Analysis of Primary and Secondary Frequency Control Challenges in African Transmission System

Abstract

This study analyzed the frequency control challenges within the West Africa Power Pool Interconnected Transmission System (WAPPITS) as it plans to incorporate variable renewable energy (VRE) resources, such as wind and solar energy. Concerns center on the ability of WAPPITS primary frequency control reserves to adapt to high VRE penetration given the synchronization and frequency control problems experienced by the three separate synchronous blocks of WAPPITS. Optimizing solutions requires a better understanding of WAPPITS’ current frequency control approach. This study used questionnaires to understand operators’ practical experience with frequency control and compared these observations to field tests at power plants and frequency response metrics during system events. Eight (8) of ten (10) Transmission System Operators (TSOs) indicated that primary frequency control service was implemented in the TSO, but nine (9) of ten TSOs indicated that the reserves provided were inadequate to meet system needs. Five (5) of ten (10) respondents answered “yes” to the provision of secondary frequency control service, while only one (1) indicated that secondary reserves were adequate. Three (3) TSOs indicated they have AGC (Automatic Generation Control) installed in the control room, but none have implemented it for secondary frequency control. The results indicate a significant deficiency in primary control reserves, resulting in a reliance on under-frequency load shedding for primary frequency control. Additionally, the absence of an AGC system for secondary frequency regulation required manual intervention to restore frequency after events. To ensure the effectiveness of battery energy storage systems (BESSs) and the reliable operation of the WAPPITS with a higher penetration of inverter-based VRE, this paper recommends (a) implementing and enforcing basic primary frequency control structures through regional regulation and (b) establishing an ancillary services market to mobilize secondary frequency control resources.

1. Introduction

The frequency of an interconnected system is directly related to the instantaneous balance between generation and load. Frequency response is the ability of an interconnected system to arrest frequency excursions and stabilize frequency when subjected to disturbances such as the sudden loss of generation or load [1,2,3,4,5]. To reliably operate a bulk power system, system operators should maintain system frequency within predefined limits, typically specified in a grid code. In the West Africa Power Pool (WAPP), system operators are required to keep a quasi-steady-state frequency deviation of <200 mHz (49.8–50.2 Hz), as specified in the network standards.

The WAPP interconnected power system comprises of the power systems of 14 countries: Benin, Burkina Faso, Cote d’Ivoire, Gambia, Ghana, Guinea, Guinea Bissau, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, and Togo. The WAPP is established to develop generation and transmission infrastructure and coordinate power exchange among member states. Its broader mandate is to integrate national power systems into a unified regional electricity market with the goal of providing reliable and competitive electricity to citizens of West Africa.

Historically, national TSOs operated their grids independently, based on their own operational procedures, frequency control practices, and grid codes. Prior to the establishment of regional coordination mechanisms, frequency control provision was either non-existent or minimal. Only a few TSOs implemented primary frequency control, while most systems operated without automatic frequency response. The lack of integrated control severely limited regional reliability and system security during disturbances.

Over the past two decades, interconnection efforts have accelerated, leading to the interconnection of the power system of all fourteen (14) mainland ECOWAS countries. The integration of the power systems has also led to the need to gradually integrate frequency control mechanisms such as primary control, fast frequency reserves, and secondary and tertiary control. This is expected to enable a more coordinated and dynamic frequency regulation in WAPPITS. Additionally, the development of a regional grid code consolidates efforts to harmonize operational practices and lay the foundation for the implementation of a regional ancillary services framework.

The currently installed generation capacity is approximately 29 GW with an average generation availability of 17 GW (60%). WAPPITS operates in three (3) synchronous areas, labeled here as Areas 1–3. The WAPP transmission network has more than 42,993 circuit kilometers of high voltage transmission lines with voltage levels of 90–330 kV.

Maintaining frequency within acceptable limits requires a system operator to have adequate available generation resources to continuously keep generation–load balance. Also, the system operator is also expected to have sufficient resources to restore frequency after the occurrence of expected, unexpected, or unplanned events such as generator trips or a major transmission line outage.

