment include: environmental and performance considerations of hydraulic versus oil-free actuator systems; environmentally friendly hydraulic fluids; oil cleanliness control; using dedicated governor controllers versus integrating the governing and unit control functions into the same controller; maintainability of the equipment; ability to make future upgrades; and, proprietary design versus open design controller platforms.
An example of plant assessment (from just before TEPID was available) was for the Raystown Hydroelectric Project (William F. Matson Generating Station) near Huntingdon, Pennsylvania. The plant was built 23 years ago, has two units (7MW and 14MW, respectively), and is owned by Continental Cooperative Services. Primarily, the station is needed to regulate the water level of the lake upstream of the dam and improvement to the system monitoring was needed for when personnel were not on site.
The control system equipment, both hardware and software, was no longer supported by the manufacturer. A control system conversion approach was chosen to replace old digital controls with latest hardware and a SCADA system that included alarm dial-out capability. The existing hydraulic actuator and hydraulic pressure supply systems were retained, as they were in good operating condition. The main challenge with this upgrade project was the integration of the existing equipment to the new digital control equipment. The station water intake control system was replaced with new digital control hardware.
Improvements to the controls of a hydroelectric generating station can result in greatly improved profitability and performance if these improvements are made with the specific needs of the station in mind. The time spent up front in considering the issues discussed above is a good investment to assure the success of the project.

Working with ‘Feedforward’

The concept of feedforward has been used (and misused) in the hydro industry for many years. Through several generations of hydro governors (mechanical-hydraulic, analog electronic, and some digital controllers), faster online response was achieved as required for dispatching generation by switching the online PID gains to much higher values (typically by a factor of 4 or so).
The problem with this approach is that the higher PID gains made the governor inherently unstable. The governor drew upon the large inertia of the interconnected power system to maintain its stability, and each unit connect this way reduces the frequency stability of the grid, although by a generally fairly small in each instance. If enough units on the interconnected power system use the high online gain approach to enhancing the power dispatch response of the governor, the frequency of the power grid can be driven to become unstable.
The instability risk due to such gain-setting of governors can lead to difficulties for smaller power grids. Such difficulties were experienced, for example, in New Zealand during a major upgrade of the hydro governors on the South Island, in Iceland, and in Brazil.
Using feedforward from the governor setpoint to the governor actuator bypasses the governor PID gains, allowing generation to be rapidly dispatched as needed into the power grid without altering the governorâ€â„¢s speed-responsive PID gains from their inherently stable offline settings.
All TEPID governors have setpoint feedforward integrated into the algorithm as a standard feature. There is no reason that any conventional PID governor could not also employ a setpoint feedforward function as well in order to benefit from the stability enhancement of the grid frequency. It should be further noted that the TEPID governor algorithm contributes more stability to the power grid than even a properly-tuned conventional PID governor with setpoint feedforward.
Some other governing strategies include â€Å“lead-lagâ€Â, â€Å“genetic algorithmâ€Â and â€Å“state spaceâ€Â. Mostly, they are at theoretical or early stages of development.