The following are the main insights from our analysis of the RMS portals:

• No unified portal: Typically, it is the manufacturers of the solar pump controller who design the RMS, as these are integrated with the controller. Solar pump installations in Chhattisgarh, since 2016-17, has witnessed seven different controllers. Consequently, the data from solar pumps is hosted across the respective portals of controller manufacturers, making it extremely difficult for the government or any agency to undertake an aggregate analysis or have a complete overview of the solar pumps across the state.
• No standardised data structure: We find that the parameters recorded and the frequency of data captured, widely differs across the portals. Table 1 summarises the data structure in different portals. The lack of a standardised data structure makes it impossible to unify the data, and make an aggregated sense of the state of solar pumps, across the state.

Table 1 The parameters and frequency of RMS data vary between different portals.Source: Authors’ compilation

• All data, no insights: Data is only as good as how it is presented to the user. The existing RMS portals do not provide any aggregated information across all pumps, for they only provide pump-by-pump information. Such information is not useful for the state administrators to understand the big picture about the usage, performance, and maintenance issues at the state- or district levels. The only statelevel data readily available in any portal are the total number of pumps installed and the number of pumps actively logging data in real-time. The remaining data are provided at the individual pump level. Figure 1 summarises the user interface of different portals.

To take a deeper dive into the quality of the RMS data, we selected one among the seven portals; the one with the largest number of registered solar pumps (designated as Portal 1). The following analysis is primarily based on this one portal, but a cursory assessment of other portals confirmed that most of the issues are common across portals.

1. Improper entries: We find that many critical parameters are entered incorrectly, and pump identifiers are missing. For instance, we found that for about 80 per cent of the pumps, the geotag was missing or pointing to the controller manufacturer’s headquarters. Similarly, other critical metadata like the pump owner’s address and the pump and PV array capacity were incorrectly entered for most of the pumps. We notice a similar issue in some of the other portals as well.

2. Poor maintenance and upkeep: We find that a majority of the RMS systems lack proper maintenance, which is reflected in two ways:

• The number of pumps transmitting data has been gradually declining, in most of the portals. Figure 2 shows the number of pumps logging data on any given day. Out of the 1,306 pumps that had logged data at least once in the portal since about mid2016 (when the state solar pump scheme had just begun), only 446 pumps continued to transmit data as of December 2020. Figure 3 depicts the period of data transmission for pumps installed in various years. Occasionally, the pumps go offline in bulk. One likely reason could be that the SIM cards for the RMSs were activated in bulk on a particular date and became invalid due to the lack of timely recharge.
• Across all portals, the number of new pumps being registered showed a significant drop; in most cases, dropping to zero by 2019, although the number of installations in the state has remained at a constant 20,000 pumps a year since 2016-17. In Portal 1, the last pump was registered at the beginning of 2019, although the scheme is ongoing and vendors use the same controller as before (Figure 3). Across all the portals, we could only account for about 10,000 pumps, which is only 16 per cent of the total number of pumps installed. In the absence of proper scrutiny by the state government—which itself is a result of poor portal design—the manufacturers are not complying with the RMS mandate, as specified in their contracts; especially that of providing RMS data during the lifetime of the solar pumps, in an attempt to corner undue profit margins.

3. Data loss: We find that the data logging of pumps is not consistent even between their installation dates controllers against the annual contracts by CREDA, the RMS devices used in 2017-18 were likely to be of better quality than those used in the 2016-17 pumps. This points towards the need for mandating good quality RMS devices and prescribing the limits of data loss for individual devices.

Figure 1 None of the portals is designed to facilitate state-level monitoring of solar pumps

Source: Authors’ compilation

Figure 2 The number of pumps logging data is consistently decreasing over the years

Source: Authors’ analysis

Figure 3 A significant proportion of pumps in the portal stopped transmitting data after the second year of installation itself, indicating non-compliance with the RMS mandate as per the contracts

Source: Authors’ analysis

Note: The number on the y-axis represents individual pumps in their chronological order of installation. Each dot in the image represents a day when the pump’s RMS transmitted data to the server. Blank areas represents days without any data from the particular pump.

Figure 4 Pumps installed in 2016-17 face a higher rate of data loss than those installed in 2017-18

Source: Authors’ analysis

Note: A missing data event is the gap between when a pump stops transmitting data and eventually restarts, signifying some technical issues in between. The duration of this event can vary from a day to months.