In modern power systems, power systems operators have implemented well-established frequency control types such as primary control, fast frequency control reserves, and secondary and tertiary frequency control to ensure system stability following disturbances [2,3,4,5,6,7,8,9]. The two core capabilities consist of primary frequency response, typically provided by automatic governor action on the prime movers of conventional power plants, and secondary frequency response, commonly managed through AGC. These methods are widely adopted in developed power systems, for example, in the continental European grid, the North American Eastern Interconnection, and systems operated by ISOs like PJM, CAISO, and ERCOT—where clear roles are defined for each balancing area, and frequency reserves are procured through structured ancillary services markets.

In contrast, whilst a traditional frequency control structure has been defined in some national grid codes in WAPPITS, and also in a recently proposed regional grid code yet to be adopted, practical implementation remains inconsistent. Unlike developed systems that have implemented well-defined reserve obligations, WAPPITS faces persistent challenges in maintaining frequency stability, particularly due to non-uniform operational practices; a lack of a standardized ancillary services framework, especially frequency control reserves; and the minimum enforcement of frequency response obligations across TSOs. In practice, WAPPITS’ recent operation demonstrates persistent challenges with frequency control [10,11]. As a result, WAPPITS is operated as the three separate synchronous blocks mentioned earlier. There has been substantial discussion about, but no final conclusions on, why frequency control has been a problem in WAPPITS.

From a purely technical perspective, the operation of system frequency control is not only well-understood but also widely implemented globally. From the discussion above, it is apparent that issues in WAPPITS represent a socio-technical challenge where technical, economic, and socio-political factors interact to obscure the fundamentals of WAPPITS’ frequency control issues. Therefore, this study used a socio-technical approach to decompose system issues. We used questionnaires sent to system operators to understand how the frequency control system was implemented in practice and an analysis of field measurements to confirm certain points from the questionnaires with operational data. Combining methods provides unique insights into the operation of a large power system in the developing world that is expected to incorporate VRE generation at penetration rates that have challenged the most developed power systems in the world [12,13,14,15]. The novelty of this paper lies in its focus on diagnosing and highlighting the unique frequency regulation challenges specific to the WAPP grid, which differs significantly from those encountered in well-developed grids. Unlike existing studies that assume certain baseline conditions typical of mature systems, our approach identifies foundational issues that must be addressed before advanced solutions, common in developed networks, can be effectively implemented. This perspective is critical for designing realistic and context-appropriate solutions tailored to the actual operational environment.

This paper is structured as follows: Section 1 presents the introduction, Section 2 outlines the methodology used for this study, Section 3 discusses the results, Section 4 consists of a discussion, and Section 5 presents the conclusions.

2. Methodology

The methodology of this paper includes two parts—a questionnaire to assess the current state of operations in each of the TSOs and a confirmatory study of these responses using frequency control tests and frequency response analysis based on metrics calculated using data primarily sourced from Phasor Measurement Units (PMUs) deployed in the Wide-Area Monitoring System (WAMS). The combination of methods provides a techno-operational view of WAPPITS’ current challenges. The questionnaires provided observations from TSO personnel which are based upon broad experience but are still subject to observational biases by the observers. To confirm questionnaire responses—i.e., to check on biases in the responses—WAPPITS conducted frequency control tests in the WAPP grid by performing governor compliance verification tests on power plants.

2.1. TSO Questionnaires

The primary data source for this analysis consisted of responses provided by ten (10) TSOs to questionnaires covering key frequency control systems. The questionnaire was sent via email to technical representatives in each TSO who were familiar with central control operations. The first author followed up with representatives to assure a response and spoke with several to clarify why the questionnaire was sent, what answers were expected, and why responding was important. During this process, the first author was careful to avoid biasing the answers by suggesting or diagnosing potential responses. While individual TSOs required different levels of engagement, responses were received from all TSOs. Questionnaire responses were received via email.

Once responses were received, the first author coded answers to the questions as follows:

• Primary Frequency Control: TSOs were required to answer “Yes” or “No” to whether power plants provide primary frequency control (PFC) service and whether the reserve amounts are adequate during a typical operating day. A code of “1” was assigned to a “Yes” response and “0” to a “No” response.
• Secondary Frequency Control: TSOs were required to provide a “Yes” or “No” answer to whether secondary frequency control is provided by power plants, whether secondary frequency control reserves are adequate or not, and whether control is implemented manually or automatically. A code of “1” was assigned to a “Yes” response to the implementation of secondary frequency control and “0” to a “No” response.
• AGC Implementation: TSOs were required to provide a “Yes” or “No” answer to whether secondary frequency control is implemented in the TSO control room using AGC. Responses were coded considering both the response to whether secondary frequency control is implemented and whether AGC was implemented, resulting in three codes:

   “1” 

  indicates neither secondary frequency control nor an AGC system was implemented.

   “2” 

  indicates secondary frequency control was implemented but without an operational AGC system.

   “3” 

  indicates secondary frequency control was implemented using AGC.

• Cost Recovery and Pricing: TSOs were required to answer “Yes” or “No” to whether cost recovery and pricing mechanisms are in place for frequency control services. A code of “1” was assigned to a “Yes” response and “0” to a “No” response.
• Penalties for Non-Compliance: TSOs were required to answer “Yes” or “No” to whether penalties are applied for not providing the required frequency control services. A code of “1” was assigned to a “Yes” response and “0” to a “No” response.

The Yes/No responses requested in the questionnaire were intentionally designed to force respondents to use their experience operating the system to make a yes/no judgment of whether a particular subsystem was operationally present. In some cases, this judgment and the stated technical condition of the system match: for example, there may be no AGC system implemented. However, this judgment may also differ substantially from the technically documented or stated condition of the system. For example, power plants may have stated a provision of primary frequency control, but in the experience of the TSO operator, no—or an insignificant quantity of—primary frequency response was available in practice. The coded answers to the questionnaire should therefore be considered encoded judgments by TSO personnel capturing day-to-day operational realities.

Table 1. Heatmap of questionnaire responses related to adequacy of primary and secondary frequency control services.

Question	TSO										Total
	1	2	3	4	5	6	7	8	9	10
Primary FC Implemented	1	1	1	1	1	0	1	0	1	1	8
Adequate Primary FC	0	0	0	1	0	0	0	0	0	0	1
Secondary FC Implemented	0	0	1	1	1	0	0	1	1	0	5
Adequate Secondary FC	0	0	0	1	0	0	0	0	0	0	1
Legend:	Yes	1	No	0

and

Table 2. Heatmap of questionnaire responses related to installation and implementation of AGC and secondary control.

Question	TSO										Total
	1	2	3	4	5	6	7	8	9	10	None	One	Both
AGC Installed in Control Room	0	0	2	1	1	0	0	2	2	0	5	2	3
AGC Implemented for Secondary FC	0	0	1	1	1	0	0	1	1	0	5	5	0
Automatic Secondary FC	0	0	1	1	1	0	0	1	1	0	5	5	0
Legend:	0	Secondary FC not implemented and no AGC
	1	Secondary FC implemented but no AGC (entirely manual control)
	2	Secondary FC implemented and AGC in control room (assisted manual)
		TSO indicated adequate primary and secondary reserves

show the numerical assignments for responses from TSOs.

2.2. Analysis of Frequency Control Implementation with Field Testing

Questionnaire responses were compared with a field test to measure the response of a subset of generation units that were self-identified as supplying primary frequency control services. Testing consisted of exercising the governors of these generating units to determine whether the settings were compliant with the required droop, intentional deadband, and frequency response requirements specified in the applicable grid code. The following parameters were assessed:

• Permanent droop setting—The required droop setting for the hydro and gas turbine generators is 4% and for other generator units is 5%.
• Intentional deadband setting: The required intentional deadband setting is ≤0.05 Hz.
• Primary frequency response—The active power output increase/decrease due to frequency change should occur within 15 s and 30 s from the steady initial condition. In addition, the primary frequency response provided by TSOs should be sustained for at least 20 min after a frequency deviation.

A third-party company was engaged to perform tests by the injection of a test signal to the unit’s governor to simulate a frequency event that would result in an increase/decrease in the power output of the unit. Using the recordings from these tests, droop, intentional deadband, and primary frequency response information was computed and compared to the required settings.

2.3. Analysis of Response of Generating Units During Selected Events in WAPPITS

The performance parameters reported on by TSOs in the questionnaire and measured in the field were then compared to the observed frequency control during three selected events that occurred in WAPPITS. Frequency response curves were generated using data extracted from selected PMUs in the WAMS application installed in the WAPP regional control center. To analyze frequency response effectiveness, the minimum frequency metric, known as nadir frequency (FC); estimated time to arrest frequency (nadir time (tC)); nadir-based frequency response; and the rate of change of frequency (ROCOF) were computed and analyzed [16,17,18,19,20,21,22,23,24,25].

The purpose of this analysis was to confirm the data indicated by the two prior analyses.

2.4. Synthesis Process

The results of all three analyses were analyzed to identify agreement or disagreement between the results of each analysis. This synthesis process identified whether the experience of TSO personnel reflected the observed frequency control and response characteristics seen in field measurements.

3. Results

This section presents an analysis of the results.

3.1. Results of Analysis of Response to Questionnaires

An analysis of questionnaire responses provided insight into the aggregate experience each TSO had with their frequency control reserves.

In Table 1, eight (8) of ten (10) TSOs indicated that primary frequency control service was implemented in the TSO, but nine (9) of ten TSOs indicated that the reserves provided were inadequate to meet system needs. With 80% of the TSOs responding positively to implementing primary frequency control, the grid is generally well-equipped to handle minor disturbances across the service territories of these TSOs. However, the lack of PFC implementation in two TSOs could potentially create localized instability in those areas, making the grid more vulnerable to frequency deviations during generation–demand imbalance. Without sufficient primary frequency control reserves, it is practically difficult to maintain frequency within acceptable limits, particularly at a higher penetration of VRE.

In the absence of PFC, or insufficient PFC resources, generation loss will likely trigger automatic under-frequency load shedding (UFLS) and local blackouts. If load shedding is insufficient or too slow, a total system collapse is likely. TSO responses confirm that the current approach to frequency management in the WAPP region is heavily reliant on under-frequency load shedding to keep the system intact since primary frequency control reserves often experience less generation loss. This approach to frequency control is typically not the best practice from either a customer satisfaction or system control perspective; generally, UFLS should be the last resort for primary frequency response. During some historical events, in the WAPP region, UFLS was unable to deal with fast frequency decline, sometimes leading to partial or total system blackouts [6].

Again, referring to Table 1, five (5) of ten (10) respondents answered “yes” to the provision of secondary frequency control service, while only one (1) indicated that secondary reserves were adequate. A lack of secondary frequency control in 50% of the TSOs poses a risk of sustained frequency deviations in those sections of the grid. Prolonged and sustained frequency deviations shift base frequency from nominal frequency, increasing the risk of unexpected UFLS, damaging customer equipment, and increasing the risk of system instability.

In Table 2, three (3) TSOs indicated they have AGC installed in the control room, but none have implemented it for secondary frequency control—i.e., secondary frequency control is assisted manually by control room operators. Two (2) TSOs indicate that secondary frequency control is implemented but have no AGC in the control room, implying that secondary frequency control is entirely manual. Without automatic secondary frequency control via AGC, system operators must manually intervene to perform secondary frequency control.

Manual intervention for secondary frequency control is largely non-existent in more developed control systems. In practice, manual intervention has a substantial time delay, as operators must observe a generation issue, decide on a course of action, contact generation operators, and wait for the generator(s) to react. As a result, the system reacts slowly to sustained frequency deviations. In some instances, whilst manual interventions are ongoing, additional frequency deviations may occur, leading to further frequency decline, UFLS, and cascading outages. From a control dynamic perspective, the controlled system dynamics—i.e., the power grid—are as fast as or faster than those of the control system—i.e., manual load balancing for secondary frequency control. This is a recipe for instability.

The red box in Table 2 indicates that only 1 of 10 TSOs indicated that either reserve type was both provided and adequate for system needs. Interestingly, one TSO indicated that they had not implemented primary frequency control but had implemented secondary frequency control. This combination is unusual. It is evident that frequency control structure and implementation within the WAPP grid is an important topic that must be discussed and addressed.

All TSOs indicated that there are no cost recovery and pricing mechanisms for the provision of primary and secondary frequency control services. As a result, revenue for generation operators is constrained to payments for energy delivered. Likely, generators may prioritize meeting the demand requirements of their customers, thus maximizing income, rather than reserving capacity for system reliability. While this cannot be confirmed, generation operators may have concluded that revenue lost due to load shedding and blackouts is less than the revenue lost from reserving capacity for primary or secondary frequency control purposes.

Similarly, all TSOs indicated there are no penalties for not providing frequency control services. This is not surprising since a framework and structure that will enable the mandatory or market-based provision of these services are not in place. Regulators cannot penalize power plants for not providing these services.

Summary: Questionnaire responses highlight significant challenges in the implementation of frequency control in the WAPP grid. In most developed grids, power system operations require the availability of well-defined primary and secondary frequency control reserves. These large interconnected systems deploy AGC to restore frequency and maintain interconnection schedules.

In contrast, TSO staff indicate that these reserves and the required control systems are not practically available in the WAPP grid, likely because these reserves are not prioritized. Primary frequency response and reserves are limited, leading to ongoing frequency variability and large frequency deviations during events. Secondary frequency response requires manual intervention by control room personnel. The resulting, obvious, frequency control problems raise concerns about the assumptions being made to increase the share of inverter-based generation in the WAPP grid. Unlike other developed grids where traditional frequency control approaches have been adopted and implemented, this assumption cannot be made for the WAPP grid.

3.2. Results of Analysis of Field Testing of Generating Units

Observations from TSO personnel are based upon broad experience but are still subject to observational biases by the observers. To confirm questionnaire responses, WAPPITS conducted frequency control tests in the WAPP grid by performing governor compliance verification tests. Tests were performed on pre-selected power plants that had indicated they were either providing some limited primary frequency regulation or have the capability to provide frequency response. Power plant selection was not based on the plant’s location in a specific TSO area. The analysis in this section examines how well existing droop, intentional deadband, and governor control modes have been implemented in these generation units; see

Table 3. Droop and deadband of pre-selected existing power plants.

No.	Plant Type	Nominal Power (MW)	Existing Droop (%)	Droop Active	Existing Deadband (Hz)	Deadband Within Standard	Primary FR Within Standard	Governor Control Mode	Setting Effective for PFR
1	Hydro	173	4%	1	±0.1	0	0	Load Controller (Preselect)	Yes
2	Thermal	115	-	0	0	0	0	Load Controller (Preselect)	No—no droop
3	Thermal	131	-	0	0	0	0	Load Controller (Preselect)	No—no droop
4	Hydro	137	4%	0	±0.3	0	0	Opening Regulating Mode OR Power Regulating Mode	No—excessive deadband
5	Thermal	147	4%	0	±0.01	1	0	Droop inactive	No—effectively 0 droop
6	Hydro	120	-	0	-	0	0	Frequency Mode/Opening Mode/Power Mode	No
7	Hydro	92	-	0	-	0	0	Frequency Mode/Opening Mode/Power Mode	No
8	Hydro	108	4%	1	±0.2	0	0	Smooth operation	Yes
9	Hydro	81	-	0	±0.2	0	0	Active power setpoint mode/Gate position setpoint mode	No—looks like it is manual
10	Hydro	94	-	0	±0.3	0	0	MW Control Mode (Primary Frequency Response Disabled)	No—power only mode
Total Count:			4	2		1	0
Fraction		5%	40%	20%		10%	0%
	Total Power:	1198					0
Legend:		Response was ‘No’
		Response was ‘Yes’

.

Six (6) of ten (10) generating units do not have droop settings implemented—i.e., 60% of generating units that have stated primary frequency response will not respond to changes in system frequency. Of the four (4) generating units that had droop settings implemented, two (2) had their droop control active. Therefore, only 20% of the generating units are actively able to provide primary frequency control. These two (2) units have a total nominal power of 281 MW, less than the 440 MW primary frequency control reserves required in WAPPITS. Assuming these generating units agree to provide their entire output for primary frequency control, which is not technically feasible, there are still inadequate primary frequency control reserves for regulation. Therefore, UFLS will be the primary resource for primary frequency response for any disturbance equivalent to the 440 MW primary frequency control reserve requirement. The analysis in the next section indicates that the frequency response during recorded events was likely provided primarily by UFLS.

Similarly, nine (9) out of the ten (10) generating units analyzed have deadband settings outside the predefined value. Analyzing the droop and deadband settings, neither of the two (2) generating units with active droop control had a correct deadband setting—one had a deadband setting of 100 mHz and the other a setting of 200 mHz. This means no frequency control action will be taken until frequency excursions exceed at least 20 times the predefined deadband value.

Considering that the frequency in the WAPP grid should be maintained within ±200 mHz, the governors of generating units should be responsive to events that cause deviations within this range. In addition to decreasing the response from primary reserves, the wider deadband also increases the effective response time of these units. As noted earlier, this contributes to a control system which operates slower than the dynamics of the system being controlled. As a result, the delay in response, coupled with insufficient primary frequency reserves, further increases frequency deviations, fails to arrest frequency decline, and likely triggers UFLS as a primary means of prevent system collapse.

In summary, the assessment of governor control mode reveals that the primary frequency response contribution from a selected subset of generation units is almost non-existent, supporting observations made by TSOs in the questionnaire phase of this study.

3.3. Frequency Response Analysis of Selected Events

A final analysis of field data applied frequency response metrics to recorded data using Grid Digital’s Phasoranalytics™ software version 2.2. Phasoranalytics imports historical phasor data from the WAMS to perform frequency response analysis. Based on frequency plots, selected frequency metrics were used to analyze frequency response. The first author carried out three case studies developed from one generation loss event in each of WAPPITS’ synchronous areas.

3.3.1. Case 1/Figure 1: Frequency Response for 300 MW Generation Trip in Synchronous Area 1

A generation loss of 300 MW in Synchronous Area 1 caused frequency to drop from 50.2 Hz to an initial nadir frequency of 49.3 Hz and then to a second nadir of 49.03 Hz, the UFLS threshold. The initial frequency was higher than the grid standard due to over-generation and the slow reaction of generator controls to regulate frequency. Subsequently, TSO actions (i.e., manual load shedding) to recover frequency worked briefly, followed by an additional frequency drop to 48.63 Hz. Frequency then recovered to 49.10 Hz through further manual load shedding. The total load shed was unknown, as reported by TSOs. This event supports questionnaire responses indicating inadequate frequency control reserves and also likely illustrates that using a combination of automatic and manual UFLS to restore frequency compounded frequency deviations.

Figure 1. Frequency response for 300 MW generation trip which occurred on 1 February 2024.

Figure 1. Frequency response for 300 MW generation trip which occurred on 1 February 2024.

3.3.2. Case 2/Figure 2: Frequency Response for 200 MW Generation Trip in Synchronous Area 2

A 200 MW generation trip in Synchronous Area 2 caused a rapid frequency drop to a nadir of 49.33 Hz that triggered an automatic UFLS of 150 MW which helped recover frequency to ≈50.2 Hz. During this event, the units involved in primary response appeared to have played no significant role in frequency recovery, a strong indication of the lack of adequate primary frequency control reserves in WAPPITS.

Figure 2. Frequency response for 200 MW generation trip which occurred on 11 January 2024.

Figure 2. Frequency response for 200 MW generation trip which occurred on 11 January 2024.

3.3.3. Case 3/Figure 3: Frequency Response for 90 MW Generation Trip in Synchronous Area 3

A 90 MW generation trip occurred on 1 April 2024, resulting in a nadir frequency of 49.24 Hz accompanied by an automatic UFLS of 66 MW, which helped recover frequency to about 50.1 Hz. However, frequency was highly variable for the next 25 min, marked by a frequency decline to another nadir of ≈49.30 Hz. The reason for the decline was unreported but could have been caused by the loss of additional generation units tripped by the frequency or voltage deviations or the re-engagement of load shed circuits before adequate secondary reserves were online. The time between the first and second nadirs was likely sufficient for operational personnel to execute the load balancing actions required for secondary frequency control. Given the extended period with highly variable and decreasing frequency, these control actions were insufficient to arrest frequency instability, likely indicating that the control room personnel were struggling to understand and respond to the situation. This phenomenon is a clear demonstration that a lack of adequate primary frequency control, and to a large extent a lack of AGC for secondary control, makes manual frequency control difficult in a system as large as a WAPPITS synchronous area.

Figure 3. Frequency response for 90 MW generation trip which occurred on 1 April 2024.

Figure 3. Frequency response for 90 MW generation trip which occurred on 1 April 2024.

4. Discussion

Summarizing the analysis, this section outlines the problems identified in the analysis of frequency control challenges in WAPPITS and proposes solutions to address these problems.

The following key problems were identified:

• There is a significant deficiency in primary frequency control reserves, which leads to reliance on under-frequency load shedding (UFLS) for maintaining system stability.
• The absence of an ancillary services market limits incentives for generators to provide frequency control reserves, leading to the prioritization of energy delivery over system reliability.
• The variability in frequency control practices among different Transmission System Operators (TSOs) indicates a lack of uniformity and coordination within the WAPP grid.
• The combination of inadequate primary and secondary frequency control resources, coupled with manual interventions, poses a risk of cascading outages and total system collapse.

To address the identified problems, the following solutions are proposed:

• Development/Implementation of Grid Codes: The uneven implementation of reserves indicates that TSOs should align frequency control requirements in national grid codes with the WAPP Grid Code. Additionally, national regulators will need to collaborate with TSOs to ensure the enforcement of frequency control requirements.
• Implementation of Ancillary Services Market: This study indicates that generators are likely prioritizing energy delivery, as evidenced by persistent over-frequency conditions within WAPPITS. This behavior is likely due to a lack of incentives to provide reserves, as many independent power producers have power purchase agreements to deliver full capacity. In most developed grids, an ancillary services market is used to procure primary and secondary frequency control reserves and is likely needed in the WAPP grid as well.
• Mandatory Provision of Primary Frequency Control: Prior to the implementation of an ancillary services market, frequency control should be improved by implementing a mandatory requirement for power plants to support primary frequency control. The class, type, or size of the plants required to implement PFC capability need to be identified. A framework to periodically test (governor compliance verification test, frequency response tests, etc.) power plants to assess and evaluate the effectiveness of frequency control post-implementation will have to be put in place.

5. Conclusions

This study was conducted against the backdrop of increasing pressure on the grid operator(s) to incorporate a higher penetration of VRE. VRE penetration in WAPPITS is modest with significant growth projected over the medium to long term. VRE typically lacks inherent inertia and traditional governor-based frequency response, which may exacerbate existing issues such as low primary frequency reserves and reliance on manual interventions. Different European demonstration projects have shown interest in the use of battery energy storage systems and other control systems as a possible solution to mitigate the variability in VRE and at the same time provide ancillary services to the grid [26]. However, most of these technologies have been developed on large, stable power systems with robust frequency control systems. The purpose of this study was to clarify the current state of control systems within WAPPITS, in contrast to other internationally recognized control methodologies.

While there is a tendency to use only technical methods to analyze system response, to implement any control system, both technical and social factors must be addressed. Technical systems cannot be implemented without robust regulatory guidance (grid code) and controls (enforcement), and sufficient economic investment is necessary (ancillary markets). Additionally, the practical experience of control room operators is often discounted: after all, they are able to keep the system running a substantial portion of the time, albeit with highly variable frequency characteristics.

This study started with the techno-social experience of control room operators, which indicated a near total absence of primary, and automatic secondary, frequency control reserves and concomitant control systems. Since human inputs may be biased, the subsequent technical analysis was used to confirm the operators’ observations: in practice, no primary response units had workable governor settings, and in typical events, automatic and manual UFLS were the primary means of arresting frequency drops.

Two recommendations can be made for WAPPITS to operate reliably at higher penetration levels of inverter-based resources: (1) adopt and enforce primary frequency response requirements through regional regulation and (2) mobilize secondary frequency control resources and other frequency control products through an ancillary services market. Additionally, when an ancillary services market is in place, TSOs should be obligated to procure and use ancillary services in their generation scheduling and dispatch and during real-time system operations